Solutions Manual BASIC HEAT& MASS TRANSFER second edition ANTHONY EF. MILs University of California-Los Ang PRENTICE HALL, Upper Saddle River, NJ 07458Acquisition Editor: Bull Stenguist ‘Special Projects Manager: Barbara A. Murray Production Editors: Dawn Blayer, Shea Oakley Production Coordinator: Pat Brown ‘Supplement Cover Manager: Paul Gowrkan © 1999 by PRENTICE-HALL, INC. ‘Simon & Schuster/A Viacom Company Upper Saddle River, NI 07438 All rights reserved. No part ofthis book may be reproduced, in any form or by any means, without permission in writing from the publisher. Printed in the United States of America w 987654321 ISBN O-13-082922-9 Prentice-Hall International (UK) Limited, London Prentice-Hall of Australia Ply. Limited, Sydney Prentice-Hall Canada, Inc, Toronto Prentice-Hall Hispanoamericana, S.A., Mexico Prentice-Hall of India Private Limited, New Delhi Prentice-Hall of Japan, Inc, Tokyo ‘Simon & Schuster Asia Pte. Ltd, Singapore Euitora Prentice-Hall do Brasil, Ltda., Rio de JaneiroPREFACE ‘This manual contains solutions to all 952 exercises given in Basic Heat and Mass Transfer. ‘The format and style is similar to that used for the examples in Basic Heat and Mass Transfer. However, somewhat more participation is expected of the student, particularly in the more advanced exercises. I do not believe that students should consult sample solutions before attempting the exercise themselves. My usual practice is to post solutions of assigned homework (rather than to hand out, or make available for copying), and I encourage students to focus on the essential elements of the sofution procedure, rather than on the detailed algebra or numerical calculations. Those exercises designated by a diskette logo have been solved using the computer software, though many are suitable for manual solution as well. Lists of answers are provided for all appropriate exercises, Odd- and even-numbered exercises are listed separately; the instructor may choose to provide either list (or excerpts) to students so as to assist self study. The use of a three-hole-punched loose-leaf form for the ‘manual was chosen to facilitate copying, and also to accommodate updating. I was assisted in the preparation of this manual by many of my past and current students, including E. Myrhe, R. Tsai, S. Kim, A. Gopinath, S. Hiebert, F. Chao, M. Friedman, C. Yuen, B. Tan, B. Youn, H. Choi, B. Chang, P. Hwang, M. Tari, J. Sigler, M. Fabri, and A. Na- ‘Nakornpanom. I gratefully acknowledge their help. Special thanks go to Mrs. Phyllis Gilbert: without her expertise in word processing the production of this manual would have been impossible. AF. Mills May, 199824 23 2-11 213 215 ANSWERS TO SELECTED EXERCISES (A) ODD-NUMBERED EXERCISES Chapter 1 (i) 83.8 MW, (ii) 33.9 ke/s (i) 14.9, Gi) 13.6 cm, (iii) 1.14 10.1 Wim’, 18.0°C and 2.0°C 1.22 cm 39.6 W (34.0 kcal/h) ( 8430 Wim? K, (ii) 42.1 W/m? K, (iii) 97.5 Wim? K 668 W 43.3°C @ 207 W, (i) $12.1 5020 W 66°C 12.5 KIW, 4.995x10? KW 1.33x10° s -29.P°C, 32.9°C 20°C 3.31 Wim? K ii) 3640 W @ 270 K, (i) 274 K, (ili) 276 K, 277K 630 s 48h @ 9170 Wim? K, (ii) 0.068 17.2 min (051°C, (i) 30°C Gii) 0.614 K, -1.509 rd; 2.0 K, 316K 50.1°C, 2.66 h 433s 19.9 s, 5.5 KW (i) 6.04 W, Gi) 37 min (i) 0.155 [W/m? KT", Gi) 0.258 clo, (iii) 6 W/m? K 37d Chapter 2 T = ~(/a-T,) + [We Dgky Mktg) 19.2°C 246 Win? K @ 329 Wim, 3.80 Wim Ty)? -T? + Qla-2y)T, ~ 2axhkga] (kytkp)/2 Gi) 31.9 Wim217 219 2-21 2-23 2-25 2.27 231 2.37 2-41 2-43 2-45 2-47 2-53 2.55 257 2-61 2-67 2-69 21 2-77 2-81 2-83 2-87 2-89 2-93 2.95 2.97 2.99 2-101 2-103 2-107 2-109 24111 2115 2417 2-119 3-5 3-7 3-1 3-17 3-19 3-21 0.12 Win K 0.134 Wim K ii) 870,000 S/yr 765 Wim (i) 646 Wim, (ii) 630 Wim, (ii) 662 Win No 19 mm (i) 2kMhy, ii) (3k/20? 2.29 em @ 4.70 W, (ii) 5.75 MW/m", (iii) 1.47 kWim? 841A @ 165 MWim’, (ii) 93.3 W 2.42 m 705°C 185°C 14.3 kW 49.4 Wim? K Yes @ 1.83 em, (ii) 3.42 om @ 58.7 Wim K, (i) #05 K 219 W 959 Wim? K 28 cm 57K 211W 221°C (ii) 7% 256A 53.5 W 323.1 K; 516.2 Wim? @ 4.68 W, (ii) 5.66 W (tem 8); 5.85 W (Item 9); 5.86 W (Item 10) 0.10 (10% of bare tube value) 0.360 440% @ 1918 W 0.132 (13% of bare tube value) Chapter 3 2030 K/m 359.8 K Gi) 365 K Gi) 3908 K Tray-Tay, = 0-3660,0/K (9.34 W/m, (i) 9.26 Win 489 K3.25 3-27 3-29 331 3.35 3.39 343 3-45, 3-49, 3-51 3-53 355 3-57 3-59 361 3-65 3-67 3-71 3-73 3-15 3-77 3-83 3-85 3-127 49 411 413 415 417 4-19 4-23 4-25 431 439 441 4.43 4-49 451 4.53 455 457 103x107 K/m @ 178 Win, Gi) 176 W/m 42.0 W @ 76.1 Wim, Gi) 138.4 Wim 502 Wim? 0.40 m 29.4°C; 189 degrees 12.6 Wim K © 969°C, (ii) 91.6°C, (ili) 70.5°C, (iv) 63.2°C 406 Wim? K 1430 s ( 38.1 s, (ii) 118 5, (ill) 148 5, (iv) 94 22s 142s 238 s, 3021 s 2840 s 281K 4.44 by; 59% () 244°C, Gi) 145°C, (ii) 147°C 2100s 4200 s (i) 86.6°C, (i) 100°C 6.580 84x10 m/s 54.1°C. Chapter 4 365.6 K 170 W/m 13,920 Wim? K; 3.98x10° Pa/m 18,020 Wim? K; 3.92x10* Pam 28.1 Wim? K; 5.09 Palm G) 12.06; 61,200 W/m? K, (ii) 14.07; 71,350 W/m? K 318.1 K 920 Wim? K; 13.85 kPa 222.5 Wim? K; 372.3 635, 631, 291 Win? K 0.294 N/m; -951 W/m 3370 W 312K 11.7 cm/s 328.5 K; 353.8 K (@ 0.615 N, (ii) 1.48, (iii) 1070 kW 19.5 cents/m day 058 W 9.62 kPa459 4-63 4-65 4-67 4-69 4n 4-73 4-79 4-81 483 485 489 491 4.93 497 4.99 4-101 4-103 4-105 4-107 411 4113 A117 4119 4-125 4-127 4-129 4-133 4-135 5-11 5-15 5-7 5-27 5-29 472°C 138 Win 530.8 K 3.47 mis, T,(0.1 m) = 322.3 K 26,000 K. 549 55 9.16 kg Air: (i) 0.091 mis, (ii) 100 (i) 55.8 W; 24.6 W, Gi) 631 120W 4.09 W from each side 804 s, 8.78 kg 99.0 min 118 W 0.00525 m, 10.94 W/m 18.4 m = 0.86, k = 0.881 W/m K 32 min G) 119 Wim?, (ii) 114 Whn?, (ii) 113 W/m? 264 Win 3.73 k; 5.63 W 42,900 W; 24,700 W (i) 1.6 mis, (iv) 7.9 mis 101 Wim? K; 712 Pa 17; 36.7 Pa 2312 Palm; 134 Wim? K; 413 m”; 41.3 m 363K (i) 154.6 W, Gi) 10.5 kW 38.2% 6,860 Pa/m; 54,500 W/m? K 3.57 mm; 15,590; 9.54x107; 359 N/m?; 3.26x10°; 3.01x107 Wm? rater: (i) 9.9x10% mis, (ii) 0.56 K 212 W Chapter 5 (i) 1.27 W, Gi) 59.7 W, (iii) 34,700 W (i) 4900 kW, (ii) 0.72 0.919, 520.3 Wim? K, 506.8 Wim? K T = T, + (/4)(uaz/i[5 +4(y/b) -(y/b)"] T = T, - (aug -@/R)) -40.7 Pal. 7.314(D2)RepPr 36.0 W; 74.9 W 1.73x10* W/m?; 21% increase Cy = 0.654Re,"; Nu, = 0.350Re,!Pr!6-5 67 6-13 6-15 6-17 6-19 621 6-23 6-25 6-29 631 6-33 6-35 6-37 6-39 6-43 6-47 6-49 651 6-53 6-59 6-61 6-63 6-65 6-67 6-69 671 673 6-75 6-77 6-79 6-83 6-85 6-87 6-89 6-91 6-93 6-95 6-99 6-103 6-105 6-107 6-109 6-111 Chapter 6 @ 9 KWim?, (ii) 21 kW/m”, (iii) 1.063 kW/m’, (iv) 22.06 kW/m? 0.0775; 0.155; 0.155, 0.0775 @ 25.0 W, (ii) 85.1 W; 5.7 W 0.25, 0.0625, 0.0278 799 W, 1346 K, 1157 K (i) 62%, (ii) 32% 1059 K; error reduced from 186 K to 109 K Zero 609.5 K @ 1473 W, Gi) 432 W 2.62 m 264 kW/m? 14.2 MW (@) 214.3 kWhm?, (ii) 210.8 kW/m? (@ 4.33x10°, (ii) 4.312x10°, (iii) 4.308x10°, (iv) 4.308x10° W/m 4.62 W; 28.79 W; 7.99 W 70.9 Wim @ 66.89 Wim, (ii) 66.86 Wim 1.18 W 441 kWin () 760.2 K, (ii) 582.9 K, (iii) 27.6 W, (iv) 100 W 17.3 kW 4.23 W 0.74; 269 K (Brunt correlation) 202.9 Wim? 370.5 K; 349.2 K 0.9385 37 Wim? 1.248 Wim? ‘Yes; maybe 247.6 K 363 K, 369 K 321.0 K, 45.7% Yes H/(R74H7/4)'? (i) 0.405 Wim, (i) 0.398 Wim, (ii) 0.398 W/m (@ 163 W, (ii) 260 W () 19.7 kW/m, Gi) 30.6 kW/m (@ 0.071, (ii) 0.24 B75 kW () 809 Wim?, (ii) 0.13, 1.6°C 0.027 0.446-113 6-115 6-117 6-121 6-123 6-125 6-127 6-129 6-131 6-133 6-135 6-137 6-139 TA 13 7:13 TAS TAT T19 1-25 733 131 7:39 TAL 7-43 7-45 TAT 753 1-55 157 759 1-63 1-65 1-61 81 83 85 8-7 8-11 1745 K @ 0, Gi) 0, (ii) 0.873 @ 0.748, Gi) 120°C 0.190, 0.370, 0.544 1952 kW/m 110.2 kW/m Quad = 17.2 KW; Quoq/L = 1.6 kWh aaa = -30.3 kW/m?; = 23.8 kW/m? () 4.56 MW, (i) 3.14 MW 2220 kW/m? aaa = 16.5 KW”; eoqy = 49.5 kW/m? 630K 4.95x10* Wim? P Chapter 7 1.32x10* W/m? K; 1.18x10° kg/m s 2781 Wim? K; 5.79x10° kg/s 0.221 @ 1.552x10" Wim? K, 3.81x10" keys, Q/L = 4180 W/m; ri/L = 3.71x10* kg/m s 1840 W/m? K; 7.59x10° kg/m s 5.52x10* W/m?, 4.29x10" kg/m s Gi) 3692 K 211m 20.7 W, 50°C (i) 15.7 em’, (ii) 390 K 59.6 om* () 1.01 K, Gi) 2.93x10° Win’, (ii) 440.9 K. (@) 131 Whe? K, Gi) 2.42x10% Wim? R-12: 2.41x10° W/m?, R-134a: 3.10x10° Wim? 41 Kis 132.9 Wim? K, 18.8 W (i) 200.1 W, 140.7 W, (ii) 200.1 W m (i) 3.84x10° kg 294 Wim? K 640 W Chapter 8 No 19.4 kg/s 2.80 Wim K 370% 135 (ii) 1.015x10' W/m? K, 2.49x10% ke/s813 B15 817 819 8.21 8.23 8-25 8-27 8-29 831 8-33 8-35 8-39 8-41 8-45 8-49 8-51 8-53 8-55 8-57 8-59 8-61 8-65 8-67 8-69 8-71 8-73 8-75 8-77 8-79 8-81 8-83 8-85 8-91 8.93 8-95 897 8-99 93 9-5 2.22x107 kg/s; 28.4 m (321 K, Gi) 1.61, Gii) 23.8 m; 14.3 MW @ 1.40 m* 6670 m? 1260 W 0331x107 kg/s 27.34°C 388 K 0.03185 kg/s; 8.53 m (i) 373 K, Gi) 0.647, (iii) 1.81 (i) 619.6 K, (ii) 319 m* 217°C (i) 62.2 m?, (ii) 26.4 m?, (ii) 32.5 m?, (iv) 27.5 m? 2.50 m (@) 423 m?, (ii) impossible, (iii) 698 m*, (iv) 581 m? 242°C 3.47x10° (W/m? Ky" 169.7 m? 52 2.05 ke/s @ 35.6°C, (ii) 4.35 ky/s 703°C (i) 2.338 ke/s, (ii) 26.5 m? 101, 2.76 m 83.6 m? (i) No, (ii) 144 m*, (iii) 22.5% Both fluids unmixed 26.3 m* L = 3.624 m, H= 0.607 m 598 Pa p= 1.2x10° m, L = 0.308 m, 0.291 m 747 m?, L = 0.47 m, H = 1.59 m 48 m, L = 0.353 m H 11W H = 0.361 m, L = 0.340 m 77.8 m®; 37,500 $/yr 12.6 m*; 17,600 S/yr 51.1 m?; 78,460 S/yr Chapter 9 @ 27.8 kg/kmol, (ii) 299 Ikg K, (ii) 127 ppm by mass (i) 0.340, 0.646, 0.014; 68.9 kPa, 130.9 kPa, 2.84 kPa; 0.252, 0.746, 0.0012, (i) 24.24 kg/kmol, 343 Ihkg K, (iii) 0.984 kg/m’, 4.06x10? kmol/m’, (iv) 0.248 kg/m’, 0.734 kg/m’, 0.0012 kg/m’; 1.38x10” kmol/m’, 2.62x107 kmol/m’, 5.7x10" ‘mol/m’9-7 9-11 913 9-15 9-17 9-19 921 9-23 9-25 9-27 9-29 9-31 9-33 9-35 9.37 9-41 9-43 9-45 9-47 9-49 951 9-53 9-59 9-61 9-63 9-65 9-69 9-73 9-15 9-83 (@ 3.21x10°; 1.804x10°, (ii) 10.56x 107; 4.32x10%, (iii) 2.53x10%; 1.340x10%, (iv) 0.25; 0.26 2.51x104 m? 1.15x107 kg/s 7.81x10"? kmol/s 5.9 m'/day @) 3.34 h, Gi) 13.4 b, Gi 2.79x10° kmol 140 hb 2.11x10" s (8 months) iii) 0.15 mm, (iv) -19.2 m/s, (v) 13.6 s os71 2.28x10° kg/s 1.94x10° s (32 min) 8.92x10* kmols; =0.1 mm (i) 21.9 s, Gil) 1.20 m, (ii) 20.7% 1.08 mg/h @) 9850 W, (ii) 51,300 W (@ 174 W, Gil) 970 W, Gi) 174 W 1331 K 5.9 kW; 262 KW; 210 KW, (ii) 60.2 KW; 179 kW; 185 kW 1.59x10° kg/m? s (@ 302.5 K; 9.16x104 kg/m? s, (ii) 302.7 K, 1.27x103 kg/m? s Gi) 114 min Yes 4.43x10" kg/m? s Gi) 8.01x10" s, Gi) ~0.8 mm 7.98x10° kg/s (@ 7.6%, (i) 137% 5.18x10° kmol/m> (i) 7.49 avs, (ii) diffusion controlled ) 6.04 m/s 9.27 ms, (ii) for Hy; 0.240 kg/m? s; 0.300 kmol/m? s, 0.0243 kmol/m” s At 290 K: 0.01895; 3.52x10° kg/m? s; 3.47x10° kg/m? s, 0.9905, ete. @ 4.8 giday, (ii) 3.05x10° Pa, (iii) 5.35x10 m/s, -1.04x10° Pasm (i) 137.6 g/day, (ii) 18.3 Pa, (iti) -4.46x10? m/s (i) 47.5 g/day, (ii) 1.03 Pa, (iii) -4.66x107 m/s () 4.61, (i) 7.07x10? kg/m’ s, (iii) 2.61 m?, (iv) 14.110? m, (v) 1.034x10%; 1.034x10% 2.73. em e.g. D = 14 cm, L = 23.5 cm, AP = 901 Pa 3.8 cm (0.889, (ii) 2.20, 5710 Wim? K; 0.12 K 30.85°C 0.0323, (iv) 3.59x10% ke/s9-113 OLS 9-119 9-121 9-123 9-125 LIG = 0.57, Ny = 3.4 (i) 5.9%, Gi) 0.498, (ii) 3.88 kg/m? s vs. 4.70 kg/m? s 95.2. KW, 0.307 kg/S, my ou = 3315x107, Toy = 30.8°C, 2.94x107 kp/s (ii) 2.262, 14.1°C, Gi) 0.413 m (i) 4.27 m, (ii) 0.85 m, (ii) 592 W vs. 304 W (ii) 4.96, 10.0°C, (iii) 1.30 m12 14 16 1-10 1-12 1-14 1-18 1-20 1-22 1-24 1-28 1-30 1-32 1-34 1-36 1-38 1-40 1-42 1.44 1-46 1-48 1-50 1-56 1-58 1-60 1-62 1-66 174 2-10 2-12 214 216 2.20 2-22 (B) EVEN-NUMBERED EXERCISES (i) 27.4 mis; 263 mm/s, (ii) 7970 W 14.0 W; 38.4°C 1.44 kg 25cm 57.2 Whn?; 992 W/m, 2.63x10* W/m*; 2.92x10° Wim? 20.68 0.161 mW Gi) -163 W TEC (i) 72.0 mW, (i) 4.28 kW, (ii) 1.88 kW, (iii) 036 KW 344 kg/day, $2945 (i) 2.32 [Wim? Ky", (ii) 143 W 0.073 m (i) 76.7 kW, Gi) 3.73 m 89.5.< Q <125W (i) 111 Wim, Gi) 95 W/m? (@ 255 K, (ii) 270 K, (iii) 272 K; 276 K @ 178°C, Gil) no 222s 71.36°C, 0.114 s; 2.71°C, 0.228 s3 0.996°C, 0.342 s (i) 2060 s, (ii) 2150 s 518s 900 kg K @ 296.7 K, Gi) 356.1 K 476°C 3710 s () 17.6°C, (ii) 35.4 (i) 1078 mbar; 826 mbar, (ii) 374.8 K; 367.5 K (@ 0.111 pVAWAn?), (ii) 0.026 K; 0.26 K, 2.6 K Chapter 2 (@ 0.0066 m, (vii) 92.8 m @ 0.0180 W/m K, (ji) 0.0221 Wim K, (ii) 0.0180 Wim K, (iv) 0.101 Wim K (i) 4.67%, (ii) 14.6% 448.2 K 10.17 kW 121.0 Win? 139.4°C 339 Wim; 98.4 W/m 1049.8 K 825 Wim2.24 2.26 2.28 2.30 2:32 234 236 2-38 2-40 2.42 2-46 2-48 2-50 2-56 2-60 2-64 2-68 2-70 272 2-74 2-82 2-84 2-86 2-88 2-90 2-92 2.94 2-98 2-100 2-102 2-104 2-106 2-108 24112 2414 2-118 2-122 2-124 2-126 38 3-12 3-20 604 Wim Tyy = 291.5 K 19.1 mm 4.6 mm; 12.1 Win (2.7 mm, Gi) 3.3 mm. 417.8 K; 348.5 K; 15.2% 58.9 W 420x104 kg/s 35 cm (ii) 84.4 Wim? K (@ 5.18 W/m, (iii) 6.68 Wim 6.6 cm 1426 K 247.4°C (i) 91.2 K, (ii) 90.6 K, (iii) 90.6 K 1870 Win 119K 8.73. W 47.9 W; 9.0 m‘/min 20.8 W 50.79C 56.3°C 82cm 331K 224 mw (38.4 A, Gi) 40.5 A 54.8°C 119.6 K 353.1 K 26K 2.08 W, 52.0°C @ Reie/Rgare = 0.604, (ii) 1.25 84.0 Wim; 0.0554 kg/m 115.4 Whn? K Gi) 371.7 K 15.1 W @ 3.48x10° m, (ii) 9.80x10° m, (iii) 1.90x10? m 254 Win 0.737 W, 192.4°C, 177.5°C (@) 0.300, Gi) 0.485 Chapter 3 () 24°C, (i) 1.93 W 27.19°C 89.7°C3-22 3-24 3-26 3-28 3-30 3-34 3-36 3.38 3-40 3-42 3-46 3-48 3.52 3-58 3-60 3-64 3-66 3-68 3-10 3-72 3-74 3-76 3-80 3-82 3-88 3-126 42 48 410 412 414 418 4-20 422 424 4-26 4-28 4-30 432 4:36 4.40 (27.7 kW 23.8°C 64.2 Wim. 7.15x10° kg/s 1025 m 36.28 -2.23°C; 0.011 m 34 K; 101 degrees 7488 0.251 m (i) 43 m; 0.23 m, 8.22 m 200 Wim? K Steel: (i) 380 K, (ii) 300 K; Brick: (i) 677 K, (ii) 323 K 861s () 383 s, Gi) 3630 s 579s Gi) 265 s, Gi) 3660 No: steady state is approached too slowly 121x108 s 3345 (i) 23.9°C, (ii) 55.8°C (i) 90.5°C, (ii) 100°C (i) 547 s, Gi) 462 8 5.184 14x10° m/s 769 i) 5.2 m; 0.27 m Chapter 4 19.4 ms; 305.9 K (@ 0.0381 mm, Gi) 7.89 mm Nap = 0.416Re," 156.8 Wim? K; 5.68x10* Palm 16.7 Wim? K; 11.0 Pa‘m 524 W (i) 12.4; 14,700 Wim? K, Gi) 14.4; 17,100 W/m? K 2.21x107 kg/m s; 2.68x10 m’/s; 212 W/m? K 272 Wim? K; 291.9 K; 0.817 MPa 220.7 Wim? K; 372.0 K; 6.52x10* Pa 24cm 2.102x10* Wim? K, 500 K, 1.77x10° Pa 5.75x104 Pa, 4857 Wim? K, 276.3 K 3.35 cm ( 82.0 kW, (ii) 154 KW4-42 4.48 4.50 452 454, 4-56 4-66 4-70 an 4-74 4-78 4-80 4-82 484 4-86 4-90 492 4.94 498 4-100 4-102 4.104 4112 4114 4-118 4-120 4.122 4.124 4-126 4-128 4130 4-136 5-4 56 5-10 5-18 5-20 524 5:26 5-28 5-32 5:34 352 W (303 kcal/h) 165 kW 174 Wim? K (i) 0.0474 mis, Gi) 4.9%, (ii) 15.6 wm (@ 356 K, (ii) 10.62 W 371.9K 36.0 Wim? K @ 7.60 Wim? K, (i) 634 Wim? K 329.4 K 12% (i) 3.54 W; 2.58 W, (ii) 28.1 W; 17.5 W 38.6 W/m; 315.5 K 85.3 W 342.9 K 353.3 K Vertical, 18.7 min (i) 375.6 K, (i) 1002 K cgay = 42.4 W, Qnyy = 44.0 W 318.0 K 63.2 Wim width 3560 W 4.17 k; 47.6 Win 2.15x10* Wim? K; 6160 Pa 44; 913 Pa 303.2 K; out of Re range 5 2750 Wim? K 35.4 m 340.2K Chapter 5 M = 5: (1126 K, (ii) 718 K T23.0K 5.385 7093 Kim (i) 18.95 Wim? K, (ii) 4.02x10* Win Cy = 0.685Re!? @) Nu, = 0.886Pr,", (ii) Nay = 1.128Pe,""; Nu, = 1.329Pe,!? C;, = 0.730Re,"”; Nu, = 0.393Re,"Pr'? 1.38 cm; 1.012 m/s; 14.5; 3.97 Wim? K; 79.4 W/m? ‘Nu, ~ 0.503(Ra,¥)"*5-36 5-38 6-8 6-12 6-16 6-18 6-20 6-22 6-24 6-26 6-30 6-32 6-34 6-36 6-38 6-40 6-42 6-50 6-54 6-58 6-62 6-64 6-66 6-68 6-70 6-72 6-74 6-82 6-84 6-86 6-88 6-92 6-94 6-96 6-98 6-100 6-102 6-104 6-106 318.8 K 334.2 K Chapter 6 0.0338; 0.045; 0.0675; 0.0225 2.36 kW; 1.57 kW 980 W; 1262 K 141.6 kW/m; 1391 K (@ 0.839 W/m, (ii) 0.40 Wim (® 1099 K, Gi) 1161 K, (ii) 1188 K 320.035 K (negligible error) 0.517 @ 9.02 kW/m, (ii) 7.77 kWh (@ 0.144 (Whm? Ky}, (ii) 7.2 mm (@ 442 W, (ii) 397 W, Gii) 153 W, (iv) 153 W 424W @ 20.2 kWhr’, (ii) 56.3 KW/m? 451m 136 Wim 283.5 kWim?; 141.8 kW/m? (@ 139.4 Whm?, (ii) 11.2 W/m? (@) 480 kW/m, 2360 K 168 W, 660 W, 50.4 W 18.2 KW 6.08 m 1903 W 216 K 0.67; 259 K (Brunt correlation) @ Yes, (ii) Yes (frost) G@ 299.0 K, Gi) 302.7 K, (iii) 309.9 K 299.6 K 319.8 K 5.52 m 573K (i 318.3 K, (ii) 299.1 K, (ii) 319.2 K, (iv) 313.2 K (i) 557 Wim’, (ii) 491 Wh’, (iii) 484 Win’, (iv) 467 W/m? 385 W (i) 1.96x10° kg/s, (ii) 1.89x10° kgs, (iii) 1.89x10° kg/s @) 337.5 K, (ii) 300.0 K @ 29.7-76.3 ym, (v) 0.47-1.21 pm G@ 0.046, Gi) 0.38, (iii) 0.27 0.036; 0.043 0.602 0.268; 0.25 0.356-108 6-110 6-112 6-114 6-116 6-122 6-124 6-126 6-128 6-130 6-132 6-138 72 Tl 7-16 78 730 734 1-36 7-44 1-46 7-50 154 1-56 7-58 810 8-14 8-16 8-18, 8-22 8-24 8-26 (i) 0.272, (i) 0.267 @ 0.27, (ii) 27s 109.8 kW/m, 0.506, 1254 K 49% 7.2 MWim?, 21.6 MW/m?, 31.1 MW/m? (@) 64.8 kW/m, (ii) 49.7 kW/m? i) -836 Kis, (fi) -108 K/s Grog = -24-9 KW; qoqy = -54.0 kW/m? 131.5 kW; 174.1 kW 83.1 KW; 94.4 kW 0.155 MW/m? (i) 1.02910 W/m?, (ji) 1.071x10* W/m? Chapter 7 (i) 0.157 m, Gi) 0.113 m, (ii) 0.030 m, (iv) 0.001 m, (v) 0.089 m, (vi) 0.139 m 4311 Wim? K; 4.63x107 kp/s 1.675x10° kg/m s 3.79x10° W/m? K; 5674 W/m? K; 0.0396 kg/m s 23.5% increase 370% (i) 1.12x10° kg/s, (ii) 2.97x10* kg/s (i) 1.093x10° W/m’, (ii) 427.2 K (@) 1.15x10? m?, (i) 315.5 K 5730 Wim 143.6 W/m? K, 2.03 W; 132.0 Wim? K, 29.9 W naw 381.1 K, 748.2 K (i) 637.1 W, (ji) steady operation impossible 3760 W 635 W Chapter 8 Yes (i) 26.2 W/m? K, (ii) 7% 14.9 Win? K 330.3 K; 2.20 kp/s 2128 Wim? K 341.0 K; 0.650 kgs (® 312.3 K; (ii) 0.647, (iii) 1.04, (iv) 6.08x10° WK lim 287.2 K; abort design 589.6 K (@ 15.1 m, (ii) 6.68 m 33.3 m?, 0.444 kg/s8-28 832 8.34 8-36 8.38 840 8-42 848 8-50 8-52 8-54 856 8-58 8-60 8-62 8-64 8-66 8-72 8-74 8-80 8-82 8-84 8-86 8.90 8-100 92 9-6 9-10 9-12 914 9-16 9-18 9-20 9-22 9-24 9-26 9-28 9-30 @ 06, Gi) 0.916, (iii) 353 m 11.3 m (@ 10.91 m, (ii) 11.77 m? 3002; 2.16 m; 1.30x10° Pa 0.70; 147.5 Whm? K; 1.204 m? 0.634; 0.677 (6.4% discrepancy) 0.454 m (i) 43.8°C, (iii) 46.5°C, (iv) 15°C 622 m?; 737 m* Gi) 717 m?, Gi) 745 m?, Gil) 741 m? 124°C 2.11 ke/s 1.53°C, 42.21°C, 9625 W 2120 m? (@ 357 WK, (ii) 9.08 m 41% @ 49.0°C, 0.9231, 6.14x10° WiK, (ii) Infeasible, (iii) Feasible 554, 4.61 m 308.3 Wim? K 0.89 m 140 Pa @ 2.68, Gi) H = 2.5 m, L=0.29m D= 734 mm, L = 1.30 m H = 0.657 m, L = 0.450 m 86.2 W ( 2.815, (ii) H = 0.511 m, L = 0.513 m 290 K and 120 kPa Chapter 9 (@ 0.023; 0.228; 0.749, (ii) 0.886; 0.089; 0.027 () Yes, (ii) No, (iti) No ( 0.368; 0.128 kg/m, (i) 166.7 kPa; no change, (ii) 249.8 KPa; 0.480; 0.208 kg/m () 1.22x10° ki/kmol; 2.19x10* m/s, (ii) 9.28x10 m/s Gi) 161x107 kmol/s 7.62x10°? kgls 2640 days (12) H2iA)expAPD 2V/4L?) @ 1.79x10?, (i) 7.10x107/t"? kmol/s G) 6.1x107 mis (@ 0.611 mm, (ii) 0.093 mm 1985 s (33 min) 1.81x10°? kg/m? s i) 0.114 m/s, (ii) 2.12x10? m/s @ 147x107 kmol/n s, (ii) 1.72x10%, (iii) 4.56x107 kmol/s9-38 9-40 9-42 9-44 9-46 9-48 9-50 9-52 9-54 9-56 9-58 9-62 9-64 9-68 9-70 9-72 9-76 9-80 9-84 9-86 9-88 9-92 9-94 9-96 9.98 9-100 9-102 9-104 9-106, 9-120 9-122 9-124 (@ Yes, (i) Yes 0.1119, 108.3% 174s 0.0558 (i) 0.0297; 72.1%, (ii) 0.0301; 71.9% 2.06x10° kg/m? s 3.56x10° kg/m? s 0428x107 kg/m? s 1kW No (7.7 h) (i) 1562 Wim’, (ii) 41.5°C (ii) 876 s, (iii) 15.1 h 0.16 liter/mile 0.196 s 295.35 K (i) 0.49%, (ii) 0.59% 8.75x10° m*/m? 2.80% Gi) 7.2 kmol/m’ s, 2x10° m 25.6 W X, = 0.140.75exp(-17502) (i) 1.08x10° g/day, (ii) 9.07x10* Pa () 0.1061 g/h, (ii) 0.1072 gfh, (ii) 0.1063 gf (@) 25 mY, (ii) 37.6 m, iii) 40.1 m? eg, D= 10 cm, L= 14.0 cm 2750 eg. b= 0.6 mm, L = 9.4 cm, AP = 1460 Pa W= 0.55, H= 1.09 m, L=091 m eg. p= 7mm, W = 0.5 m, L = 1.09 m, AP = 46 Pa 97m @ 247°C 1.30 m (i 18.008°C, (ii) 217 WExercise 1-1 Given: Five linear ordinary differential equations. Required: Solutions. @ sBy=0 Assume a homogeneous solution of the form y,(#) = Ae”, and substitute into the differential equation, 4+ B=0; hence 4=-B yx) = Ae As for case (@) we have yy(x) = Ac™ = Ac ‘We now need a particular solution, say yx) tion, BB+ 0.=0; hence B = -o/f = constant; substitute into the differential equa- 2 Gi) A yn0% yaQ) = Ae™ PH =O; s= 22 yx) = Are™ + Age™ = By sinh Ax-+B,cosh Ax 2, vy £2 422y=0; y= Ae™ o -#B+0=0; B=+o/n? yO) = Aye + Age™ + off. Comments: 1. Apply initial or boundary conditions to obtain the constants A. Yp@)=B 2. ‘These differential equations play an important role in heat transfer analysis.Exercise 1-2 Given: Inlet and outlet temperature and pressure for a helium flow through a low pressure heat exchanger, Required: (i) Inlet and outlet velocities, and Gi) heat exchange. Assumptions: 1. Steady flow. 2. Perfectly insulated heat exchanger. @ The inlet and outlet helium density can be obtained from the ideal gas law. PiM __(1000x4) RT, (8314X50) Inlet: py = 62x10 kg/m? pp = eM. 73004) 3 ohn? Outlet: P2 = “ep, = (314,350) ~ LOOxTO™ kes ‘The helium velocity V is obtained from rh = pVA,, where A, is the flow cross-sectional area. Inlet: Vj = ti/p A, = (Sx10-°)(9.62x10-°)(0.019) = 27.4 m/s Outlet: V2 = th/p2Ae = (Sx10Y(1.00x10*}(0.019) = 263 m/s Gi) The steady flow energy equation, Eq. (1.3) applies, mAh + ¥ +2)=Q4W. Using Eq, (1.60) for an ideal gas, we write Ah = i p41 = 0,(T2—T}) for ¢p constant, Also Az ~0 and W. (gsi an eas eocaae ee Hence, aml r-7) + (VI-VD] = Gx1015200050-50) + 4 263-2749) = (5x10)[1560x105 + 34.2x10*] = 7970 W Since the exchanger is perfectly insulated, Q is also the heat transfer from the hot siream to the cold stream. Comments: 1. Take particular note that W = 0 in a heat exchanger, 2. Even though the helium velocities are large, the kinetic energy gain is small compared to the sensible heat gain. This is usually the case in practice.Exercise 1-3 Given: Coolant water flow through a condenser of an ocean thermal energy and fresh water plant, Given; @ Heat transferto coolant. i) Steam condensation rate. Assumptions: 1. ‘The condenser is perfectly insulated. @ The steady flow energy equation in the form of Eq. (1.4) applies since changes in eleva- tion and velocity are negligible, Q=thth-h) From Eq, (1.6b) for an incompressible liquid, PoP Iywhy = C4(Tg-T) + » for ¢y constant . =41920985-280) + COSRE)10” =20960-21 = 20,939 J/kg ) Q= (4000)(20,939) = 8.38x10" W (= 83.8 MW) Gi) From Table A.12a the saturation temperature corresponding to a pressure of 1482 Pa is, 286.0 K, and the enthalpy of vaporization is 2.471x10° J/kg. The steady flow energy equation applied to the condensing steam requires that Qe rhAh = tity tin = Qe = 8.38x10" /2.471x10° = 33.9 kg/s. Comments: 1. The contribution (P)-P,)/p to the enthalpy change of a liquid can be neglected when calculating enthalpy changes in heat exchangers. 2. Check that the SI units of AP/p are J/kg.MOC boy Exercise 1-4 Given: Glass vessel, 5 mm thick wall, insulated with 1 om thick neoprene rubber. Required: Heat loss, interface temperature, temperature profile. Assumptions: 1. 1-D conduction since the insulation is relatively thin. glass 4 “rubber Sloe Z t 1,720 oe 25% 20° Elie ett I From Table 1.1 the k values arc ka uber respectively. Using Eq. (1.10) the heat flow is Ti-Te 40-20 tx, ts _ 0005) _, _“Oon kA kgA (1,090.04) (0,190.04) 0. ‘To obtain the interface temperature T;, consider only the rubber layer, 20% -fooes nh —corn—| ‘Comments: 1. Temperature may be expressed either in degrees Celsius or kelvins for this, conduction calculation. 2. The temperature profile in each layer is linear.Exercise 1-5 Given: Three common building insulations. Required: Thermal resistance in terms of "R" value. Assumptions: 1. One dimensional conduction. @ —_Rvalue of 10 cm thick fiberglass? Table 1.1 gives k = 0.038 Wim K, and Eq. (J.9) gives the thermal resistance of unit area of insulation as L O1m mk. = To 72.53 ype asin Table B2, 2K _, mK 1Ws. in in_| 10553 hfe °F iw w <1 “3600s *|03048m]) * pe 58a Hence 2.63 m? K/W = (2.63)(5.68) = 14.9 h ft °F/Btu iie., a 10 em thick layer of fiberglass gives an R value of about 1 Gi) Thickness of cork to give an R = 18? Table 1.1 gives k = 0.043 W/m K for cork. gr 18 mK eK Ba 368 Wo Ww ne 3.17; L =0.136 m= 13.6 em Gi) R value of 2 cm thick white pine? Table 1.1 gives k = 0.1 W/m K for white pine perpen- Comments: 1. White pine is anisotropic, and values of thermal conductivity along and perpendicular to the grain are different. For a building siding, the value perpendicular to the grain is appropriate.Bxerelse 1-6 Given: Ice melting in an icebox exposed to an ambient temperature of 30°C. Required: Amount of ice that will melt in 8 hours. ‘Assumptions: 1. Steady one-dimensional conduction. 2. No air leakage through lid seal. Ty : T1=0°C g 1, =30°C ie .028 Wim K (Table 1.1) hg = 335 Wikg a 2900) = 0.40 m? | 0.02 m ke A= (0.40)(0.20)(4) + (0.207)(2) = 0.40 m* ‘Using Eq. (1.9), the rate of heat gain i j— eT 30K caw “Lika. ~ (0.02 m)/(0.028 W/m K)0.40 m*) In 8 hours the total heat gain is Q= (16.8 J/sy(8 h)(3600 s/h) = 4.84x10° J ‘The enthalpy of melting ive is 335 kifkg, hence the total amount of ice that melts is w= Qfhay = (4.84x10° J)(3.35x10° I/kg) = 1.44 kg Comments: 1. We have assumed that the only significant thermal resistance is the polystyrene, ‘The exterior and interior thermal resistances are negligible by comparison (see Section 1.4).Exercise 1-7 Given: Inner and outer temperature of a composite wall. Required: (i) Heat flow per unit area. Gi) Interface temperatures. Gii) Temperature profile through the wall. Assumptions: 1. Steady, one-dimensional conduction. Fiberglass Ky =ke =0.10 W/m K kp = 0.038 W/m K (Table 1.1) Asim? ooamet Bocce] Preeem . @ The wall can be modeled as a circuit with 3 resistances in series. ba Ke ‘Writing Eq, (1.10a) for three resistances with A = 1 m? gives Q= @M)=ATIER = (T-To)/LalkatLa lk tc/ke) = (20 K)/[(2)(0.02 m)/(0.10 W/m K) + (0.06 m)/(0.038 W/m K)] = 10.1 Wim? Gi) = MT) g/kq) = (T2-To)Lc/ke) ° Ty =Ti-aL a ka = (20){10.1)(0.02)/0.1 Tz =TotaLc/ke = (0)+(10.1)(0.02)/(0.1) = 2.0°C Fivergtass Pine ot Comments 1. Temaperature may be expressed either in °C or K for this conduction calculation. 2. ‘The temperature profile in each layer is linear.Exercise 1-8 Given: Toner and outer temperatures and rate of heat transfer for a freezer unit. Required: Thickness of polystyrene walls required to satisfy specifications. Assumptions: 1. Steady one-dimensional conduction. 2, Base of freezer perfectly insulated. T.=-10°C T,=30°C = 0.028 Wim K (Table 1.1) A= (4)(2m)(1 m) + (1 m)(1m)=9 m? be @ Using Eq. (1.9) the heat flow is Tea! L=kACT TQ L = (0.028)(9\(30--10))/(200) = 0.05 m ‘So, to maintain T; =—10°C we need L 2 5 cm. Comments: 1. Since the base of the freezer cannot be perfectly insulated, the required thickness of insulation is somewhat greater than 5 cm. 2. Ifthe reftigeration unit has a COP of 3.6, and the electric motor efficiency is 80%, the steady power consumption required to balance this heat gain is (200/(3.6)(0.8)) = 69 W. 3. More often a freezer is controlled by a thermostat that allows the temperature to cycle between two limits, say -10°C +2 K. Then the refrigeration unit is only on intermittently, and must have a refrigeration capacity much larger than 200 W. After studying Section 1.5, you will be able to size the refrigerator unit for intermittent operation.Exercise 1-9 Given: Inner and outer temperatures of a spherical tank containing saturated liquid nitrogen, Required: Thickness of superinsulation to meet boil-off specification, Assumptions: 1. Steady, one-dimensional conduction. 2. The tank wall and outside thermal resistances are negligible. T=714K, T,=250K k=9x10 W/m K yg = 0.200x10° Jkg. tn <2mg/s A= xD? = x(1)? =3.14 m? ‘The first law of thermodynamics applied to the tank requires that the heat gain and boil-off rate are related as Q=rtithyg = (2x10 keg/s)(0.200x10F J/kg) =0.4 W Using Eg. (1.9) we can find the required insulation thickness as sO aa) ADOSO-TA). «1 225107? m (122m) Comments: 1. Since the pressure is maintained constant at 1 atm, temperature T; remains at the No boiling point of 77.4 K during vaporization. 2. Check that the outside thermal resistance is negligible, Assuming that the tank is not in a pressurized compartment, ho=hro= 456 Ty Inyo = (4) (0.035)(5.67 x 10°)(250)’ = 0.124 W/m*K Yh, A= 1/(0.124)(3.14) = 2.6K/W L/kA =1.22x107/(9 x10"°)(.14) = 432K/W Since 2.6<<432, the outside resistance is negligible.Exercise 1-10 Given: A black body radiating to a surrounding black enclosure. Required: Radiative heat flux leaving the body at various temperatures. Fora black body radiating to a black enclosure, Eq, (1.14) is an =(T{-T4) =5.67x108(T}-T3) 4 =567% (= 5 (=) wim? 100} ~ [100] @ 1) =180K, T)=80K 1 = 5.67(1.80*-0,808) = 57.2 Wim? Gi) 7, =400K, T, =300K gn = 5.67(44-34) = 992 W/m? Gi) T, =1100K, T, = 100K qi = 5.67(114-10*) = 2.63x10* W/m? (iv) 100K, Tz = 5000K qi = 5.61(514-S0*) = 2.92x10° W/m? Comments: 1. The T* dependence of black body emissive power causes a rapid rise of qy with T.Exercise 1-11 Given: Astronaut in space suit at work in shuttle service bay. Required: Rate of heat loss through suit when outer surface is at 0°C. Assumptions: 1. Small gray body in large, nearly black surrounds. Eq. (1.18) applies, with Ty = 273K, Ty = 173K. The heat loss is then Qn =eA Tord (0.05)(3){(5.67x10-*)273)'{5.67x10 *)(173)4] .05)(3)[315-S11 =39.6W T.=-100°C ‘There are 4187 kcal, 3600 s/n, thus 7 J}}_1_ kcal 3 sow-[vs Va, 22100 : Comments: 1. Notice that absolute temperature must be used in radiation calculations. 2, The kilocalorie is the “Calorie” commonly used in food technology as the unit of energy. 3, Use UNITS to check the conversion from W to keal/h. 34.0 keal/hExercise 1-12 Given: Electronic device contained in a cylinder dissipating heat by radiation only. Required: Emittance of surface coating to satisfy Ty < 80°C. Assumptions: 1. Small gray body in large, nearly black surroundings. T, =-80°C Ty, $80°C Qn=60W Ay = a{(0.1)(0.3) + 2)(0.25)0.17] =0.11m @ Using Eq. (1.18) with T, = 193 K, T = 353 K and g; as unknown,we obtain Qi2/ALo(Tt-TH) = (60)/.11)5.67)3.53-1.934) 168 Hence we need ; 2 0.68 to maintain T; < 80°C. Comments: 1. From Table 1.3 we see this condition is easily satisfied with a suitable paint. A bright metallic surface will have too low a value of e.Exercise 1-13 Given: Sensor inside a vacuum chamber. Required: Graph of power vs. emittance of sensor to maintain sensor temperature, Assumptions: 1. Small gray body in large,nearly black surroundings, Ty =25°C = 298K 190°C = 83 K o Using Eq, (1.18) the power required is PQ = Ai0IT}-TH] 1 (1x10 )(5.67)(2.98"-0.834) =0.444e; W Pw) o.aaa 04 02 o 05 10 & Comments: 1. There is a linear relation between power and emittance.Exercise 1-14 Given: A 1 cm copper cube heat sink for a semiconductor laser at 80 K. Required: Parasitic heat gain from surrounding Dewar flask at 100 K. Assumptions: 1. Radiation exchange can be modeled as a small gray body in large, nearly ‘black surroundings. 2. The presence of the laser and spreader can be ignored. Provided the Dewar flask is evacuated to a sufficiently low pressure, radiation will be the only significant mode of heat transfer from the flask to the sink (even though the temperatures are very low). If the radiation exchange is modeled as a small gray body in large, nearly black surroundings, Eq. (1.18) applies, Qy = eAOT!-0T;) A Qyp = (0.08)(6x10 4)(5.67x10 “*)(80*-100%) = -1.61x10* W = -0.161 mW Comments: 1. Since the laser dissipates heat at an overall rate of 1.5 W, the parasitic heat gain is (1.61x104/1.5)(100) = 0.0107% and is surely negligible. 2. The parasitic heat losses could be further reduced by polishing the copper~Table [A.Sb shows how € decreases to a very low value at low temperatures.Exercise 1-15 Given: Flow of three fiuids in a tube, L= 3m, D= 1.26cm, Required: Convective heat transfer coefficient h, for (2) water at 2 m/s, Gi) oil at 2 m/s, and Gi) air at 20 m/s. Assumptions: 1. A Jong tube so that entrance effects are negligible and Eqs. (1.21), (1.22) apply. 2. Fluid properties can be evaluated at 300 K. @ Water flowing at 2 m/s VD _ (0.0126) Rep => = COED = 28,970, 1 (1.22), Dy = Ga.gr10%y 72097 tuvblent hence se Eq, (1.22) VOR OE _ (0.023)(2)0.611)(096x4178)* _ gas W/m? K he = 0.023 5 oa (0.0126)(0.87x10*)* Gi) SAE SOoil flowing at 2 m/s YD _ 20.0126) : Rep = 22 =4i 7 21), == eygq 8) 742 laminar, hence use Eq (1.21, =3.66 K — 3-66)(0.145) _ 2 } ne =3.66 5 = Gorge) 421 Wim? K Git) Airat 1 atm flowing at 20 m/s Rep = me = 16,050, turbulent, hence use Eq. (1.22), (20)°*(0.0267)°51.177x1005)"* 2 = 0023 eer W) K be (0.0126)°2(15.7x10-*)0* ane Comments: 1, Notice that the h, for water is 86 times larger than that for air, even though the air is flowing at a ten times higher velocity. 2, Examine the role played by each propery, namely k, p, Cy, V.Exercise 1-16 Given: Flow in a 1 cm 1.D. long pipe. Required: Velocity dependence of the convective heat transfer coefficient for () water at 300 K, Gi) air at 300 K and 1 aum. Assumptions: 1. Laminar flow for Rep < 2300; fully turbulent flow for Rep > 10,000. @ — Flowof water Rep = VD/y, hence V = Repv/D Rep = 2300: V = (2300)(0.87x10)/(0.01) = 0.20 m/s, Rep = 10,000: V = (10,000)(0.87x10°)/(0.01) = 0.87 m/s For V < 0.2 m/s, the flow is laminar and Eq. (1.21) applies: hy =3.66 E = 3.65 9612 = 294 win? For V > 0.87 m/s, the flow is turbulent and Eq. (1.22) applies: VFO (p,)4 119° (996x4178)°4 0 ee pemiesye SOM0 AP Win?K Gi) Flow of air. Rep = 2300: V = (2300)(15.7x10)4(0.01) = 3.61 m/s Rep = 10,000: V = (10,000)(15.7x10-6/(0.01) = 15.7 mis poe 0.0267) Laminar flow: he =3.66 Top v0" 0.0267)°(1.177%1005)"* ‘Turbulent flow: h, = 0.023 Y-(0-0267)°.177x1005)"" _ 9 39 yor a (0.005.781) h, wem hn, wm K 1 Water at 300 K 104 Air at 300K, 1 atm 77 W/m? K WF 1 TT mvs, WF 11010 mvs Comments: 1. For the transition regime, 2300 = Rep = 10,000, he cannot be reliably predicted. In ‘engineering practice this regime is avoided, if at all possible.Exercise 1-17 Given: A vertical wall 1 m high at 310 Kn still air at 290 K, 1 atm. Required: (i) Local heat transfer coefficient variation. Gi Convective heat loss per meter width of wall. Assumptions: — 1. Eqs. (1.23a,b) are sufficiently accurate. 2. Transition from a laminar to a turbulent boundary layer occurs at Gr, = 10°. @ ‘We first calculate the location of transition from a laminar to a turbulent boundary layer. Gr, = 10? = BAN EEA?; B= Ts T= FAT = [ao]? _ {ao 05.7810? |” _ 9799 * [arms 0p00.81) | 0 (T-#293) X= so? _]"?__ 0315 @mte] ~ arn * es Der eer ee A he RHS | 400 of °c K K K m Wm?K £Ea.(1)| 1200 333 313 400.625 4.08 841 343-318 50 0584 4.421119. 353 323 60 05524711419. 1000 sas From the graph, the solution is T, = 66°C. Comments. 1. AtT, = 66°C, Qeat = 591 W, which is 59% of the total heat transfer.Exercise 1-26 Given: Saturated steam at 150°C flowing through a 15 cm O.D. uninsulated steam pipe. Required: Reduction in steam condensed if the pipe is painted with aluminum paint (e = 0.14), and annual savings. Assumptions: 1. Thermal resistance of condensate film and pipe wall is negligible. 2. The pipe can be modeled as a small gray object in large, nearly black surroundings. ‘The rate of heat loss from the pipe is Q = aDLh, +h) y-T) where we have used the linearized radiation heat transfer coefficient defined by Eq. (1.19) for ‘convenience, h, = 4eoT,, = (4)e(5.67x10 *)[(293+423)/2]° = 10.4e Wim 7K For the original pipe with e = 0.8, Q = ()(0.15)(20)14 +(10.4)(0.8)(423-293) = 15,090W With the aluminum paint of € = 0.14, Q = (H)(O.15)(20)[4 +(10.4)(0.14)]423-293) = 6685 W ‘The reduction in the amount of steam condensed is Arn = AQ/h,, where hy is the enthalpy of vaporization at 150°C; from Table A.12a, hy, = 2.11x10® J/kg. Hence Ath = (15,090-6685)/2.11x10° = 3.98x10" kg/s = 344 ke/day To obtain the annual savings we multiply AQ by the cost of thermal energy. AQ = (15,090-6685) = 8405W = 8.405kW Annual savings = (8.405 KW)(24x365 hvyr)(0.04 $/Wh) = $2945 Comments: 1. A more accurate estimate for the radiation heat loss is obtained from Quaq Aoe(T3.-T;). Check to sce if our approximation was justified.Exercise 1-27 Given: A 2 cm square cross-section bar consisting of 1 cm thick layers of copper and epoxy. Required: Thermal resistances for heat flow parallel and perpendicular to the layers. Assumptions: 1, One-dimensional steady conduction (end effects negligible). 2. Constant properties. 3, See cases (i) and (i) below. For heat flow perpendicular to the layers we have simple series resistances: ty, be Ag R=R,AR, WV RY 0.01, __0.01___ 9.9125+12.50 = 1251 KAW ! © GOK) O.A)(2x10) For heat flow parallel to the layer we will try two models, ! (i) Assume surfaces of constant x are isothermal. Then . the thermal circuit is as shown, and ae w WN } FRA Kg thpAy z ch a 0.02 (400)(1x10 *)+(0.4)(1x10 *) R ii) Assume that there is no transverse heat flow between the copper and epoxy; then the thermal circuit is as shown, and for parallel resistances R = 1/20.02 = 4,995x10* K/W Comments: 1. Of course, the thermal resistance perpendicular to the layers is much higher than that parallel to the layers~due to the insulating effect of the epoxy layer. 2. The two models for heat flow parallel to the layers gives the same results— because we constrained the slab sides to be isothermal, which ensures 1-D heat flow (see also Exercise 1 - 34). | | 4.99510 * KAWExercise 1-28 Given: "Suprathane” urethane foam sandwich used with a 3 1/2 in thick mineral wool to insulate a wall Required: (i) Thermal resistance per unit area of Suprathane sandwich in SI units. (ii) Rate of heat loss for T; = 20°C, T, = -20°C. Assumptions: 1. Steady one-dimensional conduction. (@) The R-value of the sandwich is given as 13.2 [Btu/hr ft? °F]. Table B.2 gives the conversion factor for heat transfer coefficient as 5.678 (Who? KV[Buw/f? br °F] ‘Thus the R-value of the sandwich in ST units is R = 13.2 [Bewhr f¢ °F}Y5.678 (Wim? KY [Btu/ft? hr “F] = 2.32 [Wm? Ky" Gi) From Eq. (1.29) the heat flow through the wall is : -Qew— ww 1 Q = VAM, -T) KOA , DR-values | 1 The R-value for the combined urethane sandwich and mineral wool blanket is R = 22.7/5.678 = 4.0 [Wim? Ky! (5)(3) = 15 m®. ‘Thus Aiea 4 1 UA” Tay” UH” OAH ‘The area for heat flow is A = (9.5+26743.3)10% UA =3.57 Q = G.57)(20-(20)) = 143 W ‘Comments: 1. Notice that the thermal resistance of the wall is so large that precise values of the inside and outside heat transfer coefficients are not needed to give an accurate value for the heat flow.Exercise 1-29 Given: A 30cm x 20cm x 15 em deep ice chest with 3 cm thick polystyrene walls and lid, filled with a mixture of ice and water. Required: Time required for ice to melt if initially 70% by mass of the mixture. Assumptions: 1. Steady, one-dimensional conduction through the walls and lid, 2. Heat loss through base negligible. 3. The mixture is initially at O°C. 4. . Negligible convective resistance on the water side. ‘The first law of thermodynamics applied to the ice-water mixture requires that wh, = Q a where W is the melting rate of the ice, h,, is the enthalpy of melting, and Q is the heat transfer into the chest. To calculate Q we require the overall heat transfer coefficient. From Eqs. (1.30b) and (1.32) Since h,, is relatively large (water), the inside resistance can be neglected. Then, using Table 1.1, 1 _ 003 1 a 0.85 Wim? K U 0.028 10 = 107401 = 117; U ‘The total area of the sides and top is A = (2){0.3)(0.15) + (2)(0.2)(0.15) + (0.3x0.2) = 0.21 m* Q = UACT,-T,) = (0.85)(0.21)(30-0) = 5.36 W From Eq. (1), Qh, = 5.36/335x10° = 1.60x10°° kg/s The initial mass of ice is (0.70)(0.24x0.14x0.09)(1000) = 2.12 kg. ‘The time taken for the ice to melt is thus w 1 = whw = (2.12)(1.60x10*) = 133x105 s (~ 37 h) Comments: 1. The walls of this ice chest are relatively thick resulting in a long time for the ice ‘to melt.Exercise 1-30 Given: A kitchen oven operating at 280°C in a kitchen at 30°C. Required: Thickness of fiberglass insulation required to ensure that the outside surfaces do not exceed 40°C when the ambient air is at 25°C. Assumptions: 1. Steady, one-dimensional conduction. 2 3, Surroundings also at 25°C. ‘Thermal resistance of metal walls negligible. Referring to the thermal circuit, we must first calculate the heat flow from the outside thermal resistance, QA = (hg *B Toy Ty) = (15)(40-25) = 225 Wim? Next we use the circuit from T; to Ty, to write Solving, L = 0.073 m (7.3 em). Comments: 1. The outside heat transfer coefficient accounts for both convection and radiation, h, = hey + Big Now h,, = 5 Wim? K, and h,, = 4 W/m? K for natural convection, Hence, a lower limit for h, is about 9 Wim? K. A value of 15 Wim? K suggests a fan is circulating air in the kitchen.Exercise 1-31 Given: A furnace wall to operate with inner and outer surfaces at 1500 K and 320 K. Required: Arrangement of two types of insulating bricks to restrict the heat flow to 1000 W/m? at least cost. Assumptions: 1. One-dimensional, steady conduction. 2. The bricks have the same cost Since ky = 2k We should use B type brick to the greatest extent possible. Since its temperature limit is 1000 K we set T, = 1000 K, then 1900 = 1500-1000 _ 1000-320 Ly20 Tyo Solving, Ly = 1.0m; Ly = 0.68 m. Comments: 1. Since the bricks are 20 cm x 10 cm x 8 cm it is possible to arrange them to give L, = 1.0 m and Ly = 0.68 m exactly (assuring no mortar is used).Exercise 1-32 Given: A composite furnace wall consisting of 0.3 m thick fire clay brick, 0.2 m thick kaolin brick, and 0.1 m thick outer layer of face brick. Required: (i) Heat loss through a 4 m x 8 m wall for a furnace temperature of 1500 K and ambient temperature of 310 K. ii) Additional thickness of kaolin brick if face brick is not to exceed 360 K. Assumptions: 1. Steady, one-dimensional conduction 2. No air gaps between the layers of brick. 1d ede a Flr U =0.50Wim? K. Q- UAT -T) (ii) Referring to the thermal circuit between T, and T,, (©.50)(8x4)(1400-310) =17.47x104 W (175 kW) Q_ Tt _ 360-310 ea wT Ke 3°75 Q = (348)2) = 1.11x10* W (11.1 kW) We see that Lg must be increased to give about a seven-fold reduction in heat flow, which does not seem too practical. The value of U required is (11.107 5)0.50) = 0.318 = 348 Wim? 1 1,03 bs 01 1 ; solving, Ly =0 34m. 4. A substitute brick for the outer layer that can withstand higher temperatures should be chosen.Exercise 1-32 Given: A composite fumace wall consisting of 03 m thick fire clay brick, 0.2 m thik Kaolin bbrick, and 0.1 m thick outer layer of face brick. Required: (i) Heat loss through a 4 m x 8 m wall for a furnace temperature of 1500 K and ambient temperature of 310 K. ii) Additional thickness of kaolin brick if face brick is not to exceed 360 K. Assumptions: 1. Steady, one-dimensional conduction. 2. No air gaps between the layers of brick. @ From the thermal circuit, 1 La be be 1 (03,02 01,1 a OK & Kb 100 17 12 13 15 U=2.01 Wim’ K. Q = UACT-T,) = (2.01)(8x4(1500-310) = 7.67x10* W (76.7 kW) (i) Referring to the thermal circuit between T, and Ty, Q_ Tat _ 360-310 PA eee ieee dOHEcaL Seeiubte t eetedsate Q = (348)(32) = Liixl04 W (11.1 kW) We see that Lg must be increased td give about a seven-fold reduction in heat flow, which does not seem too practical. ‘The value of U required is (11.1/76.7)(2.01) = 0.291 = 348 W/m? L, mee = at 03 7801 41 sotving, 1 Oar © T0017 12135 3.73 m, Comments: 1. This:thickness of kaolin:brick-is impractical. 2. A substitute brick for the jouter’ layer that. can withstand higher temperatures should be chosen,Exercise 1.33 Given: Ambient air at -10°C with a wind speed of 10 m/s, giving h, = 50 W/m? K. Required: Wind-chill temperature and skin outer temperature. -gaidtol2 Astmptions: 1, Clothing equivalent to 8 mm thick wool of k = 0.05 W/m K. 2. A3mm layer of skin of k = 0.35 W/m K, and 35°C below the skin. 3. Under calm conditions, h, = 5 Wim? K. 4. A radiation heat transfer coefficient of 5 Wim? K. ‘We must first calculate the actual heat loss from the face and then find the air temperature that gives the same heat loss on a calm day. Referring to the thermal circuit, 35-(-10) 0.003 , 0.008 | 1 035 O05 3505 = 241 Wim? where the radiation coefficient has been taken as h, = S¢ ~ 5 W/m? K. Ona calm day, 35-Tye 0.003, 0.008 1” 035 OOS 35 a, = 241 = solving, Tye = -29:7°C. Also, the actual temperature of the skin surface is obtained from the thermal circuit as TT, 15 hence 241 = QA Lyky Comments: 1. Notice that the radiation and convection heat flow paths are in Parallel |--the heat transfer Coefficients'éan be added. ae ee ‘The clothing isa Yery good inslator~ihe Skin surface temperate is only 2°C ‘ew the interior emperatie. “a! You may Wonder Why the windchill temperature does not characterize heat loss from bare skin since frostbite of exposed skin is major concer in extreme cold weather. If the above calculation is repeated for bare skin a wind chill temperature of -147°C is obtained, Better put on gloves and a ski mask!Exercise 1-34 Given: A 3 em wide, 1 m high composite wall as shown. Required: Heat flow through the wall, Assumptions: 1. Steady state 2. Constant properties. 3. One-dimensional conduction. 4, See cases (i) and (ii) below. bees} tore VV TE heel (i) Symmetry allows us to divide the wall into 10 slices, each 30 em wide, as shown, We will first assume that surfaces of constant x are isothermal: then the thermal circuit is as shown. L, 02 Ra eA Hare Oe ees A EA, TOGHODM fae Ls Ol Le ol ee eee = KA, NCHA KeAg O05 TOO RytRe = 0.5+1.0 = 1.5 K/W for series resistances a ts for parallel resistances 1 1 1 +Ryg tl = +0.176 +. = 0.280 KW nA Rat iA GB) O78 rags * 0280 Q = (TT, VER = (295-260)/0.280 = 125 (ii) For an alternative mode! we can assume that there is no transverse heat flow, and a new thermal circuit results. = 0.5 KW; Ro =. = _____ = 10 kw1-34 (continued) 1 Ry Re 1 TyAe Peete Lay, 4 Ry 4 baa heel 1 1 1 1 ER, = +Ryt = t+ 02+ = 0.5125 KW a TA, * BA, OO yy 1 Hy a 1 = +Ry*Ro* = 2 + 0541.0 + = 1.656 K/W me Tse BaAne OO Toe “4 gr = {1 _+_1_\" = 0301 Kw Oss 1.656 Q = C,-T, VER = (295-260)/0.391 = 89.5 W Comments: 1. The two estimates of Q differ by 33%, which is disturbingly large. 2. We would expect the true answer to be between these two values, say the average value of 107 Wim would be a best estimate 3. The actual problem involves two-dimensional conduction. The correct solution bbe obtained by numerical methods (see Section 3.5).Exercise 1-35 Given: ‘Thermometer suspended inside an enclosure. Required: True air temperature in the enclosure. Assumptions: 1. Small gray body in large, nearly black surroundings. “fe = 12 Wim? K €=08 (Table 1.3) Ty =273K rain oF T= 288K ™ (@ Using Eq. (1.35) gives us Qeonv + Qead =O since at steady state there is no conduction within the thermometer. Substituting from Eqs. (1.24) and (1.18), HLACT—Te) + 26A(T4—T4) 12(288-T,) + (0.8X(5.67)(2.88*-2.734) = Solving for T, yields, Te =293 K=20°C ‘Comments: 1. Ifthe thermometer reading was taken to be the air temperature, the error would be 5°C.Exercise 1-36 Given: A sleeping bag inside a mountaineer’s tent. Required: Rate of heat loss from bag surface if the emittance of the tent outer surface is (0.7, and Gi) 0.2. Assumptions: 1. The tent is much larger than the bag. 2. Ambient air circulates through the tent. 3. he = 8 W/m? K on both outside and inside walls. G) We first make an energy balance on the tent wall assuming that the heat transfer from the interior to the tent is negligible. From Eq, (1.34) with dams = Grad + Geoav = 0 oe(T$ Ty) + Ih TynTe) =O (5.67x10* (0.742134) + 16(T,,-272) = 0 Solve by iteration, Ty, = 265 K. Knowing the temperature of the tent wall is 265 K, we can now make an energy balance on the sleeping bag. Assuming the air in the tent is also at approximately 272 K, tors = Goad + Gocaw = GE(T3—Ty) + he(Ts-Te) = (5.67x10*)(0,8(2834-2654) + 4(283-272) = 67 + 44 = 111 Wim? Gi) Repeating the previous calculations for a tent emittance of 0.2 gives Ty = 269.7 K, and Qos: = 95 Wim?. Comments: 1. ‘The effect of aluminizing the outer surface of the tent is to reduce the heat loss from the sleeping bag by (111-95)/111 = 14%. 2. The effective temperature of the sky as a radiation sink depends on cloud cover and the humidity of the air close to the ground (see Section 6.4.2). 3. Ifa powerful wind gust blows the tent away, dag from the sleeping bag increases to 198 Wim”.Exercise 1-37 Given: Convective heat transfer coefficient meter comprised of two sensors. Required: Determine h,. Assumptions: 1. The sensors can be modeled as small gray bodies in large, nearly black surroundings at a uniform temperature T,, Y 1, = 300K ™ = Ty = 320 1 = 21.7 hw Ty = 320K; 2 =09 is 5-01 Using the surface energy balance Eq. (1.34) with daogg = QUA, where Q is the electrical power input, QVA = eovt Gea = BET) +68)(T4-T) QUA = Gsoe2 +I = Bega) + 98,(T3-Ty) Assuming b,, = hy = hy. 21.7x10 10 = h,320-300) + (5.67(0.9)[(3.2)* -(T, x10 74) 8.28x10 “/10 = (320-300) + (5.67)(0.1)[G.2)$ -(T,x10 )'] Let y =[(3.2)* - (T,-107)4], then 217 = 20h,+5-1y 83 = 20b,40.567y Solving, y = 29.56, h, = 3.31 Wn? K, T,, = 294.6 K Comments: 1. It is essential that the sensor temperatures equal the test surface temperature in order to measure the true convective heat transfer coefficient on the test surface. After studying Section 5.4.3, the reason will be apparent. 2. Since the surrounding surfaces are unlikely to be at a uniform temperature, T,, is an effective average temperature.Exercise 1-38 Given: A horizontal roof of emittance 0.96 losing heat to its surroundings. Required: Its temperature when (i) he = 5 W/m? K, (i) h, = 20 W/m? K, (iii) new roof of emit- Assumptions: 1. The roof "sees" the sky only, which is taken to be black. tance 0.15. Let T, be the roof temperature, T, the air temperature, and T.jy the temperature of the sky as a radiation heat sink. A surface energy balance on the roof gives, Gea + Gad =0 nT ,T,) + €0(T $14) = 0 o i) Gi) Comments: 1. The effective temperature of the sky as a radiation sink depends on cloud cover and on the humidity of the air close to the ground. Methods for calculating Tuy €=0,96, h, = 5 Wim? K 5(T,-278) + 0.96x5.67x10-*(r$—213*) = 0 T, + 1.089x10*1! = 300 Solve by trial, T, =255 K. = 0.96, h, = 20 Wim? K 20(T,-278) + 0.96x5.67x10*(T-2134 Ty +2.72x10°T$ = 284 T,=270K 1.15, he = 5 Wan? K 5(T,-278) + 0.15x5.67x10°8 (14-2134) = 0. Te. + 1.70x10°T$ = 282 T,=272K €=0.15, he = 20 Win? K 20(T,-278) + 0.15x5.67x10-°(T$-213%) = 0 T, + 4.25x10-T$ = 279 1,=276K are given in Section 6.4.2. 2. Since Tyxy (To*Te) Comments: 1. The formula for the temperature response does not give T explicitly.Exercise 1-53 Given: ‘Thermocouple immersed in an air stream with a sinusoidal temperature variation. Reguired: (i) Set up goveming differential equation, i) Solution of differential equation. ii) Amplitude and phase lag of thermocouple. Assumptions: 1. Lumped thermal capacity method valid. 2. Neglect thermal radiation, G@) Using Eq, (1.36) with T, = C+B sin ot and B =h,A/pVe Af +BT=BC + BB sino Gi) Since we are not interested in the transient response, we can ignore the homogeneous solu- tion T(t), and obtain only the particular solution. Let T(t) = D+ E sin(wt+4) and substitute in the differential equation, Bo cos(at+4) + BD + BE sin(at+4) = BC + BB sin ot o Differentiating, Easin(ot+d) + BOE cos(at+>) = BBcos ot @ Setting t= 0 in Eqs, (1) and (2), equating corresponding terms and using trigonometric identities yields cos) + Psing=| BD=BC, D=C Bos $a sing = BB/E, E= BB/(a?+B2)* Te) = C+ BB/@?+B°/)sin(we+9) , d= tan ayB) Gi) T, = 320+10sint 1. ih, =30 Wim? K, B = (30)(6/(11340)(129)(0.002) = 0.0615 s* 61 = tan !(-1/0.0615) =—1.509 rd By = (0.0615)(10/(17+0.0615°)* = 0.614 K 2, he = 100 Win? K, B= 0.205 st 2 = tan (-1/0.205) =-1.37 rd Ez = (0.205)(10(1?+0.205")* = 2.0K Comments: 1. Since the angular frequency is 1 rad/s,he lags are 1.51 s and 1.37 s respectively. 2. Ifyou were able to do part (i), the time you spent in mathematics classes has paid offExercise 1-54 Given: A body with internal heat generation Q, losing heat by convection. Reqitired: Differential equation for T(t), and solution, Assumptions: 1. Biot number <0.1. Fora Biot number <0.1 we can use the lumped thermal capacity model. Applying the first law in the form of Eq. (1.2) p¥e = O10, =H ACT) + Oy o where T, is the surrounding temperature. Equation (1) can be rearranged as which defines a pseudo ambient temperature T; = T, + Qy/iieA, at BA ye ‘dt pve (T-T,) @Q) ‘The solution is simply Eq. (1.38) with T, replaced by T;,Exercise 1-55 Given: Electronic components attached to a finned base plate exposed to an airflow. Required: Temperature 10 minutes after power is switched on. Assumptions: 1. Biot number <0.1. 2, Negligible thermal resistance between components and plate, For Bi < 0.1 we can use the lumped thermal capacitance model. Applying the first law of ther- modynamics in the form of Eq. (1.2) pcv g = 00, = AMT.) +O, where Qy is the power dissipated by the electronic components. The solution of the differential equation was obtained in Exercise 1-54 as THT+Q ied) ToTe#Qy/heA) ‘From the given data, pcV = 5000 J/K, be. =10min= 600s, T =e 101500050 (300/ 10)+(295+300/ 10) = 316 K = 10 W/K, Ty =T, = 295 K, Qy = 300 W. Hence fort Comments: 1. As1— % the steady temperature of the components is T = 325 K (which could be ‘obtained from a simple balance of heat gencration and heat loss). TK 325 1 nee een crt 295 @_____1_____. ° 10 minExercise 1-56 Given: Reactor vessel with contents reacting exothermally. Required: Temperature after () 1 minute, (ti) 10 minutes. Assumptions: 1. The contents are agitated, so that the temperature is uniform and the lumped thermal capacity model is valid. Using the result of Exercise 1-54 with i, replaced by U. Qe (4x105 W/m®)(0.008 m*) ao = 267K UA (Win? K024m?) 2 te = LAX. = £3000 Hee KY1200 kephn” (0.008 m°) _ 94 099 = 400 minutes UA G Wim? K)(0.24 m?) Q T-|Te+ GK 2) T- [T+ Te @ After 1 minute, ©) 780042667) _ T= 790-(30042667) ~*~ 09975 T=296.7K Gi) After 10 minutes, 73004266) _ 6 "i ih ae 290-(300+2667) T=356.1K Comments: 1. The appropriate definition of Biot number for this problem is, ao Intemal convection resistance . ‘Wall resistance + external convection resistance Since liquid phase convective heat transfer coefficients are generally much higher that those in air, the assumption of small Biot number is reasonable.Exercise 1-57 Given: Butane is drawn offa tank at th = 0.05 kg/h. Required: (@) Steady temperature, (ii) time taken for 80% of the temperature drop to occur. Assumptions: 1, Use the lumped thermal capacity model as the butane probably boils quite vigorously. Using the result of Exercise 1-54 with Tp = Te and Q, =—tihg, __ 474344242300 “5x02 = 5924s=1.65h Thy _ (O05/3600)3.86K10) _ 4 57 . fe ta nA. (5(0.22) i =m ‘Thus,55-T = 4.87 |1-« 165 ast», 55-T= 4.87, T= 501°C 80% of 4.87 K is 3.90 K, soo=as7 ie] , solving t= 2.66 h. Comments: 1. Since the vapor pressure of the butane decreases as the temperature falls, the performance of the bumer can deteriorate. A higher vapor pressure gaseous fuel, such as propane, is preferred for outdoor gas-fired barbecues in cold climates. 2. LPG contains a mixture of butane and propane. When used as an automobile fuel, liquid LPG is drawn off from the bottom of the fuel tank to maintain a constant composition of fuel supplied to the motor. If gas were drawn off from the top of the tank, the ‘more volatile propane would tend to be used first. 3, Notice that we have assumed a negligible change in the mass of fuel in the tank: since 0.05x2.66 << 4.0, this assumption is valid.Exercise 1-58 Given: Stirred milk tank with milk initially at 4°C. Required: Temperature of milk after 24h. Assumptions: 1. Lumped thermal capacity model valid. ‘The result of Exercise 1-54 applies with h, replaced by U, and Qy = (0.75)(400) = 300 W. The inside convective resistance and the resistance of the stainless steel wall can be neglected, then 1 U * 0026 1S 4 5 =2.881020; U=0.324 Wim? K V=(nD*/4\L) = (H/4)(2.5P°G.5) = 17.18 m? A = ADL + 2(nD?/4) = (x)(2.5)(3.5) + 2R/4V2.5P 11.5+9.8 = 37.3 m> & 30 VA ~ (0324373) = SV _ (3894)(1034)(17.18) © UA (0324)37.3) = 1590 h (66 days) =248K -T2X108 § Comments: 1. Uf the milk were not stirred the temperature rise is calculated by setting Q, /UA = 0, a = 0.9850; T = 4,39°C, ic., the temperature rise is only 0.39°C, However, the temperature of the milk adjacent to the tank wall would be higher, and the milk might spoil. 2. There could be appreciable heat leaks through connecting pipes, and through the supporting structure.Exercise 1-59 Given: Laser dissipating 1.5 W heat on 1 cm cube magnesium heat sink. Required: Time required for sink to reach 100 K, if cold finger removes heat at the rate Q. = 0,02(7-50) W. Assumptions: 1. Constant properties. 2. Lumped thermal capacity model adequate. 3. Negligible parasitic heat gain. Using a fumped thermal capacity model, Eq. (1.2) applies, ar pcv acer ~0.02(T-50) wo a = -0.02(T-50-Q,,,./0.02) Q) Comparing Eq. (2) with Eg. (1.36), b,A = 0.02, T, = 50+Qie,/0.02. ‘Thus from Eq, (1.38) '-(50+Q 02) TH +Qiel.02) ote gn pov/0.02 8) T,-(50+Qi,,./0.02) § = (1750}(450)(1x107)°/0.02 = 39.4 s TAS _ 9s T)-125 For Ty = 50 K, T= 100 K, 100-125 _ ee A, t= 83'S 30-125 Comments: 1, The quantity (50+Q,,,.,/0.02) has units of kelvins and can be viewed as an effective environmental temperature T,’. As t—> ©, TT,’ 2. Solve Ea. (1) by the techniques you learned in your mathematics courses. How would you check to see if the lumped thermal capacity model is adequate? (itis not straightforward!)Exercise 1-60 Given: ‘Test cylinder subjected to a change in flow conditions. Required: ‘Time required for steady conditions to be attained. Assumptions: 1. Lamped thermal capacity model adequate. 2. Convective heat transfer coefficient is approximately 100 W/m? K. If the power input to the electric heater is unchanged, an ‘increase in flow velocity will cause the cylinder temperature to decrease. To obtain an estimate of how long the experimenter will have to wait before taking data at the new test condition we will consider a lumped ‘thermal capacity model for the cylinder. The Biot number is bL ak Bi ‘Take L = V/A = aDb/mD = b where b is the wall thickness. From Table 1.1, k, = 386 Wim K. Hence, Bi = (100)(2.75x10°)/386 = 7.1x10% << 01 ‘The LTC model is certainly valid. From Eq. (1.38), THT, : - fae, pVchA For V/A = b, the time constant t, is {, = (8950)(385)(2.75x107Y/100 = 94.8 s ‘The temperature change will be 98% complete when 0.02 =e“; t= 3.91t, = (3.910948) = 3710 (~6 min) Comments: 1. ‘The experimenter will have to wait about six minutes before taking data. 2. A calculation of this type should be made when the experiment is designed (not after it is built!). If the cylinder wall is too thick the experiments will take too ong. On the other hand, if the wall is too thin, the cylinder surface will not be isothermal as desired. 3, Table A.la contains data for density and specific heat of metals. Notice that high density metals tend to have a low specific heat~the pe product varies less than p orc individually.Exercise 1-61 Given: 1 cm sphere to be convectively cooled from 600 K to 400 K. Required: Titwe required and minimum cooling rate. Assumptions: 1) Radiation heat loss negligible. 2) Lumped thermal capacitance model valid. Bi < 0.1. For a sphere V/A = 1 nD'D? = D/6; thus the time constant defined by Eq. (1.38) is pV, ep _ (140)(14,000)(0.01) _ 194 , bhA 6h, 6180) Equation (1.38) for the temperature response is 300 K, Ty = 600 K and {, = 18.1 s 400-300 _ 1 600-300 3 Solving, In 3 = V18.1; t= 19.9. ‘The cooling rate is given by Eq. (1.36), ar _ ha “dt cpVv _ which will be a minimum when (T-T,) is a minimum, ie., when T= 400 K, aT _ 400-300 S. = SS = 55 Kis at 18. Comments: 1. Check the Biot number, Bi = B,(V/A)/k, = (180\(0.01/6)/240 = 0.0013 < 0.1 i.e,, the lumped thermal capacitance assumption is excellent.Exercise 1-62 Given: "Tungsten fluoride is drawn off a 80 cm diameter spherical tank at 50 sccm. Required: (i) Steady temperature of the liquid WE, in the tank. ii) Time taken for Tiguid to approach with 0.5 K of its steady value. Assumptions: 1. The lumped thermal capacity model is valid since the WF boils inside the tank, 2. Negligible change of mass of WF, in tank. 3. Negligible inside and tank wall resistances. Use the result of Bxercise 1-54 with T, = T, and Q,, = EZ massxspecific heat _ (30)(434) +(4/3)(m)(0.4)°(3440)(1000) bA @amoay? = 5.82x104 5 ‘The molecular weight of WF, is (183.85)+(6(19) = 297.85. At STP the density of WE, is obtained from the ideal gas law as = PMIRT = (1.013x10°\297.85)8314)(293.15) = 124 kyl? ® Hence the mass flow rate of WF is th = (50)(10~(12.4/60 = 1.034x10> ke/s ths, _ (1.034x10 9)(25.7x10 297.85) 5.54 K nA (4nyo.4y? ‘Thus, 23-T = 5.44f1-exp(v5.82x10°)] as se @ Ast, 23-T=544; T= 17.6C ie w Gi) 23-18.1 = 5.44(1-exp(v5.82x10)}; t = 1.28x108 s (35.4 h). Comments: 1, The vapor pressure of the WF, decreases as its temperature falls (and equals 1 atm at 17°C, its normal boiling point). Thus at 17.6°C the pressure of the gas leaving the tank may not be sufficient to overcome the resistance of the delivery plumbing system. 2, higher rates of gas delivery are required the tank must be supplied with a heater to prevent the liquid temperature from falling too close to its boiling point.Exercise 1-63 Given: An 83 mm high, 45 < O.D. < 63 mm, Styrofoam cup with 1.5 mm thick walls, filled with 180 ml of coffee at 80°C. Required: (i) Initial rate of heat Joss through side walls if ambient temperature is 24°C, and outer surface temperature. Gi) Time required to cool to 60°C. Assumptions: 1. Internal convective resistance negligible. 2. Lumped thermal capacity model can be used for part (i). ‘The outside surface areas are Ap = (al4)(0.0739 = 0.00819 a? Apase = (9/4)(0.045)? = 0.00159 m? ¥ x Avides * (0)(1/2)(0.073+0.045)(0.083) = 0.0154 m? where the effect of the small side angle has been ignored. (i) Referring to the thermal circuit, the overall heat transfer coefficient is obtained from Eq. (1.30b) as where curvature effects have been ignored, and h,, has been assumed large (internal convection resistance negligible for water) 1, 0.0015, 1 0.035 10 29x10 + 10x10; U = 7.00 Wim? K q = U(T;T,) = (7.00)(80-24) = 392 W/m? Q = qA = (392)(1.54x10%) = 6.04 W Also q = h,(Toy-T,); Tow = T+ 24438 = 632°C 10 (i) We will assume that the average U for the top plus bottom of the cup is the same as for the sides, In order to check the validity of the lumped thermal capacity model, we assume that there is no circulation of the coffee, then Bi = UL/kyejgg Where v 180x107 = = a __ = 8.49x10 Fm ‘A 0.0154+0.0042 40,0016 Using the Table 1.1 for water, keyg, ~ 0.6 W/m K1-63 (continued) Bi = (7.0)(8.49x10°/0.6 = 0.099 < 0.1 Circulation due to stirring or natural convection will increase the effective value of k, and thus the lumped thermal capacity model will be adequate. From Eq. (1.38) with U replacing b, = pV/A}e/ t, = (985)(8.49x10°)(4180)/7.00 = 4990 s (83 min) (At this point we should recognize that ¢, is longer than our experience might suggest.) = @ W490; t = 2200s (=37 min) Comments: 1. Our estimated cooling time seems too long. But our various parameter values and assumptions seem reasonable--or are they? Perhaps we have underestimated the heat loss through the thin plastic lid. A carefully conducted experiment that measures the outer surface temperatures of the cup would be useful 2. Of course, when we wish to have the coffee cool rapidly, we remove the lid and perhaps stir the coffee. With the lid removed there is heat loss by evaporation, and stirring maintains a high surface temperature. The problem then involves simultaneous heat and mass transfer. (See Section 9.5.)Exercise 1-64 Given: Seven units conversions. Required: Conversion factors. @ —_ Enthalpy of vaporization, Btu/lb to J/kg, Bu _ Bw , 10553. Ib Ib 1B ub 3 045365 7 (i) Specific heat, Btu/lb °F to J/kg K oF = 4187 er sxe 3K a K (iii) Density, Ib/t¥ to kg/m? Bu wer 77926 Jb. 0.4536 ke Ae wn” ib *~ [03048m| ~ 1° ac] Gv) Dynamic viscosity, Ibffthto kg/m s “ib tb O4S36kn 1 In i Bh th * 1b ~ 03048m * 3o00 S17" (v) Kinematic viscosity, f1?/h to m/s ih * 36008 =2.581x1 es ar (vi) Thermal conduetivity, Bru/h ft °F to W/m K Bu _ Bu, 1055) ih | ift biF ~ fF < 1 Bru © 3600s < 03048m se ae (wii) Heat flux, Bru/h f? to W/m? Bru 1055 ,_1n_, |_1ft we h 1Bm * 36005 [abt] 3154 Comments: 1. Check using UNITS. 2, In ST units the symbol for hour is h, whereas in English units it is hr. However, nowadays one often sees h used in English units, as was done above,Exercise 1-65 Given: Flow rates of air in cubic feet per minute and of water in gallons per minute, Required: Corresponding values in m?/s and kg/s. 3 Amin _ a x Gps = ATIORIO4 A amoxiot ™ x 1.17 5 =5. 554x104 KE s m m CFM ms kgis 1 4.719510 5.554x10% 10 4.719x10% —5.554x107? 100 4.719x107 — 5.554x107 1000 4.719x101 5554x107 10000 4.719 5.554 6a08x10§ 2 2 6308x105 ™ 996 48 = 0.0628 8 sO s GPM m/s kes 1 6.308x10° 0.0628 10 6.30810 0.628 100 6308x103 6.28 1000 6308x107 62.8 10000 6308x101 628 Comments: 1. Check using UNITS. 2. Notice that the gallon is a measure of volume.Exercise 1-66 Given: Barometric pressure in inches of mercury at two locations. Required: (@) Units conversion to mbar and Pa, i) Boiling points @ Northway: Latm_ , 1,0133x10° Pa jnHex: 254mm lore 7 254mm ,_1ton_ 107,800 Pa Pe 31.64 inHgx " * Imm Hg * 760 tom Yam = 107,800 Pa x 2000-mbar - 1078 mbar 1x10° Pa Denver: 3387 Pa Pe 24d infigx SELES = 62,6007 = 81600Pa x 1000mbar _ 976 mbar Ix10° Pa i) From steam tables (e.g. Table A.12a), P= 107,800 Pa, Tes: = Tp = 374.8 K P= 82,600 Pa, Tyas = Typ = 367.5 K Comments: 1, Meteorologists often give atmospheric pressures corrected to sea-level, rather than the trve values, Thus a weather map may show a pressure of ~1000 mbar for Denver.Exercise 1-67 Given: Constants and formulas in ST units. Required: Conversion to English units. @ ‘Stefan-Boltzmann constant a 2x4, 1 Btwhr | (0.3048 m)* | _(.K)* = 5.6710" Win? Kt x SER x Oa b= 42,07, = (4e,)(1.712810 Brufr f2 °R‘) [ea eee aR] = 1.06¢) Btu/hr ft? °F (~ €, which is easy to remember!) Gi) Equation (1.234) i n= 1.07) SEE) wim? xm x,fi(0.3048 m/f) |“ 0.2931 W = 0.22(AT/x)" Btu/ft? hr °F, for AT in °R and x in ft ao TRC K/I = ye 1 Baur ‘Equation (1.23b) be = 1.3(AT,K)!? Win? K 7 U3, 1 Bawhe = 13(AT)°R(K/LB°R)! x Oo x = 0.19(AT)! Btu/f? hr °F, for ATin °R, 0.3048 m 1k lft Comments: 1. Note that the Grashof number ranges for Eqs. (1.23a) and (1.23b) remain the same because the Grashof number is dimensionless. 2. Many sources give o = 1.714x10" in English units.Exercise 1-68, Given: Problem statement of Example 1.3. Required (i) Convert problem statement to English units. (ii) Work the problem in English units. Gi) Convert answer back to ST units. @ = «— W=4m= (4)(3.2808) = 13.12 ft °C = (273)(1.8) = 491.4°R V= 14x10 m/s = (14x10*)(3.2808)? = 1.5x10™ fs 2 = 9.81 m/s? = (9.81)(3.2808) = 32.2 fs? i) First find the transition location, _ fe]? _ faora.seo? |" _ oy “| (AT/Mg | ~ | 8/500.4\(32.2) iH Find the average heat transfer coefficient h, (see Exercise 1-67) ca L fo ne L J 022¢01%)“ax + J 0.19(07)"Fax| x = (1/7.55){(0.22)(4/3)(18)(2.69)"* + (0.19)(18)9(7.55-2.69)] = 0.488 Brufhr ft? °R ‘Then, from Eq. (1.24) the convective heat loss to the door is, Q=hAAT = (0.488)(13.12)(7.55)(18) = 871 Brufhr Gil) Convert back to ST units. Q= (871)(0.2931) = 255 W Comments: 1. The formulas forh, are only approximate, and hence were rounded offto two signi- ficant figures in Exercise 1-67 thus we do not recover exactly the same result for Q.Exercise 1-69 Given: Problem statement of Example 1.4. Required G) Convert problem statement to English units. ii) Work the problem in English units. (ii) Convert answer back to SI units. @ La =2 m= (0.02)(3.2808) = 0.0656 ft Lp =5 cm = (0.05)(3.2808) = 0.1640 fe ‘hi = 3 Wim? K = (3)(5.678) = 0.528 Btu/hr f2 °R igo = 6 Wim? K = (6)/(5.678) = 1.057 Btufhr ft? °R. Te =2°C= 275)(1.8) = 495.0°R Ty = 20°C = (293)(1.8) = 527.4°R Bigo = 420T3, = (4)(0.9)(1.712x10° (495)? = 0.748 Brufhr f °R K, = 0.10 Wim K = (0,10)/(1.731) = 0.0578 Buu/hr ft °R kg = 0.038 Wim K = (0.038Y(1.731) = 0.0220 Bauytr ft °R Gi) The overall heat transfer coefficient is given by La = (1/0.528)+(2)(0.0656/0.0578)+(0.1640/0.022)+(1/(1.057+0.748)) = 12.17 (Btufhr fi? °RYT or = U= 0.0822 Btu/hr fi? °R ‘Then the heat flux is 4= UCT;~Te) = (0.0822)(527.4-495.0) = 2.66 Buu/hr ft” (ii) Convert back to SI units. q= 2.66 Buufhr fi? = (2.66)(3.155)= 8.40 Wim? Comments: 1. The results differ because rounded off values were used in the computations.Exercise 1-70 Given: Problem statement of Example 1:5. Required (i) Convert problem statement to English units. (ii) Work the problem in English units, Git) Convert answer back to St units. @ — T=17C= @90\(1.8) = 522"R Ty = 5°C = (278)(1.8) = 500.4°R ‘= 10 Wim? K = (10)((5.678) = 1.76 Biufhr ft? °R = 1712x107 Btu/hr ft °R Gi) ‘Substituting from Egs. (1.24) and (1.18) BACT.) + 20A(T¢—T) = 0 From Table 1.3, €= 0.8. Cancelling A, 1,76(522-T,) + (0.8)(1.712x10°)(522.0°-500.4") Solving T, = 531°R. Gi) Converting to SI units, Te = 531°R = (531)/(1.8) = 295 K Comments: 1. ‘The fact that 0°C = 273.15 K is a nuisance. We can usually round offto 273 K for engineering calculations, but, more importantly, we should be consistent in a given set of calculations. 2. Many sourees give o = 1.714x10° Bru/hrExercise 1-71 Given: Problem statement of Example 1.6. Required: (i) Convert problem statement to English units. (ii) Work problem in English units. Gil) Convert answer back into ST units, 013.2808) = 0.0328 fe 73)(1.8)= 1571°R Te = 30°C = (303)(1.8) = 545°R T= 100°C = (373)(1.8)= 671°R , = 50 Wim K = (50)(1.731) = 28.9 Brufhr ft °F (p= 7800kg/m? = (7800)/(16.018) = 487.0 lb/ft? © = 450 Ikg K = (450)(4187) = 0.107 Btu/lb °F 1, = 400 Win? K = (400)/5.678) = 70.5 Btu/hrft °F BLP) _ (70.5)0.03282) _ 994201 ww ke 28.9 ell Ade _ (487.0V0.0328/2)00.107) = 46191 he 705 Gi) t= (0.0254)(3600) = 91s Comments: 1. The | s discrepancy is due to round-off error,Exercise 1-72 Given: Equation (1.22) giving h, for turbulent flow in a tube. Required: A units check for (i) SI units, and (ii) English units. Equation (1.22) is eae noo at @ STunits Gi) English units te] [pm hritF Comments 1. 8 kg. m?> kek aa [< °* Tab ele Since the dimensions on each side of the equation are the f° Ib°F 04 pene mK 04 Bu Bu fP br °F same, the constant 0.023 is dimensionless, and is the same irrespective of whether SI or English units are used.Exercise 1-73 Given: Quadratic curve-fit coefficients for steam line insulations in English units. Required: Curve fits in SE units and graphs of K(T). Assumptions: Accuracy of data no better than 1%. For T in °F and k in Btu in/ft? h °F we have Kk = AgtA,T+A,T? a) Itis convenient to first change from °F to °R. Notice that Ay is the value of k at T = 0°F; thus, Eq. (1) can be written as k = k(0)+A,(T-0) +A,(T-0)? @Q) for T in °F. Then, for T in °R, k = k(460) + Ay(T-460) + A,(T-460)? = (Ap - 460A, + 2.116x10°A,) + (A,-920A,)T + AGT? @) or, k= By + B\T + BAT, for T in °R. @ Now we convert to SI units. Notice that the units of By, B,, and B, are those of k, k/T and K/T?, respectively. ‘Thus, using Table B.2, 1 ft 1.731 Wim K B, [Wim K] = B,[Btu in/ft7h°F] a baer = 0.1443 BofBtu in/ft? b °F] B, [Wm] = B,{Bw int?noF7}0.1443) LSE = 0.2597 B,[Btm infft? h °F?) B,[WimK}] = B,[Bu ianer-rryo es, EE = 0.4674 B,[Btu in/f? h °F} Substiouting in Bq. (4) gives k in W/m K for T in kelvins k = (A)-460A, +2.116x105A,)(0.1443) +(A, ~920A,)(0.2597)T +A,(0.4674)T? (5) | or, k= Cy + C,T + CT? © ‘The table gives Cy, C,, and C, for the four insulations.1-73 (continued) Brand Name © cg G ‘Thermo-12 0.0534 -4.67x10% | 117x107 ‘Kaowool 0.0297 -3.13x10° 175x107 Epitherm-1200 | 0.0744 | -252x10* | 467x107 Kaylo 0.0894 -1.82x104 2.92x107 Comments: 1. Three significant figures for the coefficients are sufficient to ensure 1% accuracy. 2. The original data are valid for temperatures between 100°F to 1100°F. Thus the SI curve fits are valid for temperatures between about 300 K to 900 K, and must not be used outside this range. 0.25 Thermo-12 on — Kaowool Epitherm-1200 Kaylo = 045, & z x 0.10. 0.08. 300Exercise 1-74 Given: Miniature heat flux meter with calibration constant of 0.350 LV/Buu/hr ft. Required: (i) Calibration constant in SI units. Gi) Surface temperature correction for q = 100, 1000, 10,000 W/m’. Assumptions: 1. One-dimensional steady conduction, 2. Copper and thermopile thermal resistances are negligible, @ Using Table B.2, 2 1 Be 41 BY fioss 2 }_1_ fC ft he ft?hr ft?hr Buu { 0.3048? m? {3600 s = 3.154Wim? 2 Hence, 0.350" _ = Josso_HV_|_1_ Beu/ft"hr Bro/ft7hr Buw/ft7hr } 3-154 W/m? = 0.111 pV(Win?) (or use 3.155 (W/m?)\/(Btu/f he) directly from Table B.2). (i) From the diagram of the sensor, the thermocouple is separated from the surface by 2.5 mil Kapton, 1 mil copper and the thermopile. The thermal resistances of the copper and metal thermopile can safely be assumed to be negligible. For a heat flux q and one-dimensional conduction, For L = 2.5 mil = 2.5x10° in = 6.35x10° m, and k = 0.245 Wim K, T,-Ty, = (6.35x10*)K0.245)q = 2.59x10* q q, Wim? 100 1000 10,000 T.TeK 0.026 026 26 Comments: 1. Nowadays we use commonly used computer based data acquisition and data processing programs when performing experiments. The calibration constant and surface temperature correction should be included in the data processing program.Exercise 1-75 Given: Definition of the clo in English units as a measure of thermal resistang¢. Required: (j) 1 clo in ST units Gi) Resistance of a 2 mm thick wool sweater. (ii) Heat loss across a 0.5 clo cotton shirt. ‘Assumptions: 1. One-dimensional steady conduction for part (ii. » @ 1 clo = 0.88 ft? °F hr/Bea ‘The units of clo are the reciprocal of those for the heat transfer coefficient, From Table B.2, we have 5.678[W/m? K¥[Buwft? hr °F]. Thus 2 4 Delo = ogg ft? br}, 1 (Btwit? hr Fi] Be | 5678 Twi? Ky = 0.155[W/m? Ky" (ii) A.2 mm thick layer of wool of thermal conductivity 0.05 W/m K has a thermal resistance per unit area of = 0.08{Wim? Ky"! > Ab ont onasier clo = 0.04[Wim? K]* —_— ene a [Wan? Ky? = 0.258 clo iii) From Eq, (1.9), TT, _ TT, + 31-28" TA” RIK os = 6 Win? K , ‘ Comments: 1.° in practical caleulations concerning heat transfer ‘from the human body, the thermal resistance of, clothing is usually defined to characterize heat transfer between the skin and the outer surface of the clothing. That is, it includes the resistance associated with conduction ‘and radiation between the skin and the innerExercise 2-1 Given; A slab with heat flux q imposed on one face, maintaining the face at temperature Ty. Reguiféd: ‘Temperature profiles for variable thermal conductivity k = ko{1+a(T-To)]. Assumptions: 1. One-dimensional steady conduction. From Eq. (1.8), Q ar =q=-k © ~ constant, q cacenetees exch Hence q=—koll¥act—ton1 t Jaar=— J kotteact-tonar a : axel i wna 5 “ky l(1-aTa(T-T) + aC? 9)} % @a=0: T=T,—qy/ko, the usual linear variation. Gi)’ .2#0: Rearranging gives a quadratic for T; T+ (2/a-2T9)T-T} — Ty @/a-2T) + 2qx/kya = 0 Solving, T= —(/e-To) + ((1/a-To)? ~T} + (2/a-2To)T1—2ax/koal” ‘To sketch the profiles for a + 0 we should examine the sign of d?T/dx*: Kf+a(T—To)] va K§[1+a(T-To)]° For q positive, a >0 gives d?T)dx? negative, and a<0 gives d? Tix? positive. Comments: 1. In practice one might choose To = T;, in linearizing the variation of thermal cn” ductivity.Exercise 2-1 : Given: A slab with heat flux q imposed on one face, maintaining the face at temperature Ty. Required: Temperature profiles for variable thermal conductivity k = ka{1+a(T-Tp)]. Assumptions: 1. One-dimensional steady conduction, From Eq. (1.8) Hence q=-kallvacr-Ty)) SE J n=] tttear-toner a n= [cra [s va] ekol(l-arayeren) + 4 att] Ty @ — a=0: T=T)-qwhko, the usual linear variation. GD) a0: Rearranging gives a quadratic for T; T? + (2fa-2T9)T-T} -T1 Q/a-2T) + 2qwlkga = 0 Solving, T=—(1/a-To) + {(1/a—To)? — Tf + (2/a-2T9)T1-2qnkoa]* ‘To sketch the profiles for a #0 we should examine the sign of d°Tyax*: a __4__ dx Ko{l+a(T=To)] eT @a dx? KL +aCT TO) Forq positive, a> 0 gives d?T/dx” negative, and a< 0 gives d°T/ax" positive. Comments: 1. In practice one might choose Ty = Ty in linearizing the variation of thermal con- ductivity.Exercise 2-2 Given: A 2.mm-wall thickness polystyrene coffee cup. Required: Thickness of seven materials to give the same thermal resistance, Assumptions: 1. One-dimensional steady conduction. 2. All thermal conductivities may be evaluated at 300 K for comparison purposes. From Eq, (1.9) the thermal resistance pet unit area for conduction across a plane wall is R = L/k. From Table A.3, the conductivity of polystyrene at 300 K is 0.028 Wim K. Hence, R = 0,002/0.028 = 0.0714 K/W per square meter. @ — Polyvinylchloride. From Table A.2, k= 0.092 Wim K. Hence L=Rk = (0.0714)(0.092) = 6.6x10° m (6.6 mm) i)-(vii) Similarly for the other materials. The results are presented in the table. Material k L Win K m Polystyrene 0.028 0.0020 Polyvinylchloride 0.092 0.0066, Paper 0.13 0.0093 Oak 021 0.015 Stainless Steel 15 1.07 Brass ut 7.93 Copper 401 28.6 Diamond (type Ib) 1300 928 Comments 1. A thickness of almost 100 m of diamond is required to insulate as well as 2 mm polystyrene! Of course, diamond is used when a good heat conductor is required, for example, as a “heat spreader”, See Exercise 1- 14.Exercise 2-3 Given: A slab made up of n pairs of square bars of conductivities ky and kg, as shown. Required: (j) Effective conductivity across the width. i) Effective conductivity through the thickness. Assumptions: 1. One-dimensional conduction. @ Across the width the thermal resistances are in series, AT Kaka AAT *kgtkp ot peace EEE . eA 4 Compare with Q= FF AT; W = 2nt rf an \ _ 2 ake welt LL Snpairs of pars Hence kee = Fig (ii) Across the thickness the thermal resistances are in parallel, cs 1 kA KA Kathe A = /E AT =| 4 = AT =—— = AT 7 2 [: 2412 piace - A Compare with Q= “= ar ssa OD Comments: 1. A more difficult problem is the estimation of the effective conductivity of a composite such as a liquid filled porous solid. 2. Implicit in the assumption of one-dimensional conduction is that the faces of the slab are isothermal. If heat transfer to and from the slab is by convection, the faces will not be isothermal. See Exercises 1-27 and 1-34.Exercise 2-4 Given: A simple kinetic theory formula for thermal conductivity of a gas. Required: k values for (i) air at 300 K, 1 atm; Gi) air at 400 K, 1 atm; (iv) helium at 300K, 1 atm, Assumptions: 1. Ideal gas law is valid. ir at 300 K, 0.01 am; Equation (2.3) gives the thermal conductivity from simple kinetic theory as 6 4 ai 1p-1| 3 Sram 1p- at =F eM ey + > BD 5 [as w/e (Mey + > K)= vNincy + 2 NE) ‘Now, introducing v = /p; Nm=p, N=P/KT, cy-ty =R and the ideal gas law P/p = RT, kame +4 R=we- FR, @ — Airat300K, 1 atm. From Table A.7, p= 18.43x10 kg/m s, cp = 1005 J/kg K at 300 K. Also, R = R/M = 154529 = 53.3 Jikg K. k= (18.43x10%)(1005-53.3/2) = 0.0180 W/m K. Also, from Table A.7, k = 0.0267 W/m K. i, Gi), Gv) Similarly; the results are tabulated below. Case T OP H & R kk (Table A.7) K am kgims kg kg Wim K = W/mK @air 300 1 1843x10 1005 53.3. (0.0180 0.0267 Gi) Air 400 1 © 22.52x10% 1009 53.30.0221 «0.0331 Gii) Air 300 0.01 1843x10% 1005 53.3. 0.0180 0.0267 @v) Helium = 3001 20.10x10% 5200-386 O.101 0.149 Comments: 1. Equation (2.3) gives values which are too low, due to approximations made in the simple kinetic theory model. A more accurate formula relating k to pis K/t1= (Cp/*V(L.TP-O45); Y= p/cy. For air with y= 1.4, k= 145 picy. At 300K, k= (1.45)(18.43x10")(1005) = 0.0268, which is in excellent agreement with Table A.7. 2. Notice that kis independent of pressure, increases with temperature, and decreases with molecular weight.Exercise 2-5 Giver: Aluminum/aluminum interface having an interfacial conductance in the range 150 Tpp, fogging will not occur. 2. Dewpoint and relative humidity are concepts used in psychrometry, a topic dealt with in Chapter 9, and in your thermodynamics text,Exercise 2-10 Given: A 2 om thick composite plate with electrical heating at its centerplane, and cooled by air on each side. Required: Maximum rate of heat generation if the composite temperature cannot exceed 300°C. Assumptions: 1. One-dimensional steady conduction, 2. Constant thermal conductivity. ‘ @ Te 7 @) Te For steady one-dimensional conduction the thermal circuit is as shown. For 1 m* cross-sectional 200% area we have aT.) Q= De Q 5 also, Q)+Q, = Quice hE Substituting T,, = 20°C, T,, = 100°C, h, = 40 Wim? K, L = 0.01 m, k = 300°C, 45 Wim K, T, = Poe re (300-100) _ . =~ = 5.93 KW; Q, = SD) = 424 kw ee ncipaaoT & =F o01 a 045 045 Quice = 5.93+4.27 = 10.17 kW Comments: 1. Quyq. may appear to be large, but note that it is for a 1m x 1 m plate. For a 10 em x 10 em plate, Q.j,. = (10.17x10°)(0.1x0.1)=101.7 W. 2. Always draw the expected temperature profiles and the thermal circuit for this type of problem-it will help you formulate your solution procedure.Exercise 2-11 Given: A composite wall of fireclay brick and fiberglass insulation. Required: Heat flow per unit area for given surface temperatures. Assumptions: 1. One-dimensional steady conduction. 2. The temperature dependent k-values can be evaluated at appropriate average temperatures. Equation 10) apis Ry =Lalka; Ry =Lp/kp T,9540K ‘We first guess an interface temperature of T2 = 500 K, and evaluate ky at (500+540)/2 = 520 K and kp at (300+500)/2 = 400K. Using Table A.3, k, = 1.0 Wm K, and ks = 0.066 Win K. Ra .20/1.0=0.2 [W/m KT"; Ra = 0.05/0.066 = 0.758 [W/m? KJ ATs Lalka __02_ Ty Lp/ky 0.758 ‘Asa second iteration we take T; = 490 K; then ky at 510 K-remains 1.0 W/m K (to the accuracy of the data) and ky at 395 K = 0.0645 W/m K. ‘The resistance Ry becomes Ry = 0.05/0.0645 = 0.775, and AT, __02 Tp 0.775 ‘The heat flow can now be calculated as Q _ 540-300_ 2K, A 0.240.775 an = 0.264; AT, =(0.264/1.264)(540-300) = $0.1 0.258; AT, = 49.2; Tz = 491 K, as desired, Comments: 1. ‘The listed data for the thermal conductivity of fireclay brick is given to only one significant figure, because of variations from batch to batch. 2. ‘The important idea demonstrated in this exercise is that the thermal conductivity of the dominant resistance, i.., the fiberglass, should be evaluated at a satisfactory ‘estimate of its average temperature. A precise value of k for the brick is not required.Exereise 2-12 Giver: A composite wall consisting of a stainless steel plate and two layers of laminated asbestos paper. Required: Heat flow per unit area. Assumptions: 1. One-dimensional steady flow. 2, The temperature dependent k-values can be evaluated at an appropriate average temperature. 3. ‘The thermal resistance of the steel plate is negligible. ck ate as 1 kA KA KA RA La =1mm;Ly =2em;Le=2em T, = 380K; T. =300K b= 5 Wim?K ‘From the thermal circuit —T, aR We take Ty = since the steel resistance is negligible, and guess Ts = 340 K,T4 = 300 K to cobuin kvalues. Then we evaluate kp at (380+340)/2 = 360 K, and Ke at (340+300)/2 = 320 K. Using Table A.3, ky =0.097 Wim K, ke = 0.073 Wim K. Fora unit area, wae eS Ra Demy tes tg ToIeoaTE02 [Atthis point we see that the three resistances are nearly equal, and our guess of T = 340 Kis too = 325 K as.a second guess and proceeding as above gives ka = 0.099 Wim K, ke = 0.07 Win K. FR Qo5 "077 5 ‘Ty-Ty = (Ra/ERNT Te) = (0.202/0.662)(80) = 24.4 K: Ts = 356 K TyTe = RAERYT Te) = (0.26010,662)(80) = 31.4 Ks Ty = 324 Cur second guess was very good so no further iteration is required. Finally, Q_ Th __80__ ‘A ER 0662 .202+-0.260+0.2 = 0.662 (21.0 Win”. ‘Comments: 1. Usually no more than two iterations are required to give sufficient accuracy ‘when the temperature dependence of an insulation conductivity is allowed for.Exercise 2-13 Given: Composite wall as shown. Required: Heat flow through wall using two models. Assumptions: 1. Constant thermal conductivities 2. One-dimensional heat conduction. ‘The heat flow is two dimensional; however we can obtain approximate results by assuming one- dimensional heat flow. ‘Two different models can be tried. (Surfaces of constant x are isothermal. Then the thermal circuit is: Q RyRy Rey Ry Ry = LalkyAg = 0.05/(1.0)(0.1)(1) = 0.5 KAW per meter width, Ry = Lpfkg Ag = 0.12/(0.3)(0.04)(1) = 10.0 KAW Re = Leche Ag = 0.1(0.60.06)(1) = 3.33 KW Rp = Lp/kpAp = 0.04/(10)(0.1)(1) = 0.04 KW a: 100-0 0.5+(1/10+1/3.33) 140.04 = 32.9 Wim Gi) There is no heat flow transverse to the x-direction. Then the thermal circuit becomes:2-13 (continued) e- 1,7, [OR Ry PRYOR AREA) | = uot = 319 Wim [(2x0.5+10+2x0.04) 1 +(2x0.5+3.33+2x0.04)" Comments: 1, The two models give almost the same result: the true value lies inbetween, 2. The difference in Q obtained with the two models depends in a complicated manner on the relative values of Ry, Ry, Re and Rp. In general the difference increases with IRy-Rcl, but also the difference decreases as R, and Rp decrease. For Ra Ry ~? 0, the difference + 0, See also Exercises 1-27 and 1-34. 3. The exact solution can be obtained using numerical methods (sce Section 3.5).Exercise 2-14 Given: A 1 cm thick stainless steel reactor wall insulated with a 5 cm thick fiberglass layer. Required: (i) Thermal network, Gi) Inner surface temperature when outer surface is at 32°C. Assumptions: 1. One-dimensional steady conduction. — 5°25em of — 9*35em —f — 4785om ‘The heat flow Q can be calculated from the parallel part of the themal circuit, Q=2rLrs(h+h,(TsTe) However, there is no advantage to using the radiation heat transfer coefficient in this problem: instead we use Eq, (1.18) directly to write Q=2nLrs[h.(Tj—T,) + oe(T$-TH] (2x \(2)(0.085)[6(32-25)+(5.67)(0.8)(3.05%-2.984)) = 82.0 ‘Now temperature T; can be calculated using the remainder of the circuit, Ti Rate Ty-32 in.572.5)_, __In@.5(35)_ ~ 0.00167+1.308 ANMAVI) — OXm2V0.054) where kp, the fiberglass conductivity, has been obtained from Table A.3 for a guessed average temperature of 360K. Solving, Ty = 139.4°C. Since the thermal resistance of the stainless steel ‘is very small, the average temperature of the fiberglass is (139.4+32)/2 = 85.7°C = 359 K, which is close to our guessed average of 360 K. No further iteration is warranted. 820 Comments: 1. Usually no more than two iterations are required to give sufficient accuracy ‘when the temperature dependence of an insulation conductivity is allowed for.Exercise 2-15, Given: Milk at 5°C flowing inside a stainless steel tube insulated with a 5 em thick layer of cork. Required: Calculate the rate of heat gain per meter length of tube. Assumptions: 1. One-dimensional steady heat flow. 2. The thermal resistance of the steel tube is negligible. 3. Radiation heat transfer negligible. S°C, Ty =25°C 50 Win? K Win? K cork = 1 = 002m y=007m steel t= 0.002m From Table 1.1, k = 0,043 Wim K for cork. The resistance of the tube wall is negligible so that a k value for the stainless steel is not required. Multiplying Eq. (2.17) by L and introducing the perimeter P= A/L gives re L Ineo) UP 2m(-Dhe; 2k Ory 1 +) _, 1 2HO.018)(50) * (2)(R)(0.043) " (2)(2)(0.07)(5) .176844.63740.455) 5.269 m K/W; UP = 0.190 Win K. The rate of heat gain per meter length of tube is QL = UPCT-T;) = (0.190)(25-5) = 3.80 Win ‘Comments: 1. In this exercise the main thermal resistance is the cork. If we also ignored the intemal and extemal heat transfer coefficients the error would only be about 5%. 2. ‘The assumption of negligible radiation is consistent with use of an aluminum foil sheath over the cork, both to seal the cork from penetration of moisture, and to give a low surface emittance,Exercise 2-16 Given: Saturated steam flowing through an AISI 1010 steel tube insulated with 85% magnesia, ‘Required: Compare heat loss from the bare tube to that for the insulated tube, Assumptions: 1. One-dimensional steady flow. 2, ‘The thermal resistance of the stee! is negligible. 3, ‘The temperature dependent k-value can be evaluated at an appropriate average temperature. 4, The inside surface of the tube is at the steam saturation temperature (h; 1, T, =200°C T.= 20°C { Typ Comments: 1. Condensation will not occur on the outside of the insulation. 2. The insulation must be carefully sealed to ensure moisture does not penetrate to radii where T < Typ. Condensed water or ice in the insulation will significantly erease its thermal conductivity and reduce its insulating effect. 3. The thermal resistance of the fiberglass insulation controls the heat flow. It is not necessary to have accurate values for h.. and h,Exercise 2-27 Given: R-12 flowing in an 8 mm O.D. copper tube. Required: Heat transfer Q/L as a function of thickness of foam insulation. Assumptions: 1. One-dimensional heat flow. From Eqs. (2.16) and (2.17), with k = 386 W/m K for copper. 345.5 T.111-+0.0007-+ 28.6 In(t3/0.004)-+ 1/515 8 Insulation fr} —Q/L. 7 ‘Thickness E mm mm Wan oa 0 4 6.76 Rees 1 5S 728 2 2 6 751 a4 3 7 (757 HeES: 6 10 731 8 12 7.03 2 10 14 675 1 16 20 6.05 46 50 4.47 96 1003.63, 20 40 60 80100 thickness, mm Comments: 1. A large thickness of insulation is required to appreciably reduce the heat loss: it probably should not be used. 2. Use of a constant eo is perhaps an oversimplification; see Exercises 2-29 and 2-30. 3. ‘The resistance of the tube wall is negligible.Exercise 2-28 Given: 10 GPM of brine at -5°C flowing in a 1 in Schedule 10 copper pipe. Required: ‘Thickness of insulation (k = 0.2 W/m K) required to prevent condensation om the outside if the ambient air is at 20°C and has a dewpoint of 10°C. Assumptions: 1, The inside convection resistance and pipe wall resistance are negligible. At steady state the heat transfer by conduction across ‘the insulation must equal the convective heat transfer to the surface; using Eq. (2.14), 2xkL(T2-T1) Talfainy | 7 BOMALNT Ta for Qpositive Q From Table A. 14a, the outside diameter of the pipe is 33.4 mm; hence 1 = 0.0167 m, and Solving by trial, 7) = Insulation thickness = r)—1, = 0.0358-0.0167 = 0.0191 m (19.1 mm). 5 Comments: 1. The inside heat transfer coefficient can be estimated using Eq, (1.22) and the properties for brine in Table A.13b, A value of h; ~ 3000 Win? K is obtained, so that the inside convection resistance is indeed negligible. 2, When a porous insulation becomes wet its conductivity increases. Thus “vapor barriers" are provided for operating conditions where condensation can occur. For example, a layer of aluminum foil may be used to prevent moist air from entering a porous insulation (and also reduce radiation heat transfer to the surface).Exercise 2-29 Given: A.1.5 mm diameter electric wire insulated with a plastic material (k= 0.37 Wim K). Required: Plot of wire and insulation surface temperatures as a function of insulation thickness. Assumptions: 1. One-dimensional steady heat flow. 2. Uniform temperature within the wire. 3, ‘Small gray body in large, nearly black surroundings. Equation (1.33) gives the heat loss for the bare wire as ON = Qnalhe + Qoaw/l= Dns the(T,-T.)+e0T 19] = (2n){0.75x10){(1.3)(313-293)"*(1.5x10)% + (0.07)(5.673.13*-2.93%)] = ‘When the insulation is added we can set up the thermal circuit, 358W 1 1, - 1, i fl 3 2niryh, : a ahe In/r) 21k 1 2a h, where T, and Q/L are known and T; and T, are unknown. From Eq. (1.19) h, = 42073, = 20(T.+T,)*/2. Now we can set up two equations and solve for the two unknowns: . 2reg(heth. XT Te) = QL _ ank(T\Ta Vln) = QL he where rp = r+, and tis the insulation thickness. From Eq. (1), Inte HTT IL-3, -Te)4D +21)“ + 20(T. +72] = YL T, Te © Ant 4+D11. 3. -T)"D 420 + eoTe+ TY 2I - = 1358 +293.15 ‘2n(0.75x 10-3 +1){1.3(T y-293.15)"(1.5x10%+21) % + 2.552x10-§(293.15+T, 9] ‘This equation requires iteration to solve for, for a given t. When we have T, we use Eq, (2) 10 find T;, QnA). _ .3SB)in((O.15x10H1Y075x10%) atk : 2n(my0.37) :2-29 (continued) Evaluating T, and T; for varying thickness t gives the following results. t(mm) | 01 10 SO 10 SO 100 qT, | 3073 301.7 296.7 2953 293.7 2935. v; | 3074 3022 2979 2969 296.2 296.3 critical thickness of insulation for which T is a minimum is ~70 mm. Comments: 1. The student is encouraged 2. The calculations in this exercise are tedious if done by hand. The to write a simple computer program to execute the calculations.Exercise 2-30 Given: An insulated tube losing heat by natural convection. Required: The critical radius of the insulation and corresponding heat loss. Assumptions: 1. One-dimensional steady conduction. 2, Negligible heat loss by radiation, In(r,/ry) 2mLk 2rrolhe From the thermal circuit the heat flow is TT, R Ino ft) - 1 2mLK” 2aLrg[1.3(AT/21)) where R= 3 ATT AT, Since T, is a function of T;—T,, and ry, we may not use dR/dr, = O to find the maximum of Q. Instead, we obtain the solution numerically, as follows. Since the heat flow is the same through both resistances, -T. 1 Date (1.3{(Ty—Te)/2r9]*} Eire asOate Laat ,-300 L ~ tn(r,/0.003) I 2m(0.08) nr (1.3{(T,-300)/2r5}*) For given rq the above equation may be solved numerically for T, and Q, with rg varied to find Qmax/L and the corresponding t= re,. The result is eq = 0.0046 m (4.6 mm) Qurex/L = 12.1 W/m 1, = 339.7K. Comments: 1. ‘The student is encouraged to write a computer program to execute the iterative solution procedure. 2. Atr=r, bh, =1.3{(339.7-300)/(2)(0.0046) = 10.5 W/mK. Then hr, /k = (10.5)(0.0046)/(0.08) = 0.6; o.f. a value of 1 given by Eq. (2.19) for h, constantExercise 2-31 Given: A 1mm diameter insulated resistor located in an evacuated enclosure. Required: Critical radius of insulation. Assumptions: 1. ‘The resistor temperature is uniform. 2. Heat loss by convection is negligible. —@)VWWA InChy/ %) 2TTKL 2TT rhb From the thermal circuit, 6 teh | Tet Lo lnGtoft,) = _1_ 2k 2ntoh, ‘where from Section 1.3.2 the radiation heat transfer coefficient is h, = oe(T?+T2)(T,+T.). Hence QO 450-7, T,-300 L ~ In(fo/0.0005) 1 sea 2m(0.12) —_2rRro(5.67x10°*)(0.85)(T?+3007)(T, +300) Fora given value of fo, this equation can be solved numerically for‘T, and Q. The value of r, is ‘varied to find Qmax/L, which corresponds to f> = tz. The result is Fer = 0.0161 m (16.1 mm); Quay/L = 24.3 Wim; T, = 338.2K Comments: 1. ‘The uninsulated resistor with the same emittance has a heat loss of Q/L = 2nn,oe(T}—T$) = 2m(0.0005)(5.67)0.85)(4.5*-3.0") = 4.98 Wim. 5.67 x10" 0.85) 338.2? +300) = 6.29 W/mK (6.29) (0.0161)/0.12 = 0.84; c.f. a value of 1.0 given by Eq. (2.19) for constant hy.Exercise 2-32 Given: A sheathed resistor cooled by forced convection and radiation. Required: ‘The critical thickness of insulation sheath for (i) €= 0.9 and (ii) e= 0.5. Assumptions: 1. Steady one-dimensional conduction. 2. The resistor temperature is uniform. thongant (lab) where T; = 400 K, T, = 300 K; rj = 0.001 m; k = 0.01 W/m K; hy = 1.121, by = oe(T24T2)(T,4Te) ‘Equations (1a) and (1b) are two coupled algebraic equations which must be solved numerically by iteration. A simple computer program can be written to perform the task. The results are as follows. @ =09 Tr = 0.0027 m, which corresponds to T, = 362.4 K, and (Q/2L)max = 3.784 Wim. Gi) e=05 Ter = 0.0033 m, which corresponds to T, = 359.0 K, and (Q/2tL)max = 3.433 W/m. ‘Comments. 1. The bare resistor with an emittance of 0.9 has aheat loss of Q/2nL = (0.001)[1.1(0.002)% + (5.67x10°*)(0.9}(400? +300? (400+300)] (100) = 3.353 W/m The maximum heat loss for € is only 13% larger.Exercise 2-33 Given: A 2 rom diameter wire with 2 mm thick electrical insulation of k = 0.12 W/m K. Required: (j) Effect of increasing the insulation thickness to 3 mm if h, = 12 W/m? K, (ii) Effect of a contact resistance of 5x10“ (W/m? KT". ‘Assumptions: 1. One-dimensional steady conduction. 2. h, is independent of the insulation radius. ( Increasing the outer radius of the insulation r, will increase the heat transfer provided r, 0, aQ = AQ.” ‘ax e Introducing Fourier’s law and the given expression for Q,. differential equation ar fie T. -Gk. dx? K which has the solution G TER) = eC, and assuming k constant gives the ‘The constants C, and C, are obtained from the boundary conditions: x=0; T=Ty; x=L: T = T,; and the resulting temperature profile is e 0 To) = T+ pl -e™)+| 120 100 ‘To find the location T,,,, we set dT/dx = 0, = 6 oT. Gynt S fi-e*!] = 0; hence xk Lk 2 ° 1 ee ee x | GL ca Comments: 1. y= 100°C; T= 0°C;L = 1m k=1 Wink: «=0.tm" c= s000Win 1 D 01 02 03 04 05 08 07 08 09 x(m) Problems of laser heating of materials usually involve transient heat conduction, rather than the steady conduction considered here.Exercise 2-50 Given: Naclear heating at 10 MW/m? in a 0.5 m high, 1 cm thick ceramic slab clad with 1 mm thick stainless steel. Required: Maximum temperature in slab when cooled with water at 30°C. Assumptions: 1. One-dimensional steady conduction in the slab. 2. Constant thermal conductivity. Conduction with intemal heat generation was analyzed for radial heat flow in Section 2.3.4. Hexe we have planar heat flow; proceeding as in Section 2.3.4 we apply the energy conservation principle to an elemental volume located between x and x4+Ax, Qh Onan *QyAAK = 0 Dividing by Ax and letting Ax > 0, dQ/dx = Q7"4. Introducing Fourier’s law, Q = qA = -k(dT/dx)A at __& : Q qe + integrating, T = -—- x? 4C,x+Cy. From symmetry dT/dx = 0 at x = 0, hence C, = 0. Specifying T=", at x=L gives C, = 7, + (Q) "ML? and then T-T, = (Q,"/2K)L? x) @ Atx=0: Tyg -Ty = (Q, L/2k @ To obtain T, we equate the rate of heat generation in the slab to the rate of heat loss to the water. 7 T,-T. Q)LA = —_*_<_. and A cancels. TIAA) 1,-30 (0.005) = ——_____ 10x10"(0.005) = pay xDD) 205.7°C ‘Substituting in Eq. (2), Tay = 205.7 + (10x10°}(0.005)"(2)(3) = 247.4°C. Comments: 1. We have ignored the possibility of a significant interfacial resistance at the ceramic-stainless steel interface. An effort should be made to find appropriate data: unfortunately, the interfacial conductance depends on such variables as contact pressure and thermal expansion in a complicated manner.Exercise 2-51 Given: A large slab with intemal heat generation losing heat by convection to a liquid at temperature T,, Required: Steady state temperature distributions for Q,” = (i) a constant, Gi) Q, {1-(w/L)"], el 7 Gi) atb(T-T,), Assumptions: 1. One-dimensional steady conduction. 2. Constant thermal conductivity. | | Applying the energy conservation principle to an | Lida, elemental volume Ax thick gives i Qt QO AAK = Qlesae | l Q = -kAdT/dx; dividing by Ax and letting Ax — 0, Ft aT __@& @ dx? k @ Q)” =constant, Then integrating, a _ & Fe stat, @ 8 Boundary conditions: x = 0, = = 0 (symmetry); x=L, -k s =heT-T) (4a,b) ix From Eq, (4a), C, = 0. Then substituting in Eq. (4b) . a” L = h]-_L?+C,-T, Q { OK ao he Solving, C, = . Substituting in Eq. (3),2-51 (continued) 2 AL He eee a i = © Gi) Q)” = QP T-G/L)"]. Substituting in Eq. (1), aT _ _Qvo})_(x ew KL IL ~ © om Again, from Eq. (4a), C, = 0, and substituting in Eq. (4b) gives C>, ‘The resulting temperature profile is 1fx¥ (x¥, 4 eel eeee 8) Gy (i) “| i ee por, = Qok? 1 oiteaake| 2 (ii) Q,” = a+(T-T,). Substituting in Eq. (1) GT, bip_fy -2)] « 0: solving dx? | b t(-¥] = Beos (b/k)'x *Dsin (b/k)"?x Equation (4a) gives D = 0, and substituting in Eq. (4b) allows B to be determined. ‘The resulting lemperatuze profile is h,(a/b)cos(b/k) x. r-(.-4) {afb }oos(bM) ep fee cece cece 9 oF 0 * “h,cos(b/7L — (br) *sin(bhy = Comments: 1. See Exercise 2-46 for an application of Eq. (9) to storage of explosives.Exercise 2-52 Given: A long solid cylinder with intemal heat generation, losing heat by convection to a liquid at temperature T,, Required: Steady state temperature distributions for Q, Gi) axbE-T). Assumptions: 1. One-dimensional steady conduction. 2. Constant thermal conductivity. ‘Equation (2.26) governs the temperature distribution, afaT), Q& Sfrat) = ur ale K ‘Boundary conditions: av = 0 (symmetry) ar AL = nr pT MET (® Q,” = constant. Integrating Eq. (1), aT _ 1 WG eat from Ba. 28),C, LQ Te- * ao ‘Substituting into Eq. (2b), "R -»f : = wc-r] 1 Aa : QR QR? Solving, C, = Ty» e+ Substituting in Eq. (4) = (i) a constant, Gi) Q)°11-@/RY"1, a) (2a) (2b) @) @2-52 (continued) QR? PL 2 AR T-T, = 1-(2).2) Bie 5) cian I Bij] cok ee Gi) Q)” = QTM). Substituting in Eq. (1), +C; ; from Eq. (2a), C, = 0. Integrating again, at, Ooi, at aR2 a | | seal . Gil) Q,” = axbCT-T,). Substituting in Eq. (1), df dT)\_ rt 7 EE $$) -Llasber 1); rearanging, 2d°T | aT bap cap) = r rE Pr ITH, aby) = 0 aT, oT a Let 6 = T-(T,-a/b), B* = b/k, 2 = Br; then 22470 749 .279 = o a & Table B.3 shows that Eq. (7) is Bessel’s equation of zero order (c.f. Eq. (2.49) which is a modified Bessel's equation). ‘The solution is 6 = CJ, +C,Y,@) @2-52 (continued) or T-(T,-ab) = CJ,(Br) +C,Y (Br) o 7 = -BCJ,(B) -BC,Y (Br), using Table B.S aT Die = BOO -BCY Hence, C, = 0. Substituting into Eq. (2b), ~BKC,T,(BR) = b,{(T, -a/b) +C,J,(BR)-T.] ab LC, Solving. CL = sBRy = (Bin T CBR) ‘Substituting in Eq. (9) with C, = 0, I-r,-apy = —__ HIB TBR+ BEIT BR) 0) Comments: 1, See Exercise 2-46 for issues relating to storage of explosives. If heat generation in an explosive is modeled as Q,” = a+b(T-T,), then Eq, (10) indicates that T can go to infinity when J,(BR) = -(Bk/h,)J,(BR) Hence a critical value of R = R,. can be obtained. If R > Rap an explosion can occur.Exercise 2-53 Given: A 5 KW electric heater using 1,0 mm diameter Nichrome wire to heat air to 400 K. Required: Length of wire if its temperature cannot exceed 1500 K. Assumptions: 1. One-dimensional steady conduction in the wire. 2. A minimum heat transfer coefficient of 600 W/m? K, which includes the radiation component. Equation (2.29) gives the maximum temperature in the wire in terms of it surface temperature T, a An energy balance on unit length of wire is used to relate T, to the air temperature T,, Q) Ger NA) = h,2nr (ANT, -T,); hence T, = T,+(Q,"ry/2h.) @ Combi s. (1) and (2), T, ot Hh 2] gi @) wae mt Tae 7 GP RT ML which gives the allowable Q,”. Biot number Bi = h,r\/k = (600)(0.0005)/30 = 0.01 ‘we (dor, (: +2. ‘i Q0” = 2.63x109 Wim3 If 5 kW are to be dissipated in a length of wire L, Q." nL = 5000; solving, L = 5000/(2.63x10°)(n)(0.0005) = 2.42 m ‘The resistance of the wire per unit length is the resistivity divided by cross-sectional area, R (100x10 ~¥y(2)(0.0005)? = 1.27 Qim; R = (1.27)(2.42) = 3.08 @ ‘The required current is obtained from Power = PR; I = (5000/3.08)' = 40.3 amp , and the required voltage for the power supply is E = IR, E = (40.3)(3.08) = 124 V ‘Comments: 1. The required voltage is less than the 130 V available. 2. TaacT) = Q,”17/4k = (2.63x10°)(0.0005)7(4)(30) = 5.5 K. We could have approximated T,.q, a8 T; without significant error.Exercise 2-54 Given: A thin ribbon electric heater used to boil a dielectric liquid. Required: Surface temperature in terms of average temperature determined from a résistance ‘measurement. Assumptions: 1, Bdge effects negligible to give one-dimensional steady conduction. 2, Heat transfer coefficient and ambient liquid temperature uniform. Applying the energy conservation principle to an elemental volume Ax thick gives + QAAK = Qe Ueudst, Q = -kAdT/dx; dividing by Ax and letting Ax > 0, Q OT sO integrating, o dx? k Te- = x?+Cx+C, @ Boundary conditions: x = 0, dT/dx = 0 (symmetry) Ga) x=t,-kdTidx = b¢T-T), (3b) From Eq. (3a), C, = 0. Substituting in Eg. (3b), : a” t =h}-—1?+¢,-T, Q { a GT Solving for C, and substituting in Eq. (2) gives the temperature distribution, @ The average temperature is given by T, = 4 ferro2-54 (continued) © Now electrical conductivity o is the reciprocal of the electrical resistivity p and the electrical resistance per unit length is p divided by cross-sectional area. R’ = plA, = oA, = 1(6)(2tW) An energy balance on unit length of ribbon requires (Q, )(volume) = 1702 xJength) QQ QeW() = Po(2tW) Vie pases © 4ctw? ‘Substituting in Eq. (5), T+, o ‘Comments: 1. The ribbon thickness enters the result through the Biot number, Bi = h/k.Exercise 2-55 Given: Radioactive wastes generating heat at a rate of Q,” = 3x10‘ W/m? in a 1 m diameter inner diameter spherical tank with a 1 cm thick stainless steel wall. Required: Maximum temperature in tank when cooled by air at 300 K with h, = 100 W/m? K. Assumptions: 1. One-dimensional steady conduction. 2. For the wastes, k = 2.0 Wim K, independent of temperature. ‘An energy balance on an elemental volume between radii r and 1 re+Ar requires that Qi, - Qh 4e OQ) Ame7Ar = 0 Dividing by Ar and letting Ar > 0 8 a9" With Q = Aq = ~(4nr2)kdT/dr, and assuming k constant, a (297). HOH ar\ dr. Kk at 1 QW fdr aiteeok integrating, rec Q C a, 1h GS @ dr 3k r Symmetry requires dT/dr = 0 at r= 0; hence C, = 0. Integrating again, 10," T= -2— r+ 3) SEG i @) Since the stainless steel wall is thin, we can ignore its curvature and define an overall heat transfer coefficient by w2-55 (continued) ‘Then equating the heat lost out the wastes to the heat transfer into the air, aT hE, = UGH) TI © © Bi = Unk = (92.9)(0.5)/2. T.. = 300 + 6x10 0.57 (: Heo, } boa O® Bl = 300+678 = 978 K (705°C) ‘Comments: 1. Most of the temperature rise is in the wastes (large Biot number). The precise value of the outside heat transfer coefficient, and stainless stec] wall thickness or thermal conductivity, are relatively unimportant. 2. On the other hand, an accurate k value for the wastes is needed for a reliable estimate of Trax:Exercise 2-56 Given: A 1 mm thick semiconductor laser dissipating 2 W mounted on a copper heat sink maintained at 90 K. Required: Surface temperature of chip for three different models of the dissipation process. Assumptions: 1. One-dimensional steady conduction. 2. Negligible heat transfer to the chip surface from the surroundings. 3. Constant thermal conductivity @ Since 10 pm is much less than 1 mm we can assume that the heat is released at the surface. ‘Then Eq. (1.9) applies, Q = KALYT,-T,) i or T, = T,+QLkA = 90+(2)(1x10°)/(170\(0.005x0.002) = 91.2 K Gi) Since we can assume negligible heat gain at the chip surface the steady state temperature profile is as shown and T, = T,+(Q\L/2VKA ft, = 904(2)(1x10%/2)/(170)(0.005x0.002) = 90.6 K (iii) Heat conduction with internal heat generation is analyzed for a cylindrical shell in Section 2.3.4. For a plane layer the energy conservation principle applied to an elemental volume Ax thick gives QI, FQ" AMX = Qieane Q = -KAGT/dx; dividing by Ax and letting Ax > 0, aro” a? OK @ Integrating with Q,” and k constant,2-56 (continued) @ @ Boundary conditions: x = 0: a =0 (ao parasitic heat gein) ix xeL: =T ‘Then C, = 0, and C, =T, + (Q””/2K)L?. Hence T = T,+Q, AHL?) Atx=0, Ty = T,+(Q, L7/2k) = 90+{{(2)(0.005x0.002%0.001)](0.001)°(2)(170)} = 90.6 K Comments: 1. Sometimes we are not sure of an appropriate model: then it can be useful to propose various models and examine the consequences~grating: x k 1 ) At x = 0, dT/dx = 0 because the surface is insulated, hence C, Integrating again, T @) a Substituting from Eqs. (2) and (3), QrL = oft es e-1]2-57 (continued) QL? 2k Solving, Cy = << +T,., and Eq, (3) becomes pe Se dye gt RE ba 2k B, ‘The maximum temperature is at x = 0, 1 [ath A, ‘ax = Substituting, Tye © Aso" (212) 108? + (2)(12)(0.08) | 20 800° = 165+20 = 185°C Comments: 1. Equation (3% is a "3rd kind” or "mixed" boundary condition (see Ea. (3.15)) ()Exercise 2-58 Given: Infinitely long pin fin with tip temperature equal to ambient temperature. Required: (j) Temperature distribution. Gi) Base heat flow. ‘The analysis of Section 2.4.1 gives the temperature distribution as O=T-T, =Cye + Coe which is more convenient than the hyperbolic function form for this problem. B.C. 1: T=Tp at x=0, thus Tp—T, =C)+C2 B.C. 2: T=T, at xs, thus 0=Cye™ +C2e™ F Solving, Cy = 0, C2 =Tp-Te TT _ em. g_{teP]” T TT. mene (| a=, Fo % = {Tp -TOKA Be {_________ KABOTa-T) i i i Using Eq. (2.40) for the heat lost from a fin with an insulated tip, QAR) Goes BAGBKT—T. aL. ent Comments: 1. Since 0 < tanhBL £ 1, the heat loss is always greater than for the insulated tip case. 2. For BL =3, tanhBL = 0.995: thus for BL > 3, there is no significant difference between the two cases. 3. For given B, T for the infinitely long fin is always lower than that for an insulated tip. Why then is the heat loss greater?Exercise 2-59 Given: A pin fin with tip heat loss. Required: Temperature distribution. Assumptions: 1. Heat transfer coefficient same on tip and sides. Following the analysis in Section 2.4.1, ‘TT, = Tp—Te)cosh Bx +B, sinh Bx 2 “The required second boundary condition is “Ack fh = Aghg(Plaet-Te)- Apply the left hand side of the B.C. to Eg. (1) © = Cryo sinh Bx + BB cosh Bx ~ F).-. =-ketyT.B sinh BL BB cosh BL @ ‘Apply the right hand side of the B.C. to Eq. (1) Theat =TeHTs—Te cosh BL +B; sinh BL he(T yet—Te) =he(Tp—T.)oosh BL + Byh,sinh BL Equate Eqs. 2) and (3) and rearrange, he(T~Te)oosh BL +k(Tp—T,)P sinh BL = —B, (Bk cosh BL + hesinh BL) AT p~TediMe/PKeosh BL+sinh BL ; Bi Cosh BL + (h,/BK)sinh BL + Substitute back into Eq. (1), (h,/Bk)cosh BL + sinh BL. cosh Bx ~ cosh BL + (he /fisinh BL SB cosh BL +h, /Bk)sin BL __ cosh B(L=x)Hhe/Bk)sinh BL=X) ~~ cosh BL-+(h,/Bksin BL Comments 1. In Chapter 4 it will be seen that heat transfer coefficient on the tip will ‘often be significantly different to that on the sides. See Exercise 2-66.Exercise 2-60 Given: Copper tube with aluminum sleeve and 100 pin fins/em length. Required: Hecat transfer per meter for inside fluid at 100°C, and outside fluid at 250°C. 1. Steady state. 2. Heat loss from fin tips negligibte, Assumptions: Rens In the thermal circuit for the outer surface we have two resistances in parallel, one for the fins and one for the surface between the fins. Rit Rou +Rar+(/Rins + 1/Retace)* ie 1 sees Ee Ri = aD iLhes ~ HOODOO ~ 2008S IW _ In@2/D1) _ _1n@3/20)__ dnDs/Dz) _ _In(26/23)_ _ 9 ogq096 Bako — @nj(386)(1) Dry @ny204)(1) ( “The surface area occupied by N fins is N nd? /4 = (100)(100) (n(0.0015)?/4 = 0.0177 m*. ‘The total available surface area is nD3L = (w)(0.026)(1) = 0.0817 m*. ‘Thus the prime (unfinned) area = (0.0817-0.0177) = 0.0640 m? = Asp 0.000058; Ra = 1 1 a —_——_____1___ 3) ‘The side surface area of the 54 blades is S4PL = (54)(0.1)(0.06) = 0.324 m*. Hence the heat transferred to the blades is Q=h,PLH(T.—T) = (440)(0.324)(0.251)(900-500) = 14.3 kW Comments: 1, Use FIN! to check n, Q and the guessed average blade temperature. 2. Providing adequate cooling of gas turbine blades is a major concem for the design ‘cngincer. Often coolant passages are provided inside the blades themselves.Exercise 2-62 Given: Straight rectangular aluminum alloy fins for cooling a semiconductor device. Required: The effect of choice of tip boundary conditions on calculated heat loss. Assumptions: 1. Heat transfer coefficient is constant along the fin and on both sides. k= 175 WimK 10h, $200 Wim? K Solution Using FIN]. ‘The required input is: Boundary condition = 1,2 and 3 Half-thickness, length and width = 0.0005, 0.01, 1 ‘Thermal conductivity = 175 Heat transfer coefficient = 10-200 ‘Base temperature and ambient temperature = 400, 300 x-range for plot=0.0, 0.01 FINI gives the following output for Q in watts. Tip Ihe, Wn? K Condition | 10 20 40 60 80 100 140 160 1872 264.7 3744 4585 5294 5919 700.3 748.7 199 397 789 74 1554 1929 266.2 302.1 3 209 417° «827 123.0 «162.7 2018 278.1 3154 837.1 3724 388.3 Comments: 1, For h, = 200 Win? K, Bi still too small for the infinitely long fin solution to be valid. 2. However, the error incurred by assuming an insulated tip is small (~5%). fin heat loss of Q=h,PL(Tp—). | | 3. Notice that for all three options of FINI, Tiris defined in tens of an isothermal | = [200)(2)(175}(0.001)1*(0.01) = 0.48, which isExereise 2-63 Given: Inconel X-750 straight rectangular fin, 2 mm thick. Required: Effect of tip boundary condition on estimated heat loss for various fin lengths. Assumptions: 1. One-dimensional conduction. 2. Constant thermal conductivity. Tp = 800K; T, = 300K; k= 18.8 Wink ‘h, = 300 Wim?K Solution using FINI ‘The required input is: aw @ Boundary condition = 1, 2 and 3 2400 Half thickness, length, width = 0.001, 0.006-0.020, 1 2200 ‘Thermal conductivity = 18.8 2000 Heat transfer coe'ficient = 300 1800. Base temperature and ambient temperature = 800, 300 1600. xerange for plot = 0, 0.006-0.020 FIN] gives the following output for Q in watts/meter: Tip FinJength L,mm Condition | 6 8 1 12 14 16 18 20 2377 2377 «2377-2377 2377-2377 2377-2377 1522 1822 2026 2159 2243-2296 «2328 «2347 1686 1935 2100 2206 2273 2314 2339 2354 Comments: 1. The infinitely long fin model is valid only for L > 20mm. 2. Notice that for all three options of FIN1, Nir is defined in terms of an isothermal fin heat loss of Q= h,PL(T3-T).Exercise 2-64 Given: A rectangular fin exposed to a constant radiation heat flux on one side. Required: Base temperature when ambient fluid is at 300 K. Assumptions: 1. Radiation heat loss from the fin is negligible. 2. Tip loss is negligible. ‘An additional term must be added to the energy balance used to derive Eq. (2.31) Ac ha—GAc | xeax—he PAX(T=T + Graa(P/2)Ax = 0 where (P/2) accounts for the fact that the radiative flux is incident on oiie side of te fin only. ‘The resulting differential equation is er i eA, Fn POT) + F Pana =0 @T _ MP o SE dx? kA Notice that Gaag/2h, has dimensions of temperature. ‘The solution is obtained from Eq. (2.36) by replacing T, with T, + (Gaa/2he), THTetdng/2he) _ cosh B(L-x) Ty{Te+4ne/2h.) cosh BL % 6 4 Bey =| 100 25,8, S28 = 30.000. 59K AL kt) 30)(0.005) Zhe (2)(100) coor T~(300+150)___ cosh 25.8(0.1-x) Tp-GO0H150) cosh 2.58 = : on We now require T = 400 K atx=L = 0.1m ua! 400-450 1 1.151; solving, Ty = 119K. ‘Ty450 cosh 2.58 t Comments: 1., If radiation heat loss from the fin is included, an analytical solution cannot be obtained (see Exercise 2-72).Exercise 2-65 Given: Differential equation governing the temperature variation along a pin fin. Required: (i) Solution in terms of exponential functions with coordinate origin at fin’base. ii) As in case @) but with coordinate origin at fin tip. Gi) Sotution in terms of hyperbolic functions with coordinate origin at fin tip. Assumptions: 1. Insulated tip. 2. Constant k and h,. ( Equation (2.34) is d0*/dx?-B*0 = 0, B? = h,O/kA,. ‘The solution is @ = T-T, = Cye™ + Cye** w with boundary conditions T = Ty at x = 0, éT/dx = 0 at TyT. = 0 = Solving, C, = (Lg-Tyv +e) 5 C, = (y-T +e 4); hence T-T, ef oe TyT, T+e% [+e ME Gi) The boundary conditions for Eq, (1) are now dT/dx = 0 at x = 0, T= Ty atx =L, hence eB re Be ehae He ii) Equation (2.35) is T-T, = B,sinh Bx + B,cosh Bx. ‘The boundary conditions are dT/dx = 0 at x = 0, T= Tp at x =L. BB,coshfO+BB,sinhBO = 0; B, = 0 B,sinhBL+B,coshBL ; By = (Ty-T,/coshBL coshBx ‘coshBL Comments: 1. In general, the algebra is simplified when the homogeneous boundary condition is at the origin. In this problem dT/dx = 0 is a homogeneous boundary condition. ‘Compare the algebra for case (iii) with the derivation in the text. See also Exercises 2-91 and 2-96.Exercise 2-66 Given: Pin fin with different values of h, on side and tip. Required: Temperature distribution, beat loss and efficiency. Assumptions: 1. One-dimensional heat flow. 2. Side heat transfer coefficient h,, constant along fin, Equation (2.35) gives the temperature distribution as T~-T, = B,sinhBx +B,coshBx a where B? = h.,O/kA, and B, = ', as before. ‘The tip boundary condition is aKa = A.hg(T|,-1 ~T,), which gives Khe. BB,coshBL +BB,sinhBL = -(h,/k)(B,sinh BL +B,coshBL) __ (1,71, XBsinh BL +(h, ff) coshBL) Hence, By = —~-CaBL Gh, sinh BL ie Substituting for B, and B, in Eq, (I) and rearranging T-T, _ coshB(L-x) + (h,yBk)sinh BCL—x) is Ty ~~ coshBL +, 7BEsinnBL “The heat loss is Q = FF = -kA,BB,, since cosh(0) = 1, sinh(0) = 0 (hOB) Ty-T sinh BL *(h./BK)eOshBL Rae pa: mi cosh +(h,/BK)sinh BL tee The fin efficiency is VQ... = Qh,,OL(T,-T,), ‘veal _ sinh BL + (h/Bk)cosh BL © ” BL coshBL (h,/BOsimh BL Comments: 1. Compare Eq, (3) with the temperature distribution given in Exercise 2-59. 2. Notice that in the definition of 1p Qua #8 the maximum heat loss for a fin with an insulated tip. Hence it is possible to have n> 1 for a short fin, ha 29Exercise 2-67 Given: Test rig to determine heat transfer coefficients to an array of rectangular aluminum fins that are 30 mm long, 0.3 mm thick, and at a pitch of 3 mm. 100°C, T, = 20°C and Ty, = 61.2°C. Assumptions: 1. A uniform heat transfer coefficient over the fin surface. 2. Negligible heat loss from the fin tip. Required: Heat transfer coefficient if T, Equation (2.36) gives the temperature distribution along, i 2 straight rectangular fin with negligible tip heat loss as T-T, _ coshB(L-x). 2 aL B = (h,O/kA,) ‘At the fin tip x = L, Ten 100% TT 1 coshBL yeseize 700-20 ‘cosh BL qe20 coshBL = 1.942; BL = 1.283 ° * v h2WL? — hL? KWQ) kt BL? = 1.646 = hy = 1.646kUL? = (1.646)(180)(0.00015)/(0.03)? = 49.4 Wim? K Comments: 1. Since the model assumes a constant heat transfer coefficient, this value of 49.4 Wim? K is an effective value: there is always some variation over the fin surface. 2. Chapter 4 gives correlations of convective heat transfer coefficients, from which may be gained an appreciation of how h, might vary over an array of rectangular fins.Exercise 2-68 Given: A 16 mm square chip with 16,2 mm diameter, 15 mm long aluminum pin fins. Cooling air at 25°C gives a heat transfer coefficient of 110 Wim? K. Required: Allowable power rating of chip if its operating temperature cannot exceed 75°C. Assumptions: 1, Surface temperature of chip uniform. 2. Negligible heat loss from the board 3. Negligible heat loss from fin tips. 4, Heat transfer coefficient uniform over fins and base. Equation (2.47) gives the thermal resistance of a finned surface as R= IhAn, wM where 1, is the total surface efficiency, mi = 1-(AYAV-n) ® “The efficiency of a pin fin is given by Eq, (2.42) as Te = tamhyZ, X= BL, B= h, OA)!” @ ‘The allowable power rating is the heat dissipation rate when the chip surface is at 75°C, ery P=Q=()-T)R @ B = [c110)¢m)(0.002)1(180y¢n/4)(0.002)"]"" = 34.96 BL = (34,96)(0.015) = 0.5244 = x ny = tanh0.5244/0.5244 = 0.917 Az = (16)(m)(0.002)(0.015) = 1.508x10-> m? A = 1,508x10 ® +(0,016)? -(16)(a/4)(0.002)? = 1.714x109 m? . ‘Ty, = 1 -(1.508x10 7/1.714x10 *)(1-0.917) = 0.927 R = UhAn, = U/(110)(1.714x10 7)(0.927) = 5.72 KW Q = (15-25)5.72 = 8.73 W ‘Comments: 1. An important design issue is the determination of optimal fin dimensions and spacing. The effect of fin pitch is particularly important, but cannot be considered without data for the effect of fin pitch on the heat transfer coefficient.Exercise 2-69 Giver: A pin fin Required: Efficiency with tip heat loss; condition for negligible tip loss. “The temperature distribution along the fin was obtained in Exercise 2.31 as THT, _ cosh BL—x)+(h/Bk)sinh BL—*) TT. cosh BL+(h,/BEsinh BL ee sinh B+ (he/Bcosh BL Heat loss Q=—kAe Gr x~o = (ePKAL)“(Tp-Te) ‘cosh BL+(h/Blsinh BL For an isothermal fin, Q=h,(PL+A,)(T'p~Te); introducing x = BL, = Ao/PL, _ sinh yr Coosh N= aD (cosh x+ Cxsinh D If we denote Q* as the heat loss calculated ignoring tip loss, i.e., from Eq. (2.40), then after some algebra, Qt, Ae Setanh y Q) lebytanhy Forlarge x: tanh y~1, 2-1 Q For small x; tanh x= For, of order unity a numerical demonstration that the 7 error is monotonically increasing with & is shown in the ‘Table. Notice that the erroriis of order & for small x, as a ‘expected. ‘Thus we conclude that tip loss is important rot agouas only for small values of 7 = BL, and then itis of order & Pern =AJPL. Comments: 1. In Example 2.5, 7 = BL = 0.644 and { = A_/PL = (1.2x10*)(6.8x101(0.040) = 4.4x10°; thus the error is less than 4%. 2. Notice that the fin efficiency is defined in terms of an isothermal fin heat loss of Q= (h,P+A,\Tp~Te). FINI uses an Ny defined in terms of an isothermal fin heat loss of Q=h,PL(Tp—T;), ie. as appropriate for an insulated tip (see Eq. (2.41)), in order to have a consistent definition for all three options.Exercise 2-70 Given: Heat sink assembly for mounting 36 power transistors. Required: (i) Allowable power dissipation per transistor. Gi) Required fan capacity in m?/min. @ —_Forthermal conductivity use the value for ‘Duralumin at an average temperature of 335 K; from Table A.1b, k= 178 W/m K. For one fin, a pe [22] .[600.151000010))" _ 657 nt ie KA, (178)(0.15x0.002) i BL =(16.87)(0.025) = 0.422; tanh BL =0.399 ‘Substituting in Eq. (2.40), Q = kA, B(Tp-T, )tanh BL = (178)(0.15x0.002)(16.87)(360-3 10)(0.399) = 17.97 W For 4x24 fins, Quan = (4)(24)(17.97) = 1725 W Allowable power dissipation per transistor = 1725/36 = 47.9 W. Gi) From Table A.7 for air at 1 atm and 310 K, p = 1.141 kg/m’, cy = 1005 J/kg K. The steady flow energy equation applied to the air requires that Quut = HheAT t= Quouat/CpAT = (1725)/(1005)(10)=0.1716 ke/s 1/p = (0.1716/1.141)(60) = 9.0 m? /min ‘Volume flow rate: Comments: 1. The hydraulic aspects of the design should also be investigated. The pressure drop of the airflow affects the fan design and motor rating.Exercise 2-71 Given: A straight rectangular fin of thickness 2t and length L. Required: (i) Condition for Q to be a maximum, Gi) optimal dimensions, (ii) optimal lengths for mild steel and aluminum when 2t= 1 mm and h, = 150 Wim? K. ‘Assumptions: 1. "The fin width is large compared to its thickness. 2. The tip heat loss is negligible. @ —_Themass of the fin per unit width is W = p(2txLx1) = 2ptL.. From Eq. (2.40) the heat flow from the finis Q=kA,P(Tp—T.)tanh BL, where A, = (2x1) =2r, B® ‘Substituting and eliminating L in favor of W, BP _ b(2K1) _ be KA, KDC) kt = 20ngke)* Cis Te tanh heft) W/2PH] 2 ‘We now fix W; then Eq, (1) is a relation between Q and t only. For Q to be a maximum, dQ/at must be zero £0. u,b (ha Te andl hee)" WED] — 2Qh.kt)"*(Tp—T.)oost [he /kt)" Wi 2pt)] { 3 fo | ely “CW/2p) Hence, tanh{BW/2pt] = 3 (WpMeosh-*(BW/2eH] tanh y, = 37/cosh” y, where x = BA, /2t =BW/2pt @) (i) Rearranging Eq. (2) sinh x cosh x, = 34, 0r sinh 2, = 6y. Solve by trial, x =1.419 =BA, /2t= PLshence, L/t =1.419(k/h,t)* (ii) Mild steel, k= 50 Wim K, ‘L=(1.419)(0.5x10°(50/150x0.5x10 7) = 0.0183 m (1.83 cm) ‘Aluminum, k= 174 Wim K (Duralumin alloy). L=(1.419)(0.5x109)(174/150x0.5x10-7)* = 0.0342 m (3.42 em) Comments: 1, Notice that the optimal dimensions do not depend on the density of the fin material. 2. Check that the extremum condition, Eq. (2), does indeed make Q a maximum,Exercise 2-72 Given: A long pin fin losing heat by radiation only. Given: @) Integral giving the temperature distribution. (ii) Heat loss. Assumptions: 1. Convective heat loss from fin negligible. 2. Infinitely long fin tip boundary condition applicable. 3, Transfer factor F constant along fin. cuas=8F (TA 7 Pan @) Anenesgy balance onfin clement Ax requis, 777 GAcle~4Aclaax~AgPAX=0 i sit Hence ~ fb ro- aur =0 7, Substituting q = -k ¢T/Gx, dna = OF(T*T¢), and rearranging gives or " FP ST carter; A= ax? am KA, = a0, £8 _ dp _ dp aT Letp ag! get dx aT dx p 22 = Arr); separate and integrate 1 obain ap ar Op? = (Ais T-ATETCY; or p= SE =—[raisyars-zarsr+e!] since dT/dx is everywhere negative. As x o, T+ T, and dT/dx — 0, hence Cy =-Q/S)ATS + 2ATS, and oa = ~{1@/S)ACTS-1) - 2ATSO-T,)1)* Integrating from x = Oto x, and T= Ts to T, gives t T fase eee gee cee eee -} (IQAIS\TS=13) = 2ATSCT=T)1)* bo {1(0.4) 45-3005) — 2.a(800)*(T-300)!)* where A = GFP/KA, = (5.67x10*)(0.8)(4\(0.01/(10)(w/4)(0.01)? = 1.814x10 ‘The integral can be evaluated numerically to obtain x(T), and hence T(x). Gi) ‘The heat loss from the fin can be obtained from the base heat flow as Qa Hehe Flaca =kAe {1 Q/SDATH-TE-2ATE Tp —Te)1)™ = (10) /4 (0.01)? 1(0.4)(1.814x10°)(10005-300°)-2(1.814x10-°)(300)(1000-300)1)* =208W Comments: 1. The heat loss was found without determining the temperature distribution, 2. Radiating fins are dealt with by Sparrow and Cess, pp. 180-190 (see Chapter 6 bibliography).Exercise 2-73 Given A long gas turbine blade receiving heat by convection and radiation, Required: "The temperature distribution along the blade for @) an insulated blade tip (Gi) tip and side heat transfer coefficient equal. Assumptions: 1. Re-radiation from the blade is negligible. Af the blade is at a much lower temperature than the ‘combustion gases, radiation emitted by the blade Yy | | will be much smaller than the absorbed radiation and ‘can be ignored to simplify the problem. Following the analysis of Section 24.1, an energy balance on an element of fin Ax long gives wea, +h, FE gyan —MePART ATS) + GPK =O Ae Dividing by Ax and letting Ax — 0, er MP ne? ~ fa, T+ FA, dx? KA, ‘Since Grad is @ constant,we rewrite this equation as @T _ AP Sr SE lr-|r+ dx? kA, hy which defines an “effective” ambient temperature T, = T.+4ja/he. Then the solutions previously obtained in Section 2.4.1 and Exercise 2-31 apply with T. replaced by T.. @ Insulated blade tip. From Eq, (2.36) TTetdaa/te) _ coshBi—x ToTedna/h,) cosh BL Gi) Tip and side heat transfer coefficients equal. From the result of Exercise 2-31, T+(Tetqaa/h.) _ cosh B(L—x)+(h,/Bk)sinh BOL—x) Tg—(Te#dna/tte) cosh BL+(h./Bk)sinh BL Comments: 1. If emission from the blade is included, an analytical solution cannot be obtained (see Exercise 2-72).Exercise 2-74 Given: A 0.6m long AISI 1010 steel rod welded to a furnace wall. Required: The temperature of the rod tip. Assumptions: 1. One-dimensional steady conduction. 2, The tip heat flow is negligible. 3. Constant thermal conductivity. ‘Since the rod is insulated from x= 010 x= 0.2m, the heat flow is ; 7 Theo2m-T lao Qler0 = Qln02m =e ‘The remainder of the rod is modeled as a fin with base 300 =KA.BCT] 02 m—20)tanh BL ‘The conductivity is obtrined from Table A.1b as $9 Win K at 400 K: hence pel —13)@)0.03) | gs 42 mt *[9yen/4)(0.03> | L=06-0.2=0.4m ‘Substituting in Eq. (1) 300-Th-o2, G.42)(T | 02 m~20)tanh[(5.420.4)} Solving, Thoan = = 156.1°C 1 ‘Using Eq. (2. 36), Te naa = cong Tleobm-20 _ Thosa-20 _ 1 Th-02m cosh{(5.42)(0.4} Solving, T | ,-06m = 507°C Q) Comments: 1. Use FIN] to check whether the assumption of negligible tip heat flow is justified.Exercise 2-75 Given: Jakob's rule to correct for fin tip heat loss. Required: Evaluation using FINI. From Section 2.4.5 the base heat flow depends on x = BL and Bi =hL/k. We should select test cases such that both parameters are varied independently. Since B = (h,P/KA)* does not depend ‘onLL, this can be done by varying L and h,, We choose the following parameter values: ‘Half-thickness = 0.001 m Length = 0.003, 0.01, 0.03 m ‘Width = 0.1m ‘Thermal conductivity = 125 W/m K Heat transfer coefficient = 8, 80, 800 W/m? K Base temperature and ambient temperature = 400 K, 300 K Tip size 1 = A,/P = (0.002)(0.1)/2(0.1+0.002) = 9.8x10~ m ‘Three cases for each set of parameter values will be examined: 1. "Exact", tip boundary condition No. 3 2. Jakob’s rule, L = L4,, tip boundary condition No. 2 3. Uncorrected, L = L, tip boundary condition No. 2 ‘The results for the base heat flow Q, W are summarized below. Case | Length Ihe, Wm? K m 8 80 800 1 | 0003 | 0.6494 | 6.474 | 62.84 0.00398 | 0.6493 | 6.473 | 62.80 0.003 | 04895 | 4.886 | 48.02 0.01 1.787 | 17.47 | 143.4 0.01098 | 1.787 | 17.46 | 1434 3 | oon 1.629 | 15.97 | 135.0 1 | 003 4953 | 42.11 | 1993 0.03098 | 4.953 | 42.11 | 199.3 3 | 003 4802 | 41.19 | 1989 Comments: 1. Jakob's rule works very well, even for relatively short fins of high efficiency. 2. Notice that FIN2 will give an incorrect fin mass when Jakob’s rule is used.Exercise 2-76 Given: A rectangular fin cooled by air at 20°C. Required: (i) Fin efficiency, heat loss and tip temperature for each of the three tip boundary ‘conditions given by Eqs. (2.33a,b,c) for an aluminum fin. (ii) Same as (i) for a stainless ‘steel fin. Tete Tp = 120°C T.= 20°C h, = 20 W/m? K @ Aluminum fin, k=220 W/m? K. Using FINI, Boundary | Fin Efficiency | HeatLoss | Tip Temperature Condition w °c 1 3.66 12.01 96 2 0976 3.201 116 3 0.998 3.215 116 Gi Stainless stee! fin, k= 15 Wim? K. Using FINI, Boundary | FinEfficiency | Heat Loss | Tip Temperature Condition w °c 0.956 3.137 35 0.746 2.447 &4 3 0.755 2478 2 Comments: 1. The tip temperatures were read from FINI plots. 2. Notice that the fin efficiency can be greater than unity for boundary condition 1 because Omex is based on the actual fin length L. 3, The effect of tip boundary condition is less forthe steel fin because BL is larger. 4, Boundary condition 1 is quite inappropriate for high efficiency fins.Exercise 2-77 Given: A pair of identical test alloy and brass fins. Required: (i) Thermal conductivity of alloy from measured temperatures. Gi) Error propagation analysis. Assumptions: 1. Heat transfer coefficient equal on both fins. 2. Tip heat loss negligible. Eq. (2.36) gives the tip temperature Ty, as np) bn] Ac = (W/4)(0.005)? = 1.96x10° m? ‘= (0.005) = 1.57x10? m ‘The measured temperatures for the brass fin allow h, to be determined, 642-20 1 a ioe oot + BiL=1.20; By =1.2002=60m4 100-20 ~ coshByL * Bs i By = 6 = [h,(1.57x102C111)(1.96x109))% he = 4.99 Win? K ‘Similarly for the test fin, 497-20 1, 100-20 ~ cosh B:L * Bro = 8.25 = [(4.99)(1. 57x10 7 VK(1.96x10°))*; k= 58.7 Win K BrL = 1.65; By = 1.65/0.2=8.25 mt @ To find the AT, corresponding to a Ak of +1 W/m K, it will be sufficient to find Ty, corresponding to k = 57.7 W/m K. B= [4.99)(1.57x107(57.7)(1.96x10°))* = 8.32 mt BL = 1.665 ‘T-20 1 100-20 cosh 1.665 Hence T;, should be measured to within #0.5 K. 3; TL=492K Comments: 1. Often the computer programs can be used for error propagation analysis. For example, FINI could be used to estimate the error in Nr or Q due to an uncertainty ink (or any other input parameter).Exercise 2-78 Given: Dimensions of a rectangular fin that for a given mass will give the maximum heat transfer rate. Required: Use FIN] to check this result using a 2 mm thick fin as Assumptions: 1. ‘The fin width is large compared to its thickness. ‘The given relationship for maximum heat flow is Lft=1.419(k/ht), where t= 1 mm, he = 200 W/m? K and k = 175 Wim K. This gives an. Y ‘optimal length L, L = 1.419{(175)(200)(0.001)1*(0.001) = 0.042 m Fora constant mass the product (L(t) = const = (0.042)(0.001) = 4.2x10°* m*. a base case. Solution using FINI ‘The required input is: Boundary condition: 2 Half-thickness, length and width = 1.235-0.84, 0.034-0.050, 1 ‘Thermal conductivity = 175 Heat transfer coefficient = 200 ‘Base temperature and ambient temperature = 400, 300 FINI gives the following output for Qin wans Eee mm mm mm W 1235 421022 381105421047 4201 42 1054 46 0913 42 1048 50 084 © 42-1033 Comments: 1. The analytical result is confirmed. 2, Ifno analytical result were available, the optimal length could have been easily found ‘by numerical experiment. 3, Although Q depends on (Tp—T. the optimal length does not.Exercise 2-79 Given: A stainless steel straight rectangular fin, Required: Fin effectiveness e as a function of fin length, Assumptions: 1. Tip heat loss negligible. ‘The required input to FINI in ST units is Boundary condition = 2 Half-thickness, length and width: 0,001, 0.002-0.05, 0.01 ‘Thermal conductivity = 15 Win K ‘Heat transfer coefficient = 25 W/m? K Base temperature and ambient temperature = 320, 300 (arbitrary) x-range for plot = 0.0, 0.002-0.05 ‘The heat loss without a fin is h,A,(Tg—T_) = (25)(0.01)(0.002)(20) = 0.01 W. The fin effectiveness is then Q/0.01. Results obtained from FIN1 are tabulated below. L(m) | 0.002 0.005 0.01 002 0.05 n% 0.997 0.984 0.938 0.798 0.437 Qx10? w | 2394 5.902 1126 19.15 26.23 & 239 «590 3 (192262 Comments: 1. The fin effectiveness & is always greater than unity. 2, Fin effectiveness is related to fin efficiency as e = nr(PL/A.) 3. Unless e, is sufficiently high, fins will not be cost effective, 4, FIN2, item 1, assumes W >> t, and should not be used for this exercise.Exercise 2-80 Given: A straight rectangular fin, k = 40 W/m K, h, = 10 W/m? K. Required: (j) Dimensions of the minimum mass fin that dissipates 100 W/m for T,-T, = 40 K. ii) Parametric study to find fin profiles of smaller aspect ratio that also dissipate 100 ‘W/m and do not have a much larger mass. Assumptions: 1. Heat transfer coefficient constant along fin. 2. Fin tp is insulated. (i) The heat dissipation by a straight rectangular fin of thickness 2t is obtained from Eq. (2.40) as Q = KA,B(T, -TJtanhBL W/m wo for A, = (20(1) and B = (hJkt)!?_ From Exercise 2-71 the dimensions that give a minimum mass are £ = 1.4190/h 0)? Q Substituting in Eq. (1), 100 = (40)(21)(10/40t)"(40)tanh [(10/408)!7L} (100/1600) = t"?tanh (Ly2t ¥) From Eq. (2), Lit!) = 1.419@40/10)"? = (1.419)(2) .004938 m (4.938 mm) L = 0.1994 m (199.4 mm) Solving, ii) We now use FIN2, Choosing values of L and adjusting t to give Q = 100 W/m, the hence following table is constructed. L t | WWen] on | Let mm | mm 1994 | 4938 | 1.00 | 0.625 | 202 170 6.23 1.08 | 0.735 | 13.6 160 12 117 | 0.781 | 114 150 90 137 | 0833 | 383 140 134 152 | 0.893 | 52 130 | 350 462 | 0.961 | 192-80 (continued) ‘The table shows clearly the relation between fin mass and aspect ratio, Such data can be used in design evaluations that consider the trade-off between manufacturing costs and penalties associated with increased mass. Comments: 1. Parametric studies are an essential element of good design practice. Results such as those above allow the designer to identify desirable ranges of parameter values. For example, it is clearly undesirable to have L decrease below about 150 mm since the mass penalty increases too rapidly. 2, In obtaining these results the student should have discemed why the weight penalty is large at smaller values of L-the fin efficiency approaches unity so that adding extra thickness to the fin has little effect on the heat dissipation (as 9 —> 1, Q becomes independent of t. 3. By allowing a tolerance on Q of 2% the results obtained by Zhang and Chung are a little different to those presented here~but the picture is fuzey anyway!Exercise 2-81 Given: A 10 cm square plate at 70°C with 25 mm long, 0.3 mm thick rectangular aluminum fins at a3 mm pitch. Required: Heat dissipation into air at 20°C with h, = 30 Wim? K. Assumptions: 1. Heat transfer coefficient is uniform over the fins and plate. 2. The thermal resistance of the plate is negligible 3. Heat loss through fin tips negligible. Equation (2.47) gives the thermal resistance of a finned es ow R= ot a a0 fe where , = 1-(A/A)(I-n) 2 at is the total surface efficiency. Equation (2.42) gives the efficiency of a straight rectangular fin, ny = tangy; x = BL; B = (h/t)? @) where the fin thickness is 2t. B = [30/(180)(0.00015)]"? = 33.3 mt BL = (33.3)(0.025) = 0.833 = % ‘TN, = tanh0.833/0.833 = 0.819 For a 0.1 m square plate, Ag = (0.1)(0.025)(2)(0.1/0.003) = 0.167 m? A = A,+[0.1-(0.0003)(0.1)(0.1/0,003)] = 0.167 +0.009 = 0.176 m? Substituting in Eq. (2), 7, = 1 -(0.167/0.176)(1-0.819) = 0.828 and then from Eq. (1) R = 1/G0)(0.176)(0.828) = 0.229 K/W Q = ATR = (70-20)/0.229 = 219 W Comments: 1. Without the fins the allowable dissipation rate is only (30)(0.1)(0.1)(50) = 15 W. 2. Use FINI or FIN2 to check ny.Exercise 2-82 Given: Semiconductor device dissipating 5 W on a 2 cm square aluminum plate: underside has 15 mm long, 0.2 mm thick rectangular fins at a pitch of 2 mm, Required: Temperature of semiconductor base if cooling air is at 25°C and h, = 30 Win? K. ‘Assumptions: 1. One-dimensional heat conduction across the plate. 2. Uniform heat transfer coefficient over the fin surfaces and plate. 3. ‘Thermal resistance of diamond spreader is negligible. ‘The very high thermal conductivity ofthe diamond layer (1300 W/m K) should ensure that the top surface of the | aluminum plate is at a nearly uniform temperature, and hence heat flow across the plate can be modeled as one~ dimensional giving the thermal circuit shown. Equation (2.47) gives the thermal resistance of a finned surface as R = thAny ny = 1-(AYAMI-nD Equation (2.42) gives the efficiency of a straight rectangular fin, n, = tanhyyy; % = BL; B = (hykt)!” fora fin thickness of 21. B = [(28)175)(0.0001)]'” = 40.0 m=; BL = (40)(0.015) = 0.60 Ty = tanh0.60/0.60 = 0.895 ‘Ay = (0.02)(0.015)(2)(0.02/0.002) = 6x10 m? A = 6x10? +((0.02)? -(0.02)(0.0002)(0.02/0.002)] = 6.36x10> m? n, = 1-Gx10 */6.36x10 )(1-0.895) = 0.901 R, = Uh,An, = 1(28)(6.36x10*)(0.901) = 6.23 KW Ryne ~ L’kApite = (0:002)/(175)(0.02)? = 2.86x10 > K/W ER = R,+Ryge = 6234003 = 6.26 KIW Rg is nezligible) T, = T,+QER = 25+(5)(6.26) = 56.3°C Comments: 1. How thick should be the diamond spreader to ensure one-dimensional heat flow across the plate (and hence fully utilize the fins)? A three-dimensional heat conduction model is required to answer this important design question. Computer progtams are commercially available for such tasks.Exercise 2-83 Given: Test data for a finned surface with a suspected contact resistance at the base of the fins. Rectangular straight aluminum fins, 20 mm long, 0.3 mm thick. Required: Magnitude of contact resistance if present. Assumptions: 1, One-dimensional conduction through the duct wall. 2, Heat loss from the fin tips and ends is negligible. Equation (2.36) gives the temperature distribution along a fin of constant cross-sectional area and negligible tip heat loss as T-T, _ coshB(L-x) _ | be Ty ‘coshBL 6 KA, o : ToT i At the fin tip, x=L and —_¢ = 1 _ T,-T, coshpL @ Neglecting the heat losses from the ends of fins W wide and 2t thick, O/A, = 2W/W(2t) = It B = [21/(185)(0.00015)]"" = 27.5m™;BL = (275)0.02)= 0.550; coshBL = 1.155 From Eq. (2), Ty = T,+(T-T,JeoshBL = 23.2 +(41.6-23,2)(1.155) = 44.5°C Since the duct wall inner temperature is 100°C and the temperature drop across the 1 mm duct wall thickness will be very small, there is definitely a significant contact resistance. Referring to the thermal circuit and Eq. (2.8) | 1000-445 EAA Way ~O. oy L 1 7 ‘ 1 180, _ and from Eq. (2.40) QW = K(2)B(T,-T,tanhBL = (180)(0.0003)(27.5)(44.5-23.2)tanh0.550 = 15.8 Wim ‘Substituting in Eq. (3) and solving, h, = 959 W/m? K. Comments: 1. This value of his very small (see Table 2.1). It should be easy to obtain a better thermal contact between the fins and the duct.Exercise 2-84 Given: A thin walled stainless steel themnometer pocket projecting into a gas stream at 320°C. Required: Length of pocket for the error in the thermometer reading to be less than 2°C. Assumptions: 1. Temperature variation across the stee] wall is small compared to the variation along the pocket. 2. Thermometer reads the temperature at the tip of the pocket. 3. Negligible conduction along thermometer. ‘We will model the pocket wall as a straight rectangular fin of thickness 0.7 mm, with heat gain from one side only. Ty =240°C y : 7. =320°C th 318°C $= be 20Wn? K eee eee 5 Win K ty =0.7mm ‘The effective value of B is oP beotD Noo i B= ra =i9p, 7 ke 7 (30Y(15)(0.0007) = 2857 ‘The temperature of the tip of the pocket (located at x = L) is given by Eq. 2.36) as, ite eed Tp-Te (240-320) coshBL Solving, cosh BL = 40; BL=4.38, L=(4.38)/B = (4.382857) = 0.082 m Comments: 1. A pocket length of greater than 8.2 cm is required for an error of less than 2°C. 2. This exercise shows the importance of carefully designing temperature probes. 3. Check that itis valid to neglect conduction along the thermometer.Exercise 2-85 Given: A wire with internal heat generation. Required: Location of maximum temperature for given end temperatures. Assumptions: 1. The wire is thin enough for a fin model to be valid. ‘The volumetric heat source Q, W/m? is related to the ‘wire resistance R’ {2/m by the relation Qa.=PR’ a Proceeding as in the derivation of Eq, (2.31),the energy ‘ oe balance on an element of wire Ax long becomes GAs lx-GAs Inve hePARCT=T) + Oy AeA = res which leads to the differential equation x0, B.C 1: T=T; at x=0, thus T)~T, - ra =CHO BC.2: Tp at x=L, thus T2-T, = CyeFl + Cpe FE Hence solve for C; and Cp. For the maximum temperature, a Be = CAPE — Cafe 1,& he Comments: 1. The algebraic result for C; and C2 is cumbersome. It is advisable to substi- tute numerical values for the problem of interest before solving for Cy and C, (unless a large number of similar problems need to be calculated).Exercise 2-86 Given: A 2 mm thick brass plate with coolant tubes spaced 10 cm apart, Required: ‘Temperature of hottest spot on brass plate for a coolant temperature of 315 K. Assumptions: 1. Heat flow across the asbestos layer is 1-D. 2. ‘The brass plate is thin enough for the temperature variation across the plate to be negligible. 3, Negligible thermal resistance of solder and tube wall, ‘The brass plate is treated as a fin: since dT/d (0 midway between the coolant tubes, the fin length L = 5 em: we need to find its tip tem- perature T(x=L) fora base temperature of Ty =315K. ‘600K —o.002m Heat is transferred across the asbestos layer ata rate of q = (ky /tp)(600—T) where ka is the asbestos conductivity, tz is its thickness, and T is temperature of the brass, thus (ko/tz) is analogous to a convective heat wansfer coefficient, which acts on a perimeter of P = 1 per unit ‘width of heat barrier. In our fin analysis, A, =ty per unit width, where tis the thickness of the brass plate. (&aftaX(0))" _ [ 0.16/0.015)]"* _ 693 mt [e4 kit tox - {gugaual tefet pL=0347 From Eq. (2.36) the tip temperature at x = L is i Tp) _ 600-315) _ 283. is ToT Soa BL 7 coanoze7 "O~ Toet 28 ESOC) Comments: 1. Notice how a seemingly complex problem is reduced to a simple model problem. 2. The asbestos used here is denser than the asbestos paper insulation listed in Table A.3, and hence has a higher thermal conductivity.Exercise 2-87 Given: A pressure transducer connected to a fumace by a copper pigtail. Required: Length of tube for transducer to be at 340 K when the furnace is at 1000 K. Assumptions: 1. Fin model valid for tube. From Table A.1,k ~370 W/m K for copper at 700K. 1,=1000 Fora thin walled tube P/A, ~ xD/nDt = I/t where tis the wall thickness. Thus the fin parameter B is 1K Fe a eeepc Ag it) ~ [G70)05x10%) : Eq, 2.36) gives the temperature distfbuion dlongihe 3 pigtail if end losses are ignored, fi T-T, _ cosh B(L-x) 2 ® Le? Tp-Te cosh BL ForT = 340K atx =L, and T, = 300 K, Ty = 1000K: 340-300 1 i os “Too0-200 ~ coaniacmzy + cosnl27L) = 17.5 Forx lage, cosh x~ 3 e ie 2 Solving L = 0.280 m (28 cm). 127 175, Comments: 1. The given value of h is a rough estimate. Itis difficult to obtain a precise value for this configuration (see Chapter 4). 2. A stainless steel tube would be more appropriate because of its low thermal con- ductivity.Exercise 2-88 Given: A chip connected to a circuit board, as shown in Figure 2.17. Required: Allowable rating for chip if 60% of power must be dissipated by the board. Assumptions: 1. Steady state. 2, Heat loss from surface of brass lead negligible. 3. ‘The copper conductor and circuit board are thin enough for afin type analysis to be valid. ‘The brass lead and the board are in series. The resistance of the lead is simply L/kA, oe (0.005) a Rise= Gy eM H ATT ete ‘The resistance of the board is alittle more difficult to estimate, but it can be approximated as two fin problems. Firstly, the copper conductor can be considered as a fin 10 em Jong, and the board on either side of of the conductor can be considered as a fin 5 mm long, dissipating heat from the ‘conductor, For the board of width W = 0.01 m, AP = (5)(2W) = 10 W, kA = (0.2)(Wt) = (0.2)(0.0015)W = 0.0003 W % 10W = = 182.5 8 [aca Ww 1¢= BL = (182,5)(0.005) = 0.913; tanh x = 0.723: For the copper conductor, the effective perimeter takes into account the fact that the board surface is only 79% effective, he =5, P = (2)(0.002) + 4(0.005(0.79) = 0.0198 KA, = (386)(0.002)(0.00075) = 5.79x10~ (neglecting the contribution of the board). % (5x0.0198)| = (13.1)(0.1) = 1.31; tanh x = 0.864 Sapstet | 73+ X= 13.10) = 1.315 tanh x= 0: 1 1 Bane aceon ee eHEE 1 Riout = GP/phanh yg ~ GOODEN IS IKO.B6H 3EW/KT ‘Then the allowable dissipation of the chip Q is given by AT _ 350-310 IR“ qsniss O74 mW- 0.6Q= Comments: 1. Notice how a complex system is broken into simpler model components by idemtifying the major heat flow paths.Exercise 2-89 Given: Copper-constantan thermocouple with bare, separated leads. Required: Exror due to conduction along the leads for a 5 cm protrusion and bh = 100 W/m? K. Assumptions: 1. Fin model appropriate. ‘From the analysis of Section 2.4.1, the temperature dis- tributions in each wire are: copper: T-T, = Cye™ + Coe constantan: T-T, = Cye™ + Cye™* ‘The four constants are found from the four boundary conditions. <0: Ty-T, =20=C,+C3 = Ca4Ce a2 x=L: Here the temperature and heat flow must be continuous, Bl ® yeh s Coe Ph = Ce + Ce! KpBr(Cie Coe) = kage) @ From Table A.1,k = 401 and 23 W/m K for pure copper and constantan (45% Ni) respectively. ‘Also, for circular cross-section, P/A; (ndy(nd?/4) = 4/4, Hence Bi Fal ‘(petal = 44.2 mr}; ByL = (44.20.05) =2.21 id (401(0.510x10 e-[B]- ws) = 185 mr“; ByL = (185(0.05) = 9.25 Equations (1) through (4) become Cr+ Ca#Cy=20 9.12C1+0.110C2 = 1.040x10°C3+9.61x10°°C, 1.772x104(9.12C}—0.110C2) = 4250(1.040x10'Cs—9.61x105C,) Solving, C; = 0.39, C, = 19.61 “The error is the junction temperature minus the steam temperature, Ty-Te = Cre + Coe = (0.39)(9.12) + (19.61)(0.110) = 5.7K ‘Comments: 1. Most often, thermocouple leads are not separated, and the analysis of Example 2.6 is then appropriate.Exercise 2-90 Given; Electrically heated copper rod, L = 30 cm long, in air at T, = 20°C. Required: Current if midpoint temperature is 50°C (j) neglecting radiation, (ji) including, radiation. Assumptions: 1. Steady state. T 2. Fin type analysis valid. soy} 20° Y (i) The fin analysis of Section 2.4.1 can be extended to include volumetric heating Q, , cAelx + Qy AcAK — GeAe xsd —NePAXCT-T.) =O which gives. kAg pa =hPOMT,) +07 Ac=0 oee er ae or, = PTH e+Qy Ac/heP)]=0 x ‘Suitable boundary conditions are: x = 0, T= Tg; x = L/2, dT/dx = 0 ‘The mathematical problem is identical to that of Section 2.4.1, so Eq, (2.36) applies with T. replaced by T. + Qy Ac/heP, and L by L/2, hee) @ P= (w)(0.002) = 0.00628 m B= 155 bx Ac = (80.001) = 3.14x10* m? B =1247 mt barf k= 386 W/m K; h, = 30 Wan? K BL/2 = 187 Tp = 20°C. Te = 20°C Qy AchheP = 1.67105 Qy 50-(20+1.67x10° Q,) 1 20-420+1.67R10 Q,) cosh 1.87 Power P= Qy LAg = 2.42 W Resistance R = (resistivity)(L/A,) = (1.72x10°Q m)(0.3 m/3.14x10% m?) = 1.64x10°Q Current I= (P/R)" = (2.42/1,64x10)* =38.4 A Gi) Linearize the radiation transfer in order to obtain an analytical solution; using Eq, (1.19) h, = 40eT 3, = (4(5.67x10*)(0.8)(308)° = 5.3 W/m? K. ‘Since the convection and radiation are both to the same “sink” temperature of 20°C, we can sim- ply add h, andh,,h=h, +h, =30+5.3 =35.3 W/m? K. Eq, (1) applies with h replaced by hy B? = 183; BL/2=2.03: cosh BL/2 = 387; Q, Ac/hP = LA16xI0 Qy Q, =2.86x10° Wim’; 1= 40.5 A Qy =2.570x108 Win? Comments: 1. By neglecting radiation the allowable current estimate is 5.2% too low.Exercise 2-91 Given: 1 cm O.D. tube, extemal radiation heating on one side, water cooled internally. Required: i) Wall temperature distribution, ii) effect of a volume heat source. Assumptions: 1, Steady state. 2. Temperature variation across tube wall negligible. A fin type analysis is appropriate. Symmetry allows us t0 %0 consider the half tube 0 < @ < 180°, which must be split into ‘two separate fin problems 0<§<90°: Measure x; around circumference to give an approximate fin length of L = xD/4; A, = tH for tube length Hi. Following the analysis of Section 2.4.1, a, Ht] 5, + dasHAny ~ 44, Ht yes, —RgHAxy(T)-Te)=0 reo" ich vives 2TL_ Me Sut whien Bes “pe TO #01 _ p96, <0; = T1fr+ 32! a or Gp 8 = 01 8 = TTT ne) 90° << 180°: Ge, Ht lx, = Ge, Ht 405, ~ NeHAXp(T2-Te) =O @ Ac rts P= @ which gives od B°6, =0, @ = Eqs. (1) and (2) are two coupled second order ordinary differential equations requiring four boun- dary conditions: dy 8 : gyre HO GE =O bysymmenry ® L, T,=T) or +0) <0 for continuity of temperature @ a8, dO “%, OF Gr =— Gye foreontimity of heat Mux © ‘The solutions of Eqs. (1) and (2) are2-91 (continued) = Asinh Bx; + Boosh Bx; 82 =' oa Bx2 + Deosh Bro, Ge 7 ABrosh Bx + BPs rs Ge Fer = Peas is + Bi a Since cosh(0) = 1, the symmetry en (3) require A = C= 0. Substituting in Eq. (5), BBsinh BL =—DBsinh BL, or B = -D. Substituting in Eq. (4), Grad 1 Get a _ Sad = “Fer + Beosh BL = Deosh BL, or = 5 age =-B ‘The final temperature distributions are nd _ Gna cosh Bxy Gat, Sash Bra 0< $< 90%: Ty =T, + — 5 90° << 180°: Tz = ¢ : he 2h, coshBL ’ 9 2=Te+ on cosh BL B= (he /kt)" = (1000/(20\1x10))* = 223.6 m™; T, = 300K xD/4 = ()(9.0x10")/4 = 7.07x10" m, taking D along mid-plane of wall 2.531 0< 990": Ty = 1300- 197.5 cosh Bx, 180° << 300+ 197.5 cosh xa Gi) If there is neutron absorption Q, W/m? in the wall, each elemental energy balance has an additional term + Q, HtAx, and the differential equations remain the same with = ayn, et St, and @2 = tet, ‘Thus in the previous solutions T, should be replaced by T, + Qy t/he, where Q0 Uh, = (50x10°)(1x10)71000 = 50K Ty = 1350-1975 cosh Bx, ‘Ty = 350+ 197.5 cosh Bro Comments: 1. Notice the use of hyperbolic rather than exponential functions to simplify the algebra. 2. Once again the fin model allows a two-dimensional problem to be approximated by a much simpler one-dimensional problem.Exercise 2-92 Given: A Slat plate solar collector. Required: Outlet water temperature on a clear windy day. Assumptions: 1. Fin-type model of plate is valid. 2. Back face of plate perfectly insulated. 3, Reradiation can be approximated as h(T—T.) with b, constant. 4, The water-side and tube-wall resistances are small: Ter Qasr ANEhe) age Chet MT=T) ° z 4 ‘The steady flow energy equation, Eq, (1.4), sah = Q, is applied to an element of absorber 2L. wide and Az long to give thggc(Te | sede—Te |.) = 2042, ‘where the subscript C refers to the water as a coolant stream, Q is the base heat flow per unit Jength of the plate 0 B=5.09 m7, BL =(5.09)(0.05)=0.255 350-20 1 Ts-20 cosh(0.255) Equation (2.40) gives the base heat flow as Q=kA.B(T~T.)tanh BL (386)(1.257x10°5)(5.09)(361-20)tanh(0.255) = 2.11 W 3 hence Ty =361°C Comments: 1. Use of copper ensures that the tip temperature is very little lower than the base temperature,Exercise 2-94 Giver: Cooled skin panel for a hypersonic aircraft. Required: Maximum skin temperature for given conditions. Assumptions: 1. ‘The walls can be treated as fins. 2. The comer regions are isothermal. ‘The sketch shows one half a cooling passage, divided into three "fins", For convenience we will take the fluid tempera- ture Ty = 0 for the analysis (orlet @=T-T,). ‘) Fint 7 a ‘i CORNER 2 ane #T_ &. Be ye 7 % oe iy By = (hefty) Sa FINS t BC's. x=0: kdT/dx=qi,x=W/2: kdTiix=0 | I i qucosh[B,(W/2—x)] as x Hence Ty = - a # ne =o ence TS Biksinh(B; W2)* hy a-olZ i Gi) Fin2 | ' ft a t £F _ par = 0; Bp = (hulk) hex FINT t ay? i f BC's, y=0: -kdTMy= qo: y=H: -kdTidy=q, CORNER af eae ae ean ty we =~ BaksinbeB W2) Gv) Comer 1. An energy balance gives Gslo + Git = Geta and for an isothermal comer T}(x=0) = T2(y=0), hence _-sncosh(BiW2) , a _ gzcosh(BoH)-ds Biksinh(B; W/2) " h, Boksinh(BpH) () Comer 2. An energy balance gives b= its and for an isothermal comer T2(y=H) = T;(x=0). Hence2-94 (continued) se-aycosh(PaHD__ qos W2) Byksinh(B.H) —_Bsksinh(B; W/2) (vi) From the comer conditions we obtain ABB. 4 ABA, PTET _ABicABs | te BY ABAB, Wy? Big cob, wE)+ coma Ao=- oo, Ags Bk Brksinh(B)H) re came cree LY ds rth |— 2— coth(B;W/2), Bs =—- 7——~—~—— = Fpesan(fay | PT Fe MID By COMPWP), Bo =— Fey h=h=h=4x104m; W=H=2x10%m ‘h, = 6000 Wim? K ; k= 11,7 Wim K; g,= 10° Wim? % |_| = 1132 Poeaes (a : = foth(1132x0.002/2) | coth(1132x0.002) ae IRL cuisayti.ny hx" GsDa sinner TsavO.e) ESIC A= — [1 _ ootnt1132x0.002)] , 195 = As= | S000 G13.) x10 = 8.951 Ee 5 Bi = (qaaay1T Tpsinh(I32x0.002) ~ 58710"2-94 (continued) ___coth(1132x0,002), _coth(1132x0.002/2) __, 4 Ba=-"" a1301.7) agai 71710 10° Ba=— GiramyGAt int 13250.002) 587 _ (1,587?x10" + 8.951x1.702x104 _ 53030 Wim? = 7 i’ =" 587810) + (1.702104? = 10° + 53930 = 153930 Wim? 951x1587x10-5 + 1.587x1.702x104 “1587x1057 + (1.702x104? = 14350 Wim? qa = 14350 Wim? Intuitively we know Tyyax is located in FINI and also oT sinh{By(W/2—¥)] dx k — sinh(B, WP2) >0 ‘Thus Tynax Occurs at x= W/2, where 53930 (1132)(11.7)sinh(1132x0.002/2) so 7 =293+167= 19.6 T, T,=19.6 + 100 = 119.6 K (since we put Ty = 0 for convenience in the analysis) Comments: 1. This exercise demonstrates the versatility of fin-type analyses. 2, The guessed value fork = 11.7 Wim K corresponds to T = 300 K, which is too high. A value at 100 K would be more appropriate, but Table A.1b does not go below 200 K. 3, Ift is large, qy can be negative and Tyux is then located at x= 0.Exercise 2-95 Given: An exhaust stack thermocouple inserted into a stainless steel well, Required: The exhaust gas temperature. Assumptions: 1, Temperature variation across the steel is small compared to the variation along the well. 2, Negligible radiation heat transfer. hh, = 65 Wim? K k= 16 Wim K (Table A.1b) Tp = 178°C T= 221°C ‘This is afin type problem since the temperature variation across the stec! is small compared to ‘the variation along the well. The effective value of B? is B= h,PikA, = hyxD/ker(D-1)t BP = (65)5x10°/(16)(4.3x10°)(0.7x10) = 6748, BL = (6748)"(0.20) = 16.4 ‘The temperature of the thermocouple junction located at x = L is given by Eq. (2.36) as Tee eect Tp-T, coshBL — cosh16.4 ~ 6.63x10° Hence Ty, = Te = 221°C. Comments: 1. For this thermocouple installation, BL is very large and the gas temperature is measured very accurately.Exercise 2-96 Given: A2.cm O.D. tube, external radiation heating on one side, water cooled intemally. Required: Wall temperature distribution. Assumptions: 1. Temperature variation across tube wall negligible. ‘Since the temperature varies along the wall only, a fin type analysis is appropriate. Symmeuy allows us to consider the half-tube 0 < 6 < 180°, which must be split into two separate fin prob- Jems. For 0<¢ < 90°; an energy balance on an elemental volume Ax; long leads to the ifferential equation eT, a crn + cos y= O5let = T=, &, her? aR ewan “gt Ot aos: BME = ® ‘Similarly, for 0° < 2 < 90°: fe @ Four boundary conditions are required. By symmetry Gaby 1 =0: d8,/4; =0:; G2 =0: d8p/d =0 Continuity of temperanure and heat flux require 45) O1=¢r=R2: O= 12 401/49) = 482/42 ‘The solutions of Eqs. (1) and (2) are ®, = Asinh Bo, + B cosh Bd, + (0/(1+B?))cos $1; 62 = Csinh o> + Dosh Boa From Eqs. (3.4.5): 6; B=D; B=D=(oi(1+f")/2Bsinh(Br/2) ‘The final temperature distribution is . Ty 27, + —— oe Bs_, ens, 0<9<90": T)=Te+ Bei i co cosh | 90° 8: h=T.+ SO 180%: Th Te * rsp Banh) (H.R? Aay”* = [(800)(9.5x10)7/(18)(1x10)]* = 2.00 QoR7 kt = (10°}.5x10)7/(18)(1x10) = 501.4; sinh(Br/2) = sinh r= 11.55 Hence, for 0<@ < 90°: T; = 320+2.17cosh2,+100.3cos; and 90° < @ < 180°: Ty = 32042.17eosh2¢2. Comments: 1. Notice the use of hyperbolic rather than exponential functions to simplify the algebra.Exercise 2-97 Given: ‘Test cylinder for measuring local convective heat transfer coefficients. Required: ()) Formula forh,(8) which accounts for circumferential conduction. (i) Evaluation with given data forT,(®). (tii) Evaluation of design. Assumptions: 1. A “fin" type model is appropriate. fi eras @ The temperature variation across the wall of the a cylinder is negligible compared to the circumferential we variation (there is negligible heat transfer from the inner aa < surface). Thus a fin type model can be used. An energy hf i balance on an element of wall ds long and W wide “tty Hot W 1+ QIAN WAs + (at +kot2)W oa leeae be WAS(T,-T,) = 0 ‘where t; and tz are the thicknesses of the ribbon and Teflon, respectively, and (Q/A) the power ‘input per unit outer surface area of the heater. Dividing by Was and letting ds — 0 gives tT 9 1 @T that) SF + 2 RTT) =O. or Tt) = 2 + atetkae) Je Se where ds =Rd@. Gi) Using simple two-point formulas for derivatives, a, 657-659 _ ST) | 644-657 _ 49 1 =~ “79 0.02 Kideg: “Gg !1is* 10 0.13 Kideg SE ine = RTEEOO) «9.011 Kitey? =-36.1 Kad? A @r ___36.1_ 2 a9? 7 on? AN Ke he= 5900+{(15)(2.54x10)+(0.002)(0.38)]-3.61x10°) _ 5900-412 i 65.7-25 © 657-25 Comments: 1. The "correction" for circumferential conduction is 7%. 2. For an accurate result, temperatures should be measured at small intervals, and a more sophisticated numerical differentiation formula used. 3. The oylinder was designed to have a low value of 21k. In particular, the heater ribbon is very thin and has a low thermal conductivity. 4, Circumferential conduction in a 3 mm thick brass tube would be so large that there would be no measurable variation of T, around the cylinder. An average heat ‘transfer coefficient for an isothermal wall will be measured.Exercise 2-98 Given: A 4 mm diameter, 25 cm long aluminum alloy rod, with 10 W of heat supplied over the central 5 cm length. ‘Required: End temperature if exposed rod cooled by air at 300 K giving h, = 50 Wim? K. Assumptions: 1. Fin-type analysis appropriate. 2, The insulated-tip fin boundary condition is adequate. 3. For the alloy, k = 190 W/m K. % ‘The heat flow into each 10 cm long exposed %. $ WEE portion is Q = 5 W. The base temperature Ty can be found from Eq, (2.40) as =T,+_2 Ta =T. KA PtanhBL A, = (wi4)(0.004)? = 1.257x10° m?; © = (m)(0.004) = 1.257x10? m B= (H,O/A,)!? = (S0x1.257x107/190x1.257x10°)" = 16.22 m' BL = (16.22)(0.1) = 1.622; tanh BL = 0.9249 Ty = 300 + (5)/(190)(1.257x10)(16.22)(0.9249) = 439.6 K Now we can find T,, from Eq, (2.36), T1, we-_1_-__1 _ _380 T,T, coshpL — ‘cosh622 T,, = 300 + (439.6-300)(0.380) = 353.1 K Comments: 1. Although the rod is not a cooling fin, a fin-type analysis is valid because the rod diameter is small compared to its length—the temperature variation across the rod is small compared to that along the rod. 2, Can you determine the temperature variation along the rod inside the insulation?Exercise 2-99 Given: Hlectrically heated wire, 1 mm diameter, 20 cm long, located in an airflow at 290 K. Required: Current when mid-point temperature is 400 K and ends are at 300 K. Assumptions: 1. Steady state. 2. Fin-type model. ‘The fin analysis of Section 2.4.1 can be extended to include volumetric heating Q,”, : w= 280K Ach *Qy AAR“ OAc heats =(h, +h, Ax(T-T,) = 0 ite j | 4 which gives KA, st ~(h,thO(T-T,)+Q,"A, =O Fax \ or, s BIT-(T,+Q, ASH, *H IO = 0 Suitable boundary conditions are: x = 0, T= Ty; x =L/2: dT/dx = 0. Eq. (2.36) applies with T. replaced by T,+ Q, A/C, +h,)®, and L by L/2, (1, Q) AL Ane) _ 1 Ty (T,+Q, Ashe) CoshiBL/2) © =(n)(0.001 141x10 ?m; A, =n(0.005)"=7.85x10 7m?; k=386W/mK; h, = 40eT,2 = (4)(5.67x10°)(0.8)(350") = 7.8 Wim? K QZ AM.) = (Q,’\7.85x10-727.8).141x10 9) = 8:99x10 Q," B = [dh +h )O/KA,}! = [(20+7,8)(3.141x10 7VG86)(7.85x10 yy? = 17.0 BLI2 = (17.0)(0.2)/2 = 1.70 400-(290+8.99x10 *Q.") 1 = —! __ = 0.3536; solving: Q.” = 1.832x107 W/m? 300-(290+8.99x10 *Q,") cosh 1.70 Power P= Q, LA, = (1.832x107)(0.20)(7.85x10 7) = 2.88 W Resistance R = (0.2)(2.2x107) = 4.4x10° Current I: (P/R)'? = (2.88/4.4x109)'? = 25.6 A Comments: 1, It is essential to recognize that this is a fin-type problem.Exercise 2-100 Given: A 3 em O.D., 1.25 mm wall thickness, copper tbe spanning a 20 cm wide wind tunnel, with 25.W power input to a 10 em long heater. Required: (j) Fin type analysis to determine temperature distribution along tube. Gi) AT between midplane and end of heater. Assumptions: 1. Negligible temperature variation across tube wall. 2. Negligible heat loss through inside tube wall, and tube ends. ky le ‘Symmetry allows us to consider one half bon an of the tube. An energy balance on an element of the heated portion Ax, long is Tone GAL, Ada ate, “BOAKYT, ~T)) (oS +q,OAx, = 0 where the electrical heat generated by the heater has been modeled as a volume heat source distributed in the copper tube. ‘The resulting differential equation is aT, ho! 1_ Sella, 7, +28 |] - 0 ax? KA, hy Le 0 = T,-(T,+4gh) 8° 9 = 9 dx? Hence, T-(I,+gg/h,) = @ = B,sinh Bx, +B,coshBx, BC’s: x, =0, dT/dx =0; T= Tygy Evaluating B, and B, gives T,-(T, apf) = (Tyg, ~(T, + ap/h,)}oosh Bx, T, = 300K A, = (n/4)(3x10 )? -(m/4)(3x10 * -2.5x10 3? = 1.129x10 4m? h, = 145 Wim? K © = x10) = 0.09425 m2-100 (continued) dg = (25)(0.09425)(0.1) = 2653 W/m? gh, = 2653/145 = 18.30 K B = (145x0.09425/401x1.129x10%)'? = 17.37 Hence, ‘Ty = 318.3 + (Ty - 318.3)cosh 17.37x, qd) Equation (2.36) applies to the unheated portion, T,-T, cosh B(L,~ SoanBer 1 = 0.05 m ToT, coshBL, Ty = 300 + (Ty300)(0.7135)eosh 17.37(0.05-x,) @ From Eq. (1), Ty is obtained by setting x, = 0.05 m to obtain Ty = 1.4017Tngg = 127.8 ® ‘A second equation is obtained by equating temperature gradients at x, = L;, and x, = 0. From Eqs. (1) and (2) aT, dx, 1, = Tags ~318.3)(17.37)sinh(17.37)(0.05) aT, Fok (14 ~300)(0.7135)(-17.37)sinh(17.37)(0.05) Hence, (Tyq¢318.3) = ~(Ty-300)(0.7135) @ Substituting for Ty from Eq. (3) and solving for Tyas Typax = 318 K and Ty = 309.2 K; AT=26 K Comments: 1. Fin type models are appropriate for a surprisingly wide range of problems. 2. Notice that numerical values were substituted before attempting to determine Thy, and Ty-analytical results for these quantities would be very complicated.Exercise 2-101 Given: A flat-plate solar collector with no coverplate. Required: Plate temperature distribution and heat transfer to the water at a location where the water is at 40°C. Assumptions: 1. Fin-type model of plate is valid. 2. Back face of plate is perfectly insulated. 3. The radiation heat transfer coefficient is constant across the plate. 4. Surroundings at the air temperature, For a plate of thickness t modeled as a fin, Gorae Chet hy T-T,) an energy balance on an elemental volume (Ax)(1)(0 leads to the differential equation for the plate temperature T, Eh hYT-T) “Gye = 0 dx? ‘where daar is the solar energy absorbed per unit area of plate, and h, is the radiation heat transfer coefficient defined by Bq. (1.19), and accounts for radiation exchange between the plate and surroundings (assumed to be at the ambient air temperature ,). Rearranging Eq. (1), at Dele be, Gee |] 2 9 @ ax? bn, ss x=0,T=T.; x=L, dT/dx = 0 (symmetry) Let BF = (hethkt T," = Ty + doia/(heth,): then ® From Eg, (2.36) the temperature at the midpoint (x = L) is T, = Te +(Te-Teeosh BL @ and from Eq. (2.40) the heat flow into the coolant water is Q = 2kt(1)BtanhBL(T, -T,) (6) ‘The radiation heat transfer coefficient is |, = 4aeT, ~ (4)(5.67x10 *)(0.2)(305)? = 1.3 Wim? K2-101 (continued) B? = (15+1.3(200)(0.004) = 20.38; B = 4.51 BL = (4.51)(0.06) = 0.271 T) = 295 +(750)(15+1.3) = 341.0 From Eq. (4), Ty, = 341 +(313-341)/eosh0.271 = 3140 K 41.0°C) From Eq, (5), Q = (2)(200)(0.004)¢4.51)(341 -313)tanh0.271 = 53.5 Wim Comments: 1. Solar radiation is dealt with in Chapter 6, Section 6.4. A more complete analysis, of a flat plate solar collector is given in Example 6.12. 2. Assuming h, to be constant at a guessed average temperature incurs a negligible error since h, >> hyExercise 2-102 Given: A 20.2 cm long aluminum rod transferring heat between two air streams. Required: leat transfer, and temperature of mid-section of rod. Assumptions: 1. ‘The 2 mm long mid-section can be assumed to be isothermal and adiabatic. 2. Heat transfer from the tips of the rod is negligible. Tw ‘The rod can be modeled as two pin fins, fin 1 on the hot side, fin 2 on the cold side. Since the heat flow is continuous across the wall, Eq. (2.40) gives Q = KAB(T.-Tp)tanbB,L = kA,B,(T,-T_,)tanh BL L=t0em ‘where the temperature drop in the rod as it passes through the wall has been assumed to be Leven Te negligible, ic., Tp B, = (hy OKA)! = [(48)(4)/(190)0.01)] = 10.1 m7; BL, = 1.01 Similarly, B, = 7.11 m'!; BpL = 0.711. Substituting, (10.1)(70-T,)tanh 1.01 = (7.11)(T,~20)tanhO.711 Solving. Ty = 52.0°C Q = (190)(7/4)(0.01)°(10.1)(70-52)tanh 1.0 2.08 W Comments: 1. Check the temperature drop in the rod as it passes through the wall: at = QAL . _ (2.08)(0.002) ee 028 KA, (190)(n/4)(0.01)? which is small enough to be neglected. (If you wish, rework the problem allowing for a temperature drop through the wall.) 2. The hot-side fin efficiency is n, = tanh 1.01/1.01 = 0.758,Exercise 2-103 Given: A 0.3 mm thick space radiator with heatpipes at 8 cm pitch. One side sees outer space at 0 K, the other is insulated. Required: Minimum plate temperature and heat rejection per unit area for heatpipes at 330 K. Assumptions: 1. ‘Temperature variation across the plate is negligible. 2. Radiation heat transfer can be modeled as q,,4 = 6e(T‘-0). An energy balance on an element Ax long requires FAI,“ GA coax IraAX = 0 a . Hence, ~£(QA,)“dai? = 0 Tl om Substituting q = -k dTidx, A, = Wt, © = W, > aT _ Int = 9 aK o dx? kt or HT Oe p4 2g @ a? it Boundary conditions are: x= 0, T= Tp, Ga) K=L, aTidx = 0 (symmetry) oo There is no convenient solution to Bg. (2) with boundary condition (3b). Thus we will try an approximate solution in which we assume q,,3 is a constant~since we expect T to vary only a few kelvins along the plate, Then integrating Eq. (1) aT _ Gna & Using Eq. (3b), 0 = AstL+c,: CG = “Hoy @ ead (1 2. T = 4) “x2 41x bc, Kt ea +}2-103 (continued) ‘Using Eq. (3a), Ty = C,. Thus the profile is } () ‘The minimum temperature is at x = L, Graal? (6) a © T-T, Sef To use Eq. (6) it will be sufficient to evaluate gy a8 a mean value, Gay = O€T* = oe(TZ+T, 2. GEL? Ts _ 499 - (5.6710 *)(0.8(0.04) Fz kt (2)(200)(0.0003) T = 330-6.048x10 1°T*%; T* = a30'eT 2 Solving by iteration, T,, = 323.12 K. Using Eq. (4) the heat flow at the base of the fin is :] QUWL = (5.67x10 *)(0.8)(330"+323.124/2 = 516.2 Win? ‘The fin efficiency is Q = ew ST), = ew) - = GagWL dx where yaa Gras oa oly = 96.0% Comments: 1. Check this solution by numerically integrating Eq. (2) 2. Attempt to obtain a solution to Eq. (2) for the boundary condition T -> 0 as x > (an infinitely long fin).Exercise 2-104 Given: Triangular profile cast iron fins, 2 om high, Required: Effect of fins on surface resistance for (i) he = 1000 Wim? K, (ii) h, = 8000 W/m? K. From Table A.1a, k = $1 W/m K for cast iron at 300 K. 2m} (@ —_FIN2, Item 3 gives the efficiency ny = 0.741 From Table 2.2 the surface area of one fin is 8/ = 2¢P-4.?)% = 2(0.01740.02)4 = 0.0447 m/m For I m length of channel, Re HS (1000)(0.0447)0.741) For an unfinned surface the surface resistance is simply T/heA = 1/(1000)(0.02) = 5x10 K/W Ren/Roare = 3.02K107/5x10 = 0.604 =3.02x107 K/W Gi) FIN2, Item 3 gives ny = 0.357 1 Re = OOO KODATIOSST ~ Lf, A = 1/(8000)(0.02) = 6.25x10 K/W Rén/Roare = 7.83x107 /6.25x107 = 1.25 83x10 K/W Comments: 1. We see that adding fins of a relatively poorly conducting material when the surface heat transfer coefficient is high, actually increases the resistance to heat flow. Note the similarity to the concept of a critical insulation thickness, dealt with in Section 2.3.2.Exercise 2-105 Given: Parabolic, rectangular and triangular straight fins of same base thickness and mass per unit width. Required: A demonstration that the parabolic fin dissipates the most heat. From Example 2.9, t= 0.003 m, L = 0.02 m, k = 187 W/m K, p = 2770 kg/m?, h, = 2800 Win? K, Tp = 500K, T, = 300K, Q= 11,370 W/m. Denoting the parabolic, rectangular and triangular profiles as 4, 1, 3 respectively, Table 2.2 gives the mass per unit width as 4. dtem4): Q/3)tLap 1. tem 1): 2tLyp 3. (tem 3): tsp 1y= 4 Ly =@/3\002)=001338m From Example 2.9,Q, = 11,370 W/m Using FIN2, Item 1, Q1 = 6,960 W/m Using FIN2, Item 3, Qs = 11,010 W/m Hence Q, > Q3 > 1. Comments: 1. Check that FIN2 does give the same mass per unit width for each of these fins. 2. Notice that the triangular fin is almost as good as the parabolic fin. (A triangular profile is usually easier to manufacture than a parabolic one.)Exercise 2-106 Given: A straight Duralumin fin with a parabolic profile. Required: Heat dissipation and fin mass. Assumptions: 1, The heat transfer coefficient is constant over the fin surface. Since the fin is expected to be very efficient, the thermal conductivity will be evaluated at 400 K; from Table A.1b, k= 187 W/m K. Using Table 2.2, Item 4, B= (hfe) = (40/187(0.003))* = 8.44; BL = (8.44)(0.01) = 0.0844 2 2 —) __ . __2 ___ cag BLY +1)* +1 (4(0.0844)*41)* +1 peed B= (14+4(UL)?)* = (144(0.003/0.01))" = 1.166 S°=LB + L7/21)In(QuL+B) = (0.01)(1.166}+(0.017/2(0.003))ln(2(0.003)/0.01+1.166) = 0.02114 m Q=h.S(Ta-Te)n = (40(0.02114)(400-300)(0.993) = 84.0 Win From Table Ala the density of Duralumin is 2770 ke/ta°; Fin mass = A,p = ] Lp = (2/3)(0.003)(0.01)(2770) = 0.0554 kgim n= Solution using FIN2. ‘The required input is: Selection number = 4 k= 187,p=2770 FIN? gives the output: 11 = 093; Q= 84.0 (watts/meter) ‘mass of fin = 0.0554 (kg) Comments: 1. The fin efficiency is very high, as assumed.Exercise 2-107 Giver: Various profiles of aluminum spines (circular cross-section fins). Required: () Heat dissipated by a4 mm diameter, 20 mm long pin fin, (Gi) Heat dissipated by parabolic and triangular spines of the same base area and mass. @ Eq. (2.40) applies: (nP)* _[4ne]*_[_ @a0_]*_. ga gy — = B: (| -(| =| Geoxax103) =23.6 m™; BL= (23.6)(0.02) = 0.472 Q=kA,BeTp=T,)tanh BL = (180)(17/4)(4x10 )°(23.6)(600~4000.440) = 4.70 W Check: FIN2, Selection No. 7, gives Q = 4.68 W. i) ‘The spines are items 7, 8, 9 and 10 in Table 2.2. Alll have a circular cross section with base diameter of 2t. Hence taking t = 2 mm will give the same base area for each. From Table 2.2 the mass of each profile is Item 7: pxtLy Tem 8: p(t/2)7Ls item 9: p(n /3)7Ly Ttem 10: p(n/5)7 Lo ‘Hence we require Ly = 2L7 = (2(0.02) = 0.04 m; similarly Ly = 0.06 m, Lyo = 0.1 m. Using FIN? the heat dissipations are aw Stem 7: Pin Fin 468 * tem 8: Parabolic spine, y =(3| 5.66 Ttem 9: Triangular spine, y = (ey 5.85 ao {3 5.86 Comments: 1. Q.is maximum for the item 10 parabolic spine, but the triangular spine is almost ‘as good. 2. See also Exercises 2-120, 2-121 and 2-125, Optimization of finned surfaces is usually a rather complicated task with some subtle issues involved. Item 10: Parabolic spine, y=Exercise 2-108 Given: An annular perforated plate heat exchanger. Required: The overall heat transfer coefficient, Assumptions: 1. A fin type model is appropriate. 2. The fluid temperature is constant across each stream. 3, One-dimensional conduction. Heat is transferred from the hot to the cold stream through three resistances in series: a fin on the hot side, the plate between the spacers, and a fin on the cold side. ‘The overall heat transfer coefficient is given by A, AOD) UAT hAng atk heAmc ‘where, following Example 2.7, tis the plate thickness (not the half thickness). We first determine Ds to give equal flow areas for each stream, (n/4)(D3-D4) = (/4)D} D3 - (0.053) =(0.05" ; D3 = 0.073 m. ‘The heat transfer area is the same for both streams, for the cold stream Az (2)1-0.3)(m/4)DF + (0.3(D /d)? mat = Q2)(1-0,3)(e/4(0.05 + (0.3)(0.05/0.0015)*(n)(0.0015)0.001) = 4.32x10- m? Next we need to determine the fin efficiencies. For the cold stream we have an annular fin with the efficiency given by Eq. (2.52) and B evaluated as follows. For an element of disk between r and rtAr we have a cross-sectional area for heat flow of A. = 2nrt(1-0.3), whereas the surface area for heat transfer is 2mear Arta HO3) CTs To get the perimeter we divide by Ar, P=2nr(2)(1-0.3)0.3)(8xrt/A) AP _ hel2ar2)(1-0.3H(0.3Gan/A)] _ hel2-03+4(0.3)(t/¢)] dt pe ete Mant ‘ate KA, ‘K(2nrt(1-0.3)] ‘kt(1-0.3) 2 _ 300 {1.44(4)(0.3)(0.001/0.0015)] _ yogp. g = Hence B* = 55 O00 = 4962; B=70.4 ‘Then Br, = 1.866, brs = 2.570,and using Eq. (2.52), _ 2r2/B) Ki Brah Grs)-h Gr)Ki Bra) Me = ha) Ko(Bra)h Gra) loBr)Ki Gs)2-108 (continued) = — (053/104) _(0.1673)2.681)-(1.402K0.06778) _ » 549 (0.073?—0.0537)/4 (0.1345)(2.681)-+(2.082)(0.06778) ‘Next we must obtain a formula for the effectiveness of the hot stream fin, Eq. (2.51) is the solu- tion for the temperature distribution, = Cylo(@) + C2Kolz) with boundary conditions z= @: d0/dz = 0; z= 2;:6= 1. ‘Evaluating the constants gives 0 = In(2)/Io(z1). Also the heat dissipated is the negative of the heat flow at r= Oy =H, Bn, = KOR, I-O3\T eT IB ei, Biv, =IleMlolahs 4) =Br1 = 1.76 Since the maximum possible heat transfer is hA(Tp—Te), the fin efficiency is me = k(2nr 11-03) B/B ANT (@)/loC2) = (190)(1x0.05x0,001)(1-0.3)(70.4/300x4.32x10°3\1.267/1.940) =0741 ‘Thus the overall heat transfer coeflicient is obtained from oe 1 In(0.053/0.05)_ 1 UA (300)(4.32x10*}(0.842) (2m)(0.001)(190) (300)(4.32x10* 0.741) 0.9164 + 0.0488 + 1.0413 006 KW U= 1154 Wim? K Comments: 1, Aluminum or copper are used for these perforated plates to ensure a high fin efficiency. 2. New manufacturing methods developed for MEMS application allow hole diameters as small as 1m. Then performance is limited by the fin efficiency of the plates and hence by the k-value of the material. Single crystal silicon then becomes a candidate material (k = 2500W/m K at 50K). See Exercise 2-126Exercise 2-109 Giver: 2.cm O.D. heat exchanger tube with a 0.2 mm thick, 5 om O.D. steel spiral fin. Required: Elect of fin on outside thermal resistance. Assumptions: 1. Steady state. 2. Approximate spiral as annular fins. ‘The bare tube has a thermal resistance per meter of length of ‘The half thickness of the annular finis t= 0.1 mm, 4. _09)___ 3571; 5 =598- 60m Kt (42\(0.1x10) Br; = (0.01)(60) = 0.60 Bro = (0.025)(60) = 1.50 ay fer 0.60 1.0920 0.3137 0.7775 1.3028 150 0.9817 02774 Using Eq. (2.52), = —2(0.01)/60_1,3028)(0.9817)-(0.3137)02774) _ 9 749 (0.025?—0.017) (0.7775\(0.9817)+ (1.0920\(0.2774) ‘Equation (2.46) gives the total surface efficiency, m= HAVANA) Ata pitch of 3 mm there are 333.3 fins/meter length of tube Ag = (2)()(0.025 -0.017)(333.3) = 1.099 m? ‘A= 1,0994(70)(0.02)(1-{333.3)(0.0002)) = 1.099+0,0586 = 1.158 m? the = 14(1.099/1.158)(1-0.710) = 0.725 Reinned = Uh AM, = 1/(15)(1.158)(0.725) = 0.0794 ne Comments: 1. Rggaea/Roare = 0.0794/0.796 = 0.10, a tenfold reduction. 2, Spiral annular fins are widely used in engineering practice.Exercise 2-110 Given: Thin disk heated on one side, cooled at its edge. Required: Temperature at center. Assumptions: 1. Steady state, 2. Temperature variation across plate thickness negligible, A fin problem: Eq, (2.48) applies with t taken as the disk thickness since only one side is exposed to air. The 7 solution for the temperature distribution is Eq. (2.51), 7, w TT B=Cylo(z)+CaKo(zy, O= Bors 2-0, 2 = > 2=Br 0 4 Rae 2=BR,O=1 0= Ci @)-CoKy(O); Co =O since Ky(O)= = 1= CyIg(BR); Cy = 1/1g(BR) : WG ‘Substituting back, 6: TGR) WO 1 Graaiaaaia FBR) ~ TBR) ‘Thus T| no =Te ~ (Te-Ta)/To(BR) Comments: 1. A table of the zero-order modified Bessel function of the first kind, Tp, can be found in Appendix B.Exercise 2-111 Given: Rectangular fin with temperature difference (1(x)-T,) constant. Required: (@) Finefficiency formula, }) Value of ny appropriate to Example 2.7. Assumptions: 1. Insulated tip. @ Equation (2.31) remains valid, &T ea? Setting T-T, = AT, a constant, and introducing B* = h,P/kA, gives a Par 0 x L ae T= 5 BAT? +Cix+ Cy -bPCT-T,)=0 = Ateconstent ‘As before we use as boundary conditions T = at x = 0, and dT/dx = Oat x= L, then C; = -#P ATL and Cz = Ty. The resulting temperature distribution is TpT= BAT fe 7 + An appropriate definition of fin efficiency is 8 h.PLAT aT. MO PLTyE) PLT.) To-Te where T, is the average temperature of the fluid. TpT, = Te -(T-AT) = (Ty-1) + AT 1 1 a = rear fu 7 2 fowsr = @LYAT3 + AT fouee AT 1 G@LYATZ+AT — G2/3)+1 andr Gi) Substituting x =2.31 gives ny = 0.360. Comments: 1. This value of ng is 15% less than the value obtained for T, = constant in Example 2.7.Exercise 2-112 Given: A cylindrical transistor cap. Required: (j) Heat dissipation. Gi) Base temperature for 400 mW dissipation. Assumptions: 1. Heat loss from outside surface only with a constant heat transfer coefficient over the surface. 2, ‘The metal thickness t is small so that a fin model is appropriate. @ We split the cap into two fins, (1) a straight fin of width 2xR and length L, and (2) a disk fin of radius R. For the straight fin with T= at x =0, Eq. (2.35) gives 1 L ‘TT, = BysinhBx + (Ty—Te)cosh Bx; B= say For the disk using Eq. (2.51) with dT/dr=0 at r= 0, : ToT, = Co(B0 B= (hesk”* e . ‘The constants B, and C, are determined by matching the temperature and heat flow at the join, TWL)=TR): ATyixle=—AToldtln i) since kA, is the same for both fins at the junction. Substituting Eqs. (1) and (2) in Eqs. (3) and ©) gives Cylo(BR) = By sinh BL + (Tg~T.)cosh BL ~Ciln(BR) = B cosh BL + (Ts~Tsinh BL Jo(BR)sinn Ly (BR)cosh BL To(BR)cosh AL, (BR)sinh BL ‘The heat dissipation is the base heat flow of fin 1, . aT; Q=-kA, = Lao =-KA_BBy = 2nkRIB(T—T,) Solving, By =~(Tp-Te) oR HL GR.oxh HL, Io(BR)cosh BL+h; (BR)sinh BL Gi) B= Jka) = [25(50)(0.0003)}* = 40.8 mt BL = (40.8)(0.009) = 0.367; BR = (40.8)(0.004) = 0.163 ‘Ig(0.163)sinh(0.367)+1, (0.163)cosh(0.367) ‘Io(0.163)cosh(0.367)+I, (0.163)sinh(0.367) (0.4 = (2)(50)(0.004)(0.0003)(40.8)(Ts-310) Solving, Ts = 371.7K Comments: 1. Once again, use of the fin model allows a seemingly complicated problem to be reduced to a much simpler problem.Exercise 2-113 Given: A straight fin with a parabolic profile. Required: (i) Governing differential equation. (ji) Temperature profile. (ii) Efficiency for a linear profile. @ For unit width of fin, A, = 2y = 20(1-x/L)* = 22?/L?, and P = 2. Substituting x = L-z in Eq. 2.30) and introducing Fourier's law, Y av 2) 4 pert) =0,0r-—+ Ack “| BPCI-T,) =0, 0 + B= (yy Uz (1) 2 OT yo, AT _ pep rer 2p eG, PLT = eee or x}_[z} ‘ i @ Fen [TL L ar 2)" (1), er. 2)" a] Lo arytyp z| [2]: a -eto20-[3] t Substituting in Eq. (1) and cancelling terms, p(-1)+2p-BL2=0 or p%p-B°L? =0; solving, p=- ++ J caput Gi) Forp=1,f*L? =2. From Table 2.2, Item 4, Comments: 1. In setting P = 2 in part i) the curvature of the profile has been ignored. The analyses giving the formulas for ni in Table 2.2 are based on this assumption. In fact, the true surface area for convective heat transfer is, we a ys = BY 21 Pax =20x2oy2y% or P= 2tna( SL) 7 Y= t[1-4/L)P; dyfdx = -QUL)[1-(x/L)] P=2(144/L2)[-@L)P)* C.J. Maday ("The minimum weight one-dimensional straight fin" J. Engng. for Industry, 96, 161-165 (1975)) obtains the optimal fin profile taking into account the profile curvature. However, the one-dimensional heat conduction assumption should ‘also be questioned for such profiles. 2. Notice that fins or spines for which A. —> 0 as x —>L, require that T—> T, as x—>L. ‘That is, the tip boundary condition is T, = Te.Exercise 2-114 Given: Annular steel fin of uniform thickness cooled by air at 20°C. Required: "The rate at which heat is dissipated by the fin, Assumptions: 1. Heat wansfer coefficient constant over the fin surface. 2. Heat loss from tip negligible. t= 1mm k= 60 Win K h, = 24 W/m? K Ts = 110°C T.=20°C For an annular fin B* =h,/kt BF = 24)/(60)(0.001) = 400m? B=20 mr ‘The fin effectiveness is given by Eq. (2.52) y= LEB KxGrIh Br2)-L, BK Gr) 7) Ko(Brh Br)Ho(Br Ki Bre) Br: = (20)(0.02)=0.4 Br =(20(0.04) = 0.8 ‘From Appendix B, Table B.3 we find the required values of the Bessel functions and substitute in Eq. 252) (2y(0.02)/(20) (2.18435(0.43285)(0.20402)00.86178) _ 9.939 "= "(o.042-0.02%) (1.11453)(0.43285}+{1.04040)(0.86178) ‘The heat dissipation is the efficiency times the dissipation for an isothermal fin. methe)(2}¢n)(3-19 )Tp—Te) = (0.930)(24)(2)¢x)(0.047-0.02")(110-20) = 15.1 W Comments: 1. Check this result using FIN2,Exercise 2-115 Given: A surface to be provided with triangular fins. Required: Percent increase in heat dissipation. Assumptions: 1. Heat transfer coefficient unchanged by addition of fins. ‘rom Table 2.2, Item 3 the fin efficiency is aye NOHL) BE To@BLy B= (het) = [20/(50)(0.003))* = 11.55 m™ 2BL = (2)(11.55)(0.03) = 0.693 Interpolating in Table B.3 of Appendix B gives From Eq, (2.46) the total surface efficiency 7, is T= H{Ae/AMI-M) For 1 m square of surface the fin surface area is Ar = (1/p)S’(1) where p is the fin pitch and S’ is the surface area of fin per unit width, From Table 2.2, 8’ = 262417)" = 2{(0.003)°+(0.03)"1* = 6.03x107 Ag =(1/0.015)(6.03x107) = 4.02 m> ‘The exposed area of the surface is 1-(1/0.015)(0.006)(1) = 0.6 m?. Hence the total heat transfer surface area is A = 4.0240.6 = 4.62 m*. ty = 14.02/4.62)(1-0.945) = 0.952 Any (Tp~Te) = (20)(4.62(0.952)(180-80) = 8796 W ‘Without the fins the heat dissipation is simply Quntenes = (20(1)(180-80) = 2000 Hence Q/Quatanes = 8796/2000 = 4.40, ie., 440% increase. Comments: 1. Check ny using FIN2. 2, The large increase in Q suggests that use of the fins may be cost effective.Exercise 2-116 Given: A thin disk exposed to a high temperature radiation source. Required: Temperature difference between center and edge. ‘Assumptions: 1. Steady state so that the temperature is constant across the plate thickness. 2. Negligible reradiation. ‘A fin type analysis is appropriate, An energy balance on an elemental volume of thickness t between radii r and red requires that ar aT eC ered Cer “h,2nrAr(T=T,) + Gag2Rtbr = 0 4 Dividing by 2nikAr and letting Ar > 0, Ef at|_ he Saat le Ot or? Buk - PPE) <0 Tate Following the analysis in Section 2.4.4,we let z = Br and @ = (T-T)/(T(R)-T-), where TR) is . Then which again is a modified Bessel’s equation of zero order and has the solution 8=Cylo(2) + CaKo(2) B.C.1: z=0: dT/dr=d0/dz=0 by symmetry C11) — C.K) =0; C2 =0 since Ky(0) =~. B.C.2: z=BR: T=Tg ord 1=Cylo(BR): hence C; = 1/Io(BR) ‘Substituting back, @ = Io(B1)/Io(BR) [At the center Br=0, 1o(0) = 1, hence 6(0) = 1/la(BR) Finally, the required temperature difference is TO-(Tetdnalhs) 1 TR-Tetqaa/h.) ~ To(BR) Comments: 1. A table of To, the zero-order modified Bessel function of the first kind, may be found in Appendix B.Exercise 2-117 Given: Air cooled intemal combustion engine with hyperbolic annular fins. Required: (i) Heat loss from one cylinder. (i) Comparison of the efficiency of the hyperbolic fin with a rectangular fin of same base and length. Se aeees T. = 300K ope. st Tp =485K 7 a h, = 60 Win K 4 6emd k= 174 W/m K (at 400K, Table A.1b) ieee p=2790 kg/m? @ The number of fins per cylinder is N = 0.12/0.012 = 10. Use FIN2 to find the heat lost, from one fin, The required input in SI units is: Item number = 6 ‘Thermal conductivity and density of fin = 174, 2770 Heat transfer coefficient = 60 Base temperature and ambient temperature = 485, 300 t= 0.003, 1 and rp = 0.06, 0.08 FIN? gives the following output: Fin efficiency = 0.981 Base heat flow = 191.8 W Mass of fin = 0.1253 kg ‘The total heat flow for one cylinder is 10)(191.8) = 1918 W. (ii) Using FIN? for a rectangular fin of the same base and length gives: Fin efficiency = 0.983 Base heat flow = 191.9 W Mass of fin = 0.1462 kg Comments: 1. Although the hyperbolic fin dissipates essentially the same heat as the rectangular fin, its mass is 14% smaller. 2. See also Exercises 2-120, 2-121 and 2-125. Optimization of finned surfaces is usually a rather complicated task with some subtle issues involved.Exercise 2-118 Given: Heat exchanger core consisting of finned passages between parallel plates. Thickness of fins giving 95% fin efficiency for 6) Aluminum alloy, k = 180 Wim K ii) Mild steel, k= 64 Win K ii) A plastic, k= 0.33 Wim K Assumptions: 1, The heat transfer coefficient is constant along the fin, Rw L=W=Smm hg = 160 Wim? K @ Aluminum alloy, k = 180 W/m K. Equation (2.42) gives the fin efficiency: nie = tanh x3 % = BL = (hePL?/KAc)* = (hel? /4kt)* since the fin length is L/2 = [(160)(Sx107)7/4 (1800) = 2.36x10* 095 =tanh yy; x= 0.40 : | 236x107 5 nao 348x105 m Gi) Mild steel, k= 64 Wim K 423.9610 t= page 96810" | 8010-5 m Gi) Plastic, k= 0.33 Wim K y= 551x107; t= [sue 551s10* | 90x10 m Comments: 1. For both metals the required thickness is so small thet a thicker fin would be used to simplify the manufacturing process. 2. Since t> W for the plastic fin, a 95% fin efficiency cannot be achieved for these core dimensions,Exereise 2-119 Given: A 2.cm O.D. heat exchanger tube with a4 em O.D. steel spiral fin. Required: Effect of fin on outside thermal resistance. Assumptions: 1, Approximate spiral fin as annular fins. 2. Tip heat loss negligible. ‘The bare tube has a thermal resistance per meter of length of = 0.796 K/W m Use FIN? to give the fin efficiency. The required inputs are: Item 5: annular rectangular fin kp =42, any value n= 150 Tp,Te = any values t= 2x104 ry = 1x10? f= 2x10 ‘The output is: ny = 0.923 Equation (2.46) gives the total surface efficiency, m= MAPA) Ata pitch of 3 mm there are 333.3 fins/meter length of tube Ag = (2)(a7)(0.027-0.017)333.3) = 0.628 m? A=0.628+(R(0.02)11~(333.3)(0.0004)] = 0.628+0.05446 = 0.682 m? ‘Ny = 1{0.628/0.682)(1-0.923) = 0.929 Ronea = 1/heAMy = 1/(15)(0.682)(0.929) = 0.105 W/K Comments: 1. Resnea/Rpare = 0-105/0.796 = 0.132, i.c., the resistance has been reduced to 13% of the bare tube value. 2. Spiral fins are widely used in heat transfer practice.Exercise 2-120 Given: Straight fins of rectangular, triangular and parabolic profiles. Required: Optimal proportions. Assumptions: 1. One-dimensional conduction in the fins. 2. The tip of the rectangular fin is insulated. 3. Fin surface area for an element of fin Ax long is 2WAx. (i) From Eq, (2.40) the heat loss for a rectangular fin of width W and thickness 2t with W >> t and B= (h, /kt) is Q = kQWOB(T,-T.)tanh BL a) arses QL LE CEES Een UO EE op ALE op fee BE QKW(T,-T 229 (hk)? ~[ee%@ta]? sony [pt]? tanh, ang 7 GAL} aay? } Fy? Differentiating, dQ ‘dy, = sech*y~-” ~ hx Meanhy dQ “dy = 0 gives 3 = sinhycoshy for Qua Solving by iteration, 7.5 627. A graph of Q “(x) follows. 419, Qiae = 0.791, Mog = (ii) From Table 2.2, item 3 _1 10” a x LD and Q = h,QW)L(T,-T,ny @ where the “length of arc” assumption has been used to write © = 2W (see Comment No. 2). ©) i x) he (x) Wao Q _ hL| 20h? )” Ley) _(2Ve len 2WK(T,-TUL/20h ‘The optimal value of x is most easily found numerically from data used to produce the Q “(y) ‘graph that follows. The result is Yo. = 1.309, Quay = 0.895, Xroe = 0.594.2-120 (continued) (iii) From Table 2.2, Item 4, 2 ae © Tae? Using Eq. (4), Q* is Q 2WKCT, TILK) 3 bL| 3k? 210 = | } 2 = 260) a kK tL 144241)? 1+(4y2+1)? Again, the optimal value of x is most easily found numerically from data computed to produce the Q°(X) graph that follows. The result is Zope = 1.414, Quix = 0.909, Nope = 0-5. (iv) Its seen from the graph that the parabolic fin gives the highest heat loss for a given fin ‘mass at all values of x. The maximum attainable value of Q* is 0.909. Comments: 1. See also Exercises 2-71 and 2-113. 2. Standard fin analyses use the so-called Murray-Gardner assumptions that include the approximation dS = dx. The expressions for nin Table 2 for non-rectangular profiles have this approximation. However, the expressions for S’ and S in Table 2.2 are exact. Use of exact values of S’ and S improves the accuracy of the calculated heat transfer. Of course, this issue is of no concem for the slender fins usually encountered in engineering practiceExercise 2-121 Given: Rectangular and triangular profile aluminum fins with a base thickness of 3 mm, and heat transfer coefficient of 50 Wim? K. Required: Graph of heat loss versus fin mass. Assumptions: 1, Heat loss from rectangular fin tip is negligible. 2. Constant heat transfer coefficient. Using FIN2 with k = 180 Wim K, p = 2770 kg/m’, Ty-T, = 1 K, the following results are obtained. Note that Q is the heat loss/K m and W is the fin mass/m. Fin Length Rectangular Fin Triangular Fin Q w Q w ae WKm kgm = WKm kgm 0.994 0.0831 «1.002 0.0416 1 2 1952 0.1662 1.935 0.0831 3 2844 0249-2778 0.125 6 4947 0499 4611 0.249 9 6180 0.748 «55810374 2 6808 0.997 6084 0.497 15 7405 «1.247 6.369 0.623 From the graph we see that for a Ma 12] [== Reotengular . given Q the triangular fin has the —»— ‘Triangular smaller mass. (However, see Comment No. 2.) W (kgm) (WK m) Comments: 1. Although useful, these results are somewhat misleading since we are not comparing optimal fins. For a given heat flow there are optimal fin dimensions (see Exercises 2-71 and 2-120). Exercise 2-125 revisits this problem to explore the consequences of using optimal fin dimensions. 2. If FIN2 is used to extend the table to larger fin lengths, say up to L = 30 cm (rather unrealistic!) then the Q versus W curves cross and the rectangular fin gives a smaller mass. Is this result believable or is it due to numerical error in evaluating the heat flows?Exercise 2-122 Given: A 10 mm O.D., 9 mm LD., R-22 condenser tube with 50 mm O.D., 0.2 mm thick fins. Required: Heat loss per unit length for R-22 at 320 K and cooling air at 300 K. Assumptions: 1. Outside heat transfer coefficient constant of the fin surface and tube wall. 2. Heat loss from fin tips is negligible. Equation (2.47) gives the thermal resistance of a finned surface as \ hAn, | nat as | where vm i SEA eli co nm, = 1-(AYA)—p) is the total surface efficiency. The fin efficiency ny is obtained from FIN2: the required input is ‘tem number: 5 (Annular rectangular); k, p: 180, any value; h,: 20; T, and T.: any values; t: 0.0001; r, and r,: 0.005, 0.025. The output gives ny = 0.7586. For 1 m length of tube, Ay = (1/4)(0.05?-0.017)(2)(1/0.002) = 1.885 m? A = 1.885 +(#)(0.01)[(1)-(1/0.002)(0,0003)] = 1.885 +0.0267 = 1.912 m? 1, = 14(1.885/1.912)(1-0.7586) = 0.7620 1 1 —— = —__ = 3.43x10? K/W nA, — (@0)(1-912\(0.762) R= From the thermal circuit, ini Mfr) aR, 2a 1 t0(5/4.5) cea 2 s a = —____+__“~_—~_ +3.43x10 7 = 4.42x10 * +9.3x10 © +3.43x10 HONGO) ComyIs) i: a UA = 12.7 W/K (per meter length of tube) QIL = (12.7)(320-300) = 254 W/m ‘Comments: 1. Notice how the use of fins has reduced the outside resistance to a value comparable to the inside resistance. 2. The tube wall resistance is negligible and can be ignored.Exercise 2-123 Given: A 1 mm half thickness, 10 mm long, straight rectangular fin, k = 100 W/m K, h, = 40, 1000 and 25,000 W/m? K. Required: Values of h, so as to simulate with a 10 m long fin. Assumptions: 1. Convective heat loss from fin tip with the same value of h,. ‘The similarity principle must be satisfied in the simulation. In Section 2.4.5 it is shown that this principle leads to Eq, (2.58), namely 8 = 06 x,Bi) for a fin with convective heat loss from its tip, where Wa. h, fae =|] L: Bie it K For two fins to be similar, the parameters x and Bi must be the same; then 6(€) and n)¢(x,Bi) will be the same. Consider h, = 1000 W/m? K on the 10 mm long fin, 2 _[ne]”, (100% oon ‘ (:] i (arta fone hd _ 2000)(0.01) KE 100 Bi =01 ‘Now let us match these values for a 10 m long fin. We require 2 be} 10 _ ( aol @ (410) _ Ol 3) r @) From Eq. (3) we obtain h/k = 0.01. Substituting in Eq. (2) then gives cooy?t® -1; t=10m a Since only the ratio hy/k appears in Eqs. (2) and (3) we are free to choose k and calculate h,. Choosing k = 100 Wm K as for the 10 mm fin,2-123 (continued) h, = (0.01)(100) = 1 Wim K. ‘So we see that similarity requires geometric similarity (Lit = 10 = constant), but a heat transfer coefficient 1000 times smaller on the 10 m long fin. Similar calculations can be made for the other two values of h,: the results are tabulated below. L=10mm L=10m hg Whn? K x Bi hJk, m' | h,, Wim? K 40 02 4x10" 4xio4 0.04 1000 10 4 oor 1 25000 50 25 0.25 25 Comments: 1, If the same k value is used, h, must be reduced by a factor of 10° for the 10 m_ long fin, 2. Because of the geometric similarity, the Biot number based on thickness b,t/k will also be matched in the simulation. This is important if two-dimensional conduction effects are to be explored (see Exercise 3-18). 3. The numerical error problem is peculiar to the particular commercially available software package, and the way it requires problems to be formulated. It is not an inherent difficulty associated with using numerical methods of the kind described in Section 3-5.Exercise 2-124 Given: A 2.cm long aluminum alloy pin fin with a reduction in diameter from 2 mm to 1 mm, 8 mm from its base. Required: Heat loss, and temperatures at contraction and tip. Assumptions: 1. Negligible heat loss from tip. 2. Negligible heat loss from exposed contraction area. Equation 2.40 gives the heat loss from the thin portion of the fin as Q) = kA B,(Tp, -T tanhBL, oO cL =20'c in terms of the unknown temperature Typ. Equation (2.35) gives the temperature in the thick portion of on Peeeed| the fin as T-T, B,sinhB,x+B,coshB,x; B, = Tyy-T, *™ Hence Ty,-T, = B,sinhB,L, +(Tp,-T,JooshBL, 2) ° and a = B,B, coshB,L, +B,(T'y, -T,)sinh BL, Rha, ‘0.008 Also, if we assume that there is negligible heat loss from the exposed contraction area, dT| aL. ® To solve Eqs. (1)-(4) for the unknown constant B, we first substitute the given data, B,L, = 0.1704 B, = (4h /kD,)" = [(4)(42(185)(0.002)]"" = 21.3; B, = Gh, /kD,)!? = [(4)(60)/(185)(0.001)}! = 36.0; ‘Substituting in Eqs. (1) through (4): Qj = (185)(01/4)(0.001)*(36.0)(T y) -20)tanh 0.432 = 2.130x10 (Tg) -20) Tpa-T, = B, sinh 0.1704 +(200-20)cosh0.1704 = 0.17128, +182.6 BL, = 0.432 002 @ @ o ©.2-124 (continued) = (21.3)B,cosh0.1704 +(21.3)(200-20)sinh 0.1704 = 21.61B, +6564 o a@ -1.25x10 *B, -0.3815 = Q, @) ~(185)(m/4)(0.002)'(21.61B, +656.4) Substituting Eq. (6) in Eq. (5), and the result in Eq. (8) gives 2.130x10 °(0.1712B, +182.6) =1.256x10 7B, -0.3815 B 59.6 From Eq. (6), Tyy-20 = -(59.6)(0.1712)+182.6; Taz = 192.4°C ‘The heat loss from the fin is Qe 6A = “KA,sB,B coshB,(0) ce Q, = ~(185)(9/4)(0.002)°(21.3)(-59.6)(1.0) = 0.737 W ‘The tip temperature can be obtained from Eq. (2.36), Typ = T.+(Tp-T.veoshB,Ly = 20+(192. -20Veosh0.432 = 177.5°C Comments: 1. To check our assumption of negligible heat Joss from the contraction let us guess +h, « 50 W/m?; then Qeoneastion = heA(Tg. ~T.) = (50)(n/4)(0.002? -0.001)(192.4 -20) = 203x107 W which is 2.7% of the total heat loss. 2. At the contraction we have two-dimensional conduction as the heat flow is squeezed” from the thick to the thin portion of the fin.Exercise 2-125 Given: Aluminum fins with h, = 50 W/m? K to dissipate 4.947 Wim K. Required: Optimal dimensions and resulting optimal mass for rectangular and triangular profiles. Assumptions: 1. Constant heat transfer coefficient. 2. Tip of the rectangular fin is insulated. 3. The ‘length of arc assumption’, dS = dx. From the problem statement for Exercise 2-121, we define a nondimensional heat loss as e i 2KW(T,-T IA /2(k/h, mT A,=2CiL, C 1, 1/2 for rectangular and triangular fins, respectively. Also the fin parameter x is defined in the usual manner, x = BL, B = (h/kt)!?, For a rectangular fin, Q* has a maximum value of 0.791 at y= 1.419: we will find the corresponding dimensions. 4.947 O79] eee eee (2)(180)(1)(1)(2Lv2(180/50)° 72> or Lt = 6.79x10°* a (so/180))'"L = 1.419; or L = 2.6912 @ Solving Eqs. (1) and (2), t = 8.60x104 m, L = 7.89x10? m. Using FIN2 we check that Q = 4.947 W as required, and W = 0.3759 kg/m. For the triangular profile Q* has a maximum value of 0.895 at x = 1.309. Hence, Lt = 9382x104 ® = 2.4831"? @ Solving, t = 1.126x10% m, L = 8.33x10? m Again, using FIN2, we check Q = 4,947 W as required, and W = 0.2598 kg/m. Comments: 1. ‘The optimal triangular fin has a mass that is 69.1% of the rectangular fin 2. However, the optimal triangular fin has a mass that is 90.2% of the original triangular fin (t= 1.5 mm), whereas the optimal rectangular fin has a mass that is 75.3% of the original rectangular fin. 3. This exercise has examined optimal dimensions of a single fin. A related ‘question is the optimization of a finned surface, ic. if the heat flow per unit length of surface is specified, what is the optimal fin pitch and fin dimensions.Exercise 2-126 Given: “A perforated plate heat exchanger with rectangular plates of fin length 20 mm, 0.5 mm thick and 90 um diameter holes in a square array of pitch 130 jum, Required: Fin efficiency for h, = 800 W/m? K and (i) copper (k crystal silicon (k = 2500 Wim K). Assumptions: 1. Plate thin enough for a fin-type analysis to be valid, 90 Wim K), (i) single 2. One-dimensional heat conduction. Use the formulas derived in Example 2.7. The convec- tive heat transfer area A is mt sea = (0.08)¢0.02]2-%{ © . (2X90500) |= 5380192 mq? 2\13) ~~ Gaoe ‘The fin perimeter is? = A/L, = 1.538 x 107/0.02 = 0.769m. ‘The cross-sectional area for heat conduction along the fin is approximated as A. = Wi(L-e,); €, = 0.376 for d/p = 9/13 Ag = (0.08)(0,0005)(1-0.376) = 2.50x105m*, ‘The fin efficiency will be taken as given by Eq. (2.42), Ty = tanhy/x; X= BLy: B* = h,O/KA,. (i) For copper with k = 890 W/m K, B= 1663 Uh,OL yn; = 0.271 KW X 3.326 W(kEW/L,) = 0.112 KAW ny = 0.300 UA = 1.529 WK U=994 Wim? K i) Por single crystal silicon with k = 2500 Wim K, B= 99.21 Ih, OL ny = 0.168 K/W 984 LUKE WIL,) = 0.040 KAW 485 UA = 2.660 WK U=173 Win? K Comments: 1. 1a comparing these results with the result of Example 2.7 we must compare the UA product. In Example 2.7: UA = 0.248 WIK. Present, copper: UA = 1.52 WK Present, silicon: UA = 2.66 WK ‘The advantages of going to smaller hole sizes and using silicon are clear.Exercise 3-1 Given: Cylindrical coordinate system. Required: Derivation of the general heat conduction equation. ‘The figure shows the elemental volume rAg-ArAz, r Applying the energy conservation principle, Eq. (1.2) Td gives a. y P(cAgArAz)c S-= Q+Qy ee ‘The net conduction into the element is = Grae LetaeleardAGA2 + (49 Loa long) + Gila-Gal eeaa)rAQAr where ripletielnar== 2 (ranars qy=k ar olor Splonan == (ap; 94 =e SE pres ea eoa ed Gale-Gelanae=— 3 G)M2s q=-k ‘Substituting and rearranging, 12 a aT aerial fe zh. 23h] 2 hal ‘The internal heat generation is Q. =Q (ehodraz) ‘Substituting into Eq. (1.2) and dividing by rAgArAz gives, Bes sehar stg <Exercise 3-2 Given: Spherical coordinate system. ‘Required: Derivation of the general heat conduction equation. ‘The figure shows the elemental volume rA@-rsin®Ag-Ar. : Applying the energy conservation principle, Eg. (1.2) ary gives an see SY prAO-rsinbAg-Anc z =Q+0,, xf t ‘The net conduction into the element is Q = (Parl Par lvar)AOSINO AG + (Ap 1qMp | peng AOA + (sin dg | e—sin gg | eyao)KAGAT where qe] 1dr lets =~ 2 (Pq)ars ae a ply 4p gong = Fo: = 555 7 sing o-sinOag onan = 2, (Sin0q9)40 ; gp =k SE ‘Substituting and rearranging, sea taf, am, 1 afar] a= Penge oF fe a Faint Op jx ‘The internal heat generation is =O, (AG rsindAG:Ar) ‘Substituting into Eq. (1.2) and dividing by the volume of the element rsin6ArA@Ad givesExercise 3-3 Given; A solid medium. Required: Derivation of heat conduction equation using vector calculus. Consider a control volume V enclosed by surface S n located in the solid. ‘The element of surface dS has a ‘unit normal vector n directed outward. ‘The energy con- servation principle for a closed system, Eq. (1.2) requires that Rate of storage _ rate of inflow rate of generation of intemal energy ~ of heat * of thermal energy s Applying it to the control volume V Fe fouw fanas+ [rav Use of the Gauss divergence theorem allows the surface integral to be converted into a volume integral, [anss-[vaav 8 Substituting du = eT and rearanging, J fe = +%a8ijav=0 4 ‘Since the volume V is arbitrary, the integrand must be identically zero, hence a . Hey, pc a q+Qy Introducing Fourier’s law of heat conduction, Eq. (3.2), gives oe Lavarn+e, If the thermal conductivity k is constant, pe Fane Comments: 1. Application of this equation requires that the operator V? be expressed in terms of a suitable coordinate system.Exercise 3-4 Given: V, V- and V? operators in general orthogonal curvilinear coordinates. Required: V7 in cylindrical and spherical coordinates. @) Cylindrical coordinates Trg.2) x= 10030 im ds? = dr? + Pag? + da? * ‘ Hence uy =r, hy = li up =6, hy = 1 Us =z, hy = I (see Fig. 3.38) tabs, aT toby ar ofS VI= hy dus Tnibsbs [Buy [hy z)-2 amy {he Tera) = dr? + Psin*@dq? + 17d0? (see Fig. 3.3) Hence uj=r, hy=1 Comments: 1. A derivation of these equations for V, V- and V? can be found in many applied mathematics texts, for example, L. A. Pipes, Applied Mathematics for Engineers and Physicists, 2nd ed., McGraw-Hill Book Co., New York, 1958, pp. 395-398.Exercise 3-5 Given: Melting ice block, one face receding at 0.22 mm/minute. Required: Temperature gradient in water adjacent to the ice. Assumptions: 1. The ice has been melting for some time. From Eq. 3.18), ceo ee Oa TY, i oy ‘Since the ice has been melting for some time it can be ‘assumed isothermal at the melting temperature 0°C, so AT/Ax|y = 0, then ah, If the interface is located at x = ar ox FRE? 2, One-dimensional heat flow. = ats & = 0.22 mm/minute = 3.67x10°° m/s = 910 kgm’, the density of ice at 0°C = 3.35x10° J/kg, the enthalpy of fusion of ice = 0.552 Wim K, the conductivity of water at 0°C G67x10°)(910)(3.35x10°) _ a o352) = 2030 K/m~2.K/mm. Comments: 1. The temperature gradient is quite large: to experimentally measure the temperature profile,a probe of size much Jess than 1 mm would be required for precise results. 2. The value for the density of ice at 0°C is taken from Table A.2. A value of 917 kg/m? is also seen in the literature.Exercise 3-6 Given: A thin rectangular plate with T= 20+100 (x/a)°C at y = b, and the three other edges at 20°C. Required: Temperature distribution T(x,y). Assumptions: 1. Two-dimensional steady conduction. ‘The temperature distribution is govemed by the two-dimensional form of Laplace’s equation. Let @ = T-20, then wo , Fo | ae "ay 7° : with x=0, O Acsin 2 Ty =T,= 5 Agsin 7 cosh ForT, 100°C, 100= F A,coshne sin = 8cm nex 8 sing Ea, 629), Aensine= 2 J aoopin BE dx = 22 p1-eay; Bo 8 1K Hence Tx.y)=200 F SESE sin Boosh HE whereO. —__|___ = 89, Wa{0.00874 +0.00009 +... Ee Z ome Comments: 1. Notice how rapidly the series for Q, converges. 2. Can you suggest a physical system in which heat could be extracted uniformly along a surface?Exercise 3-14 Given: Long rectangular bar, one surface heated uniformly, the opposite surface insulated, the other two surfaces losing heat by convection. Required: Temperature distribution, Assumptions: 1, Steady two-dimensional conduction. 2. Constant properties. Symmetry allows us to consider 0< x 6, 3. Uniform temperature at base. ‘Symmetry allows us to consider 0 201-4 sina sinh ya Ro noe coshanb/a ‘The maximum temperature occurs at x = a/2, y= b. Trae Taat = Sy Ay CO sin tanh BE ne za Comments: 1. For b = a the maximum temperature is Tal, AF reo40s;170996) + (0.04503)(~1)(1.000) +... [0.4034 -0.04503+0.0162-0.0083-... ~ 0.3665 K © © OM @ 2. Notice how the substitution 6 = T-T,,, gives homogeneous boundary conditions for x =0,b. The x component of the solution is in terms of the multivatued sine functions, which gives an infinite series for the Fourier expansion of the prescribed function at y = b.Exercise 3-18 Given: A straight rectangular fin of thickness 2r. Required: (i) Heat dissipation for two dimensional conduction, Qop, and (ii) a computer’ program to evaluate Qop/Qup, with Qip given by Eq, (2.40). Assumptions: 1. Two-dimensional steady heat conduction. 2. T=Tp atx=O and IT/Ox=Oatx=L. y @) Let 0=T-Tp and 0, = T.~Tp, with the coordinate -« 2 -ngo-e9 system as shown. Laplace’s equation in two dimensions os eo lies, Be, Fo ere . at + ay? =0 Qa with 6(0,y) = 0; 38/Ax(L.y)= 0; 36/Ay(x,0) = 0 Q) 20) y(x,0) = (He M}82)-B,] ® From Eq. (3.23): 6(x,y) = (Beoskx+Csindx)(De7+Ee™) Applying Eq. (2), B= 0; D=E; cosiL = 0, which has roots ty = na/2L,n= 1.3.5... (x,y) = DAssin Sp eosh Soh a = 1,3,5,Tun3 Ag =2C,Dq Substituting boundary condition Eq, (3) and mara inh BE + cosh gles =0,5 Bie ‘A Fourier series expansion gives the constants A, as 40, ee ae ox [85 +1 cost or Hence the temperature distribution is 40, oy)=>, sin OE cosh Os n=1,3,5,.. * | OR tanh SE + 1 cosh SE 2LBi 2L ‘The heat dissipation is p= 2] 32 foty= aTyT) E ea a ott [ ne e 2 Bit a +! (i) From Eq. @.40) Q=KA,B(Ts—T.)tanhBL3-18 (continued) where ne |*_ fe [2w)]*_1 {Ret)*_ Bit eed. es Q BiML Qe a Bil wy 72k" (TeTe anh Bi“tanh a A simple computer program gives the following results. vL=0.05 Bi=O1 Bi=5 Bi Qap/Qp(%) | VL Q/Qw(%) n/p) 0.01 998 0.01 985 BO 0.02 99.6 0.02 98.5 74.0 0.05 99.2 0.03 98.5 74.0 o1 98S 0.04 98.5 74.0 02 973 0.05 98S 74.0 Os 942 0.07 98.5 74.0 90.2 0.08 98.5 77.0 2 844 Ol 98.5 74.0 BO 015 983 4.0 10 64.7 02 98.1 74.0 Comments: 1. The error incurred when using the 1-D model is strongly dependent on the Biot number Bi =h,t/k. For Bi= 0.1 the error is about 2%, while for Bi = 5 it is about 30%, almost independent of the geometric ratio VL. We conclude that it is the conduction at the base of the fin which is 2-D when Bi is large. 2. Fins are usually used for gas flows when h, and Bi are small. In Example 2.5, (0.002)(8)/175) = 9.1x10+. 43, Since it is 2-D conduction at the base of the fin which causes the problem at large Bi number, itis unrealistic to specify a uniform temperature T= Tp at x= 0: conduction in the wall will also be 2-D. See D. K. Hennecke and E. M. Sparrow, Int. J. Beat Mass Transfer, 13, 287-304 (1970).Exercise 3-19 Given: Glass panition in a natural convection test rig. Required: Heat flows at bottom and top assuming (i) 1-D conduction, and (ii) 2-D conduction. Bq. @.31) applies, with boundary conditions x =0,T'= Tp; x=L, T=. Let 0= ‘T-T,, then the solution is 8=Cye™ + Cpe To = 340K with boundary conditions x= 0, @ = @: x=L, = ®. @zsinh Bx + Sosinh BL) sinh BL. By symmetry the heat flows at x =0 and x= L are equal, BrA, Jan p00 BL=OL) B= (h/t) = (6.0/0.78x0.005)" = 39.22 m-); BL = (39.22)(0.04) = 1.569 ahead 22.230-7810.9)) 149 320)00sh 1.569-(300-320)] = 9.34 Win sinh 1.569 (i) The symmetry of the problem requires that midway 1 along the glass the temperature is uniform and equal to ‘AT,+Tp) = Te. Also the temperature gradient at the center- plane of the glas is zero. With some hindsight, a coordinate system is chosen as shown, and a dimensionless temperature @ = % 0/8 = (T-T.)/(To-T.) is introduced to give homogeneous boun- 1 dary conditions at x= 0 and y= 0. With dimensionless coordi- ete nates 1) = x/(L/2), & = y/(L/2), and ¢ = v(L/2), the problem ; ‘transforms into %3-19 (continued) a & =0: 39/85 =0 (2a,b) 89/05 = Big, Bi=h(L/2)/k(3a,b) Using separation of variables gives the solution as 2b nig XE (1,0) = (Asinhin + BooshAn)(CsinXg + Deosh&) From Eq. (2a), B = 0, and from Eq. (2b), C = 0, thus,with E= AD, (7,8) = EsinhAn cosh, From Eq, (3b), cota = M/Bi, or cote = 01/Bi* where MS = a, Bit = Bic, which is recognized as being identical to Eq, (3.67). Hence the roots are those given in the prob- tem statement, ‘The general solution is then on= 5 Easiness From Eq. (3a), 1= 5 Esintconas, which leads to fener sas ed E, sinh), = Saeeareaea foostygde 3 + Ty, Sindacoshas 3 2 sino, and (7,8) = z s Sink, (ay#sinc, 6050) sinbAgnCosags ‘The heat flow per unit width into the partition is § 2 2sin?o, [L2dy = 2k(To~Te) { BB atanica-ta] E ae © ae : at} Q-2 fk 5 ToT. = 340-320 = 20K; Bit = Bi C =h.t/k = (6)(0.5x107)/(0.78) = 3.846x107 Interpolating in the table, oy = 0.191, Op = 3.154, 0 = 6.289, ay = 9.429. 2sin?a, Q= 20.7820) z ee Comments: 1. The discrepancy between the 1-D and 2-D solutions is (9.34-9.26)/(9.26) = 0.8%, which is very smallExercise 3.29 Given: A sphere below the ground dissipating 500 W. ‘Required: Maximum temperature the sphere might attain. Assumptions: 1. Steady state. 2. Deep below the surface the ground is also at the surface temperature of 30°C. 3, The soil conductivity is 0.6 W/m K. ‘An upper bound on the sphere temperature can be 30% obtained by assuming that sufficient time has elapsed for QITITITIR]. asteady state to be attained. Then item 7 of Table 3.2 applies, m Am GU) S= aay ~ -cs28sy = 196 FF a =2m QekS(T-T2): Ty =T2+Q/kS Ty = 30+500/(0.6)(13.96) = 89.7°C. Comments: 1, If this temperature is below the allowable limit then the problem is solved; if it is not,a more realistic answer must be obtained, which recognizes that the ‘ground temperatures vary both with depth and with time. ‘The timewise variation is particularly complicated: there is a diumal (24 hr) component, an annual com- ponent (1 yr) and variations due to the weather. Also, even if a steady ground temperature is assumed, the time for the sphere to attain its steady value is very Jong (on the order of years!).Exercise 3-21 Given: Exhaust gases entering at 500 K, flow up a chimney at 4.50 kgs. Required: Outlet temperature if the ambient temperature is 300 K. ‘Assumptions: 1, The inner and outer heat transfer coefficients are constant up the chimney. Proceeding as in Example 3,3, the steady flow energy equation, Eq. (1.4), is applied to an element of chimney Az high, thah= AQ icy (Tl yearT) = UPA2(T.=T) where U is the overall heat transfer coefficient from the exhaust gases to the ambient air, and P is the perimeter of the chimney (AA = PAz is the elemental heat transfer area). Dividing by mc,Az and letting Az —> 0, ar_ UP SL _ WP ina 2 titey biti 500k Integrating with T = Tq at z= Oand T= Toy at 2=H, [eet cai Tao = Te{Fe-T OH ie Modifying Eq. (2.171) to apply to a square duct, fee eyes ee 300K + UPH ~ haPiH” kS * hoPoH Ee Using ‘Table 3.2, Item 5, 2 ‘ 2 _ 0.940.340. QOH) | b 09 ~ 1657 8= D531 667-0.0502 ~ 9 1 1 1 Seer eee ea UPH ~ GG.H08 * CanaEeay * G2,626 ~1O7E2205+05%) UPH = 281.8 , Tom = 300-(300-SO0)e-814/5510100) = 489K Comments: 1. Itis important for Tay. to be as high as possible; the resulting large buoyancy force will give a strong thermal plume above the chimney, and allow the exhaust 10 be dissipated high in the atmosphere, 2. Although smaller than the conduction resistance, the outer thermal resistance isnot negligible, Thus the outer surface of the chimney is not perfectly iso- thermal, and the shape factor value is somewhat in error. (Also, if a wind is ‘blowing the outer heat transfer coefficient will vary around the chimney, see Chapter 4, Figure 4.27, for the variation around @ circular cross-section chimney.)Exercise 3-22 Given: A kiln operating at 700°C. Required: Heat loss to ambient air at 25°C. Assumptions: 1. Outside heat transfer coefficient is 5 W/m? K. 2. Steady state. 3. The outside surface is approximately isothermal. ‘There are two thermal resistances in series; the overall heat transfer coefficient times area product is 25m Ag = (2x4) (4x5) = 28 mi? oa k=2.4 Wim K for zirconia brick Sis evaluated for 6 sides, 12 edges and 8 comers; using Table 3.2, Items 1, 10 and 11, 8 =4 OED 42 OO 4 4059901. 5)+8(0.54048)015)0.5) 05 = 12444324 4432406 =24.2m ey 1 1 1 UA Bea * Gaeaay ~ 0007400172 UA=41.1 W/K, Q=UAAT=, }1.1)(700-25) = 27.7 kW. Comments: 1. The outside resistance is not negligible (30% of total). Thus, the outside surface of the kiln will be too hot to touch.Exercise 3-23 Given:: The buried cylinder shown in Item 8 of Table 3.2. Required: ‘Temperature field and shape factor. Assumptions: 1, Steady heat conduction. 2. Along oylinder. 3. The cylinder surface is isothermal. 4, ‘The ground surface is isothermal and equal to the soil temperature at great depths. ‘We superimpose the temperature fields of a line source Q below the ground and a line sink -Q above the ground in a micror image position, as shown. The excess tem- a perature at any point P(x.y) is defined as ©= TT, and the excess temperature at the cylinder surface is 0, ‘TT. From Eq, (2.14), the heat flow between two con- centric cylinders of radii ry and r at temperatures T; and Tis = 2M | 261-8) Ine) Antti) ‘The temperature excess at point P(x.y) due to the line source can be written as gece eet 6, =0)- 5 ne and due tothe line sink, 2-645 - n =O + Se ‘Since the heat-conduction equation is linear, superposition can be used to get the temperature excess @p at any point P(x),3-23 (continued) oO x ice. circles with origin b. ria 5 and radii oie Notice that the centerline of the cylinder must be below the line source since (1-+c)/(1-c) > O fore > 0. For an isotherm on a cylinder of radius r; at depth h, = 200% le @) ® Equations (2) and (3) fix a unique value of c corresponding to r; and h, Substituting Eq, (2) in Fa. @), e+ p-« Solving for ¢ and taking the positive sign since T, has to be larger than T>, befor TT RT aT ‘Substituting in Eq. (1), 2nkL(T\-T2) 2rkL(T-T2) TT. 4 oT) fe ley] =F] no [tn n Comparing with Eq, (3.32) the shape factor is S = 2xL/cosh™! (h/t) \ Forh> 3n, the shape factor can be simply written as S = 2nL/in(2hvny) Also, for y= 0 (the surface of ground), ¢= 1, which means that it is an isotherm. Comments: 1. This solution is based on an analysis given in Analysis of Heat and Mass Transfer, E. R..G. Eckert and R. M. Drake, Jr., McGraw-Hill, New York (1972). 2. Do not confuse the depth of the line source, a, with the depth of the cylinder, h. 3. ‘The choice of radius r, and temperature T; to define the strength of the line choice were arbitrary, and are not requited to relate to the cable itself: we could have just as well used ro and To, since both cancel as the analysis proceeds.Exercise 3-24 Given: Oil pipeline 200 m long, 8 om ©.D., encased in 15 om square concrete and submerged in sea water at 10°C. Required: Tous for Tig = 120°C and th = 0.6 kgs. Assumptions: 1, The only significant thermal resistance is the concrete, 2. Steady two-dimensional conduction. Proceeding as in Example 3.3, tgp (Tl esax—T ba) = KASCTe—T) where AS = 2xAx/In(0.54a/1) from item 6 of Table 3.2, since a/2r = 15/8 > 1 as required. Substi- tuting for AS, dividing by rc,Ax, and letting Ax —> 0 av 2nk at ___?sk _(q-)=0 dx the, mosey ) Integrating with T=Tjp atx=Oand T= Tyo: atx =L, Tog =TeTe—Tin)e T 2a 2900 198) mk S="qo.stalt) ~ In(0.54x15/4) ioe Tyg = 10-(10-120)¢rHN7EDIOHAWO — 73,°C 1.4 Wim K from Table A.3. Comments: 1. Since the inner and outer convective resistances are relatively small, the shape factor should be quite accurate.Exercise 3-25 Given: Oil pipe buried 1 m below ground surface level. Required: Oil temperature drop in K/m. Assumptions: 1. The only significant thermal resistance is that due to conduction through the soil. 2, Steady state. 3. ‘The deep ground temperature is also 25°C. Item 8 from Table 3.2 applies, 1 Boas 44471 a 3n 2nL ae (2n)(1) te S=Thqanyny ~ Tapso.o7sy 24m for 1 ming Using Eq. (3.32), the heat loss per meter length of pipe is Q=KS(Tpige—Totace) = (1)(1.914)(95-25) = 134.0 W/m ‘The mass flow rate of oil is th = (0.5)(800)/60 = 6.67 kg/s. _ 2 134 2 a aT na EENAIS) = 1.03x10* K/m (~1K/100m) Comments: 1. Since the oil temperature is so much higher than the soil temperature, the expected variations of soil temperature with depth and time are of secondary importance. 2, The k-value of soil is strongly dependent on its moisture content.Exercise 3-26 Giver: 0.03 m diameter insulated power cable, buried 1 m deep. Required: Allowable power dissipation if the insulation temperature must not exceed 350 K. Assumptions: 1, Steady state, 2. Both the surface and deep soil are isothermal at 300 K. 3, The soil conductivity is 1 W/m K. 300 « ‘The shape factor is given in Table 3.2 as item 8. ee Vth h/3n, = 1/G0.015)=22.2>1 s: 2nL heim inhi) 7 =350k Substituting in Eq. (3.32), & ae Q=KSAT = 2RKAT__ 2€8)(1)(350-300) aro Smee 682 Ne Comments: 1, Often buried power cables are subject to large current variations over a typical 24 hr period. Since the thermal response of soil is relatively slow, the above estimate of Q is conservative. 2. Uncertainties in soil conductivity are a major problem when rating buried power cables. A useful reference is “Backfill materials for underground cables", EPRI EL-506, Electric Power Research Institute, Palo Alto, June, 1977.Exercise 3-27 Given: An oil line buried underground, Required: Heat loss when the oil is at 90°C, @) ignoring the inside thermal resistance, and (ii) ‘accounting for an inside heat transfer coefficient of 300 W/m? K. Assumptions: 1. The soil temperature some distance from the pipe is steady and uniform at 10°C (including the ground surface). @ The shape factor is given by item 8 of Table 3.2; B/Bry = 60/(3X12.5)= 1.6>1 =2.78 for 1 mength QE KSAT = (0.8)(2.78)(90-10) = 178 Wim y= 125m (i) With an inside heat transfer coeflicient of 300 W/m? K, if 1 * (emy(0.125))G00) * (0.8)2.78) = 0.0042 +0.450 UA=2.20 WK Q= UAAT = (2.20)(80) = 176 Win Comments: 1. The inside convective resistance is negligible, 2. The k-value of soil depends strongly on moisture content (see Table A3).Exercise 3-28 Given: Saturated steam flowing through a 20 cm O.D. steam pipe which is buried 1.5m below the ground. Required: Steam condensation rate per 100 m of pipeline. Assumptions: 1. The deep ground temperature is also 15°C. 2. Steady state. 3. The only significant thermal resistance is that due to conduction through the soil. ‘Equation (3.32), Q= kSAT, can be used for the heat loss, with the shape factor obtained from Table 3.2, item 8, 2k _ 2x)(100) S=Trawn) ~ inna ~ "847" From Table A.3 we find that for dry soil, k= 1.0 Wm K. From Table A.12a at Pra = 1.1x10° Pa, Tea = 375.3 K, byy = 2251x108 Ice. 1.0)(184.7)(375.3-288.2) = 16,090 W ‘The steam condensation rate is now, Tieendenne eg = Q, Tondensate = Ob = 16,090/2.251x10° = 7.15x10 kg/s Comments: 1. The condensation rate is ~0.7% of the steam flow. 2, Notice that because the steam is saturated, its temperature is constant along the pipeline, which makes the calculation of Q simple. We do not have 10 use the single- stream heat exchanger equations. 3. The thermal conductivity of soil depends strongly on its moisture 2.0 W/m Kis more appropriate. content: for wet soil a value ofExercise 3-29 Given: A cubical oven insulated with 6 cm of medium density fiberglass. Required: Power supply required to maintain an interior temperature of 440 K when the ambient temperature is 20°C and b, = 7 W/m? K. Assumptions: 1. Steady state. Ifthe outside thermal resistance were negligible we could use Eq. (3.32), Q = KSAT to obtain the required power supply directly. To include the outer convective resistance we use the following thermal circuit. ‘Now, the heat loss is given by Q = AT 1/kS+ 1/hgA where k = 0.057 Win K (at 370K from Table A.3) and h, = 7 W/m? K. ‘The shape factors given as items 1, 10 and 11 of Table 3.2 should be used, assuming independent parallel paths for heat flow through the 6 sides, 12 edges and 8 comers of the enclosure. ‘Thus, if L is the insulation thickness and W the inside edge, S= [6W?/L4(12)(0.54)W4(8)0.1)L] = [(6)(0.20)? /(0.06)+(12)(0.54)(0-20)+(8)(0.15)(0.06)] = [441,296+0.072] = 5.368 m A= 6 (W42L)? = (60.32)? =0.614 m* ‘The required power supply is then, 6 440-293 147 = TAOS 368)+ 1110614) 3.2740233 2.0W Comments: 1, In practice the outside heat transfer coefficient would not be constant over the surface of the oven, and indeed, would be rather difficult to estimate accurately. Fowever, the outside resistance is only about 7% of the total resistance, so that ‘we should not be concemed.Exercise 3.30 Given: Hot water flowing at 500 gallons/min through a 20 cm O.D. pipe buried 1 m deep, with. the ground surface at 10°C. Required: Length of pipe required to lower the water temperature by 1°C. Assumptions: 1. Steady state. 2. The deep soil temperature is also 10°C. ‘The heat loss is equal to the enthalpy given up by the water. From Table A.8 at 70°C = 343 K, p 178 kg/m?, cy = 4186 J/kg K, Q =the, TaTou) (500 gal/min)(1 min/60 s)(1 m? /264 gal)(978 kg/m?)(4186 J/kg K\(1 K) 29x10" W To find the required length of pipe we use Eq, (3.32) KSAT, S = Q/KAT = (1.29x10°)/(1)(60) = 2150m_ with the shape factor obtained from Table 3,2, item 8: b 1 we Oop 38?! 2nL a Hence, S= TOR iy» L=S In@h/ny)/2n L = (2150)In(2/0.10)/2x = 1025 m Comments: 1. Although heat is being lost at the rate of ~100 kW, the pipe length fora 1°C. ‘temperature drop is long because of the large flow rate, 2, The thermal conductivity of soil depends strongly on its moisture content (see Table A.3).Exercise 3-31 Given: A power cable with a 20 em O.D. outer sheath buried 1.2 m deep. Required: Maximum PR dissipation if allowable sheath temperature is 60°C for (i) dry sand, (i) sand with saturated water. Assumptions: 1. Steady state problem. 2. The sheath outer surface temperature is uniform. 3. Ground surface and in-depth soil at 25°C, Table 3.2, item 8, gives the conduction shape factor for a buried cylinder of radius r, at depth h as 2m 2a cosh"(hir,) — In(2h/r,) S- for h> 3r, Equation (3.32) gives the heat flow as Q = kSAT; thus Q __ 2nkaT = ifor b> 3r L in@bi,) = 1.2m, 1, =0.1 m; hence h > 3r,. Q _ QlwK(60-25) = = 692k Lo n@x1.2/0.1) @ Dry sand: QUL = (69.2)(1.1) = 76.1 W/m (ii) Saturated sand: Q/L = (69.2)(2.0) = 138.4 W/m Comments: 1. Tt has been assumed that the conductivity of the soil outside the trench has a ‘minor effect on the result. A more precise result will require numerical solution of the heat conduction equation allowing for a different value of k in the surrounding soil. 2, The practical problem of establishing allowable I°R dissipation rates also requires ‘consideration of transients associated to variations of electrical current, and of daily weather and seasonal thermal environment variations. 3. Data for the thermal properties of back fill may be obtained from: Mitchell, J. K., Kao, T.C., and Abdel-Hadi, O. N., "Backfill materials for underground power cables", EPRI ELSO6. Flectric Power Research Institute, Palo Alto, CA, June 1977.Exercise 3-32 Given: Slab with step change in surface temperatures and a uniform heat source. Required: ‘Temperature response. Assumptions: 1. One-dimensional unsteady conduction. 2. Constant properties, Equation (3.5) applies; for one-dimensional conduction, o % parr ceece ke Boundary and initial conditions are LI T.0)= Tos TCLs) = Ty; IT(,1/Ox = 0 (due to symmetry) a. Let @=T-T,, then Bea a2 ‘o-T,; O(L,1)=0; 20(0,0/0x=0 2 [Now waite @as a sum of transient and steady solutions, @(%1) = (1) + 03(8) a ‘Then, if 6; and O» satisfy a0, #0; Peo Hae aE pe an with 00,/0x = 80,/0x=0 at x=0, and 6; = 6 =O at x =L, (6,402) will satisfy Eq, (2) and its boundary conditions. Solving Eq. (4b) gives, eae Ty From Eq. (3.40) the solution of Eq, (4a) is. =F Ae ornreh, AS; poet Eaerieronle ef: we ® 2k where @; must satisfy the initial condition 046.0) = (0-35) = 0 - 2 *h-3] ak which requires a Fourier series expansion. Comments: 1. If you feel the need for some mathematics practice, find the constants Ay!Exercise 3-33, Given: Equation (3.38), namely 7 “fr = 3 “gaz = Constant C Required: Show C cannot be positive or zero, @ C=0 Z=Cy; H=C.+Cn or O=A+Bn Applying boundary conditions Eqs. (3.36b) and (3.36c) gives A=B=0 = 0, and is a trivial solution, H= Cpe" + Gye" or Gn) =e (act + Be*} Applying boundary conditions Eqs, (3.36b) and (3.360) gives AB=0 and Act +Be*=0 Solving for A and B, A=B, A= 0 and we cannot proceed further. Comments: 1. C= 0 comesponds to a steady-state solution. 2. C>O comesponds to temperatures increasing exponentially with time without limit, which is physically unacceptable.Exercise 3-34 Given: A thick slab of glass initially at 350 K, one side surface suddenly dropped to 300 K. Required: Time for location 1 cm below the surface to drop 5K. Assumptions: 1. The slab can be modeled as a semi-infinite solid. From Table 3.1, ¢= 0.51x10~ m?/s for pyrex glass at 300 K. Eq. (3.58) applies x fe Ga 300-350 Using Table B.4, erfe 7 = 0.1 corresponds to 1 = 1.163 lem Shee oor Gor) 4x0.51x10%)* 1.163, Solving, t= 36.2 s. Comments: 1. At location 3 cm below the surface, the temperature change after 36.2 sis TT 0.03 erfe 0.03 _ Tato x0. 5110362)" i.e., the temperature has not changed. ‘Thus a slab thickness of 3 cm is sufficient for the semi-infinite solid model to be valid. 2, Use CONDI to examine the temperature profile in the slab at t= 36.2 s. erfe 349-0Exercise 3-35 Given: A 4 cm thick slab initially at 100°C, suddenly has one surface temperature lowered to 50°C and the other to OPC. Required: Heat flux out the cold surface at t = 000 s. Assumptions: 1, Constant properties, 2, Slab large enough to have a one-dimensional problem. Check the Fourier number at t= 2000 s. at _ (0.15x10 (2000) Fo = — 1? 0.02 = 0.750 Since Fo > 0.1, a semi-infinite solid solution is not valid. ‘The series solution must be obtained (though ‘one or two terms in the series should suffice). This situation is dealt with in Section 3.4.1 under “Nonsymmetrical Boundary Conditions". Using the results of the analysis, TT x TeTer | 0-50 x 0450 _ of) _ x eaebaeee 20 2 L. Y Age O27 cosa oak (00-251-* | = 75-25% L L To determine the A, we multiply both sides by cos(n + 1/2)rex/L and integrate with respect to X/L from -1 to +1. Then, since Joole-tetendaad}etax «0; foataettad 1 : A, = [75-25m)eosfn+t}mdn form = w/L 4 275 fone +2 man -25 fre +dfman 4 43-35 (continued) = 71521" 9 T=T,+T, = af i} 2-1" = n+) Z on codn ote x wL at. 25 asy 2 TS elle soning (FF sofa Lssi0§ e oune| = At t= 2000 s, Fo = 0.75; hence the heat flux is = 17.5[25 +150(0.0247 + 1.511x10 $+...)] = 17.5(25 +3.70) = 502 Wim? Comments: 1, At steady state, q, = 25 W/L = 437.5 Wim? 2. ‘The analysis of this problem involved superposition of the solutions of a steady problem and a transient problem. The principle of superposition of solutions of a linear differential equation is a very important one for heat conduction analysis. It is used in many different ways. See, for example, G. E. Meyer, Analytical Methods in Conduction Heat Transfer, Chapter 4 (Chapter 3 Bibliography).Exercise 3-36 Given: Ground initially at 20°C, ambient air temperature suddenly reduced to -15°C, with h, = 20 Win? K. Required: Surface temperature after 4 hours, and depth of penetration of freezing temperature. ‘Assumptions: 1. Radiation heat transfer can be ignored, 2. Constant properties for the soil. Solution using CONDI ‘The required input in ST units is: 1, Boundary condition: 3 2. Plotting option: Tvs. x 3. xrrange for plot: 0-0.02 4, Thermal conductivity = 1.2 5. Thermal diffusivity = 4x107 6. t= 14400 (4h) 7. Ty =20 8. T.=-15 9. he =20 ‘The output is: Surface temperature T, Penetration depth of freezing temperature: ~0.011 m Comments: 1. Only the top ~1 cm of soil has frozen in 4 hours. Ts = ~2.23457+10° qs = —2.55313+10" stm sw 7 veo vs veo Se wooo x (610)Exercise 3-37 Given: Semi-infinite solid with surface temperature varying periodically. Required: In-depth temperature response, Eq. (3.63). ‘The goveming equation is Eq. (3.49), a es o 1.C.: none, solution must be periodic. B.C’s: x=0: T,-Tp = (T{—To)sin wt: x =; T-> To Assume a solution of the form T = Ce"P'sin(at-qx) +D or “gy = Ce Pre0stot-ax) = —Cpe'P*sin(ot—gx) — gCeP*cos(cot—gx) x os = Cp?eP*sin(ot-qx) — q’ Ce F*sin(eot—qx) + 2Cpqe*cos(«ot—-gx) ‘Substitute in Eq, (1) and rearrange, &c0s(cat-ax) = (pe? sin(at-qx) + 2pgeos(t-gx) At @t-qx = 2nm, n= 0,1,2,...: sin(cot-qx) = 0, and hence ¢ =2pq Ab at-gx= 204-3) 9 O,1.2eui cO5(@H-Gx) = 0 and 0 = (6-42) Hence, p=q= Va/2a, T= Co? ssinot — Vofa x) +D For x—«,T=0+D; hence D=Ty ‘sin @t+To; hence C= T;-Ty GPA Xsin¢aot — Vora x) Comments: 1. Notice thet only the positive value of p was used.Exercise 3-38 Given: A thick slab with one surface temperamure varying periodically in a sinusoidal manner. Required: Amplitude and phase lag of the temperature variation 1 cm below the surface: Assumptions: 1. The slab can be modeled as a semi-infinite solid. Equation (3.63) applies. 1 11257 rad/s (0.1257(2)(2x10°°)* = 177 mt Hence, T”=To = 0.170(T;—Tp) = 0.170(200) = 34 K. ‘The phase lag is x(@/20)* = (0.01){177) = 1.77 rad = 101 degrees Comments: 1. Use COND! to examine the variation of surface heat flux with time, and to examine instantaneous temperature profiles.Exercise 3-39 Given: Air temperature change from 5°C to —10°C for 48 hours, Required: Minimum depth for a buried water pipe to prevent freezing. Assumptions: 1. Soil thermal diffusivity &=0.5x10* m?/s. 2. T, =—10°C for a conservative estimate. ‘We will assume that the ground surface temperature itself drops to —10°C for a conservative estimate of the required depth. CONDI can be used; the required input in ST units is: 1, Boundary condition: 1 or anise Plotting option: T vs. x Lhd x range: 0,1 8 ‘Thermal conductivity = 1 (not required) ‘Thermal diffusivity = 0.5x10* m*/s j. L= 1.728x10° s (48 h) 7. To =278K 8. T, = 263K ‘The output is a graph of T vs. x from which it is found that T = 273 K at x = 0.40 m (the x-range av apy can be adjusted for more precision). Comments: 1, To check the validity of our assumption that T, ~~10°C, CONDI can be run with boundary condition No. 3. For a brisk wind h, = 20 W/m? K might be appropriate; COND1 again gives x = 0.40 m. Forno wind, a value of h, = 5 Win? K might be appropriate: COND! then shows that a depth of less than 0.3 m is adequate. 2. Actually, the above result is not quite correct. The freezing “front” will keep pene- trating for a short time after the air temperature returns to 5°C. The principle of superposition can be used to obtain the exact solution (see Exercise 3-86 or an advanced text): the result is then x = 0.43 m. 3. The effect of moisture content on soil thermal diffusivity is far less than its effect on thermal conductivity. :Exercise 3-40 Given: A 5 mm thick steel wall initially at 300 K exposed to a radiant heat source at 1000 K. Required: Time for surface temperature to reach 500 K. Assumptions: 1, An inexpensive steel-AISI 1010 carbon steel with properties at 400 K. 2. Re-radiation and convective heat loss negligible. ‘An appropriate model must be chosen. Based on Example 3.6 one is tempted to treat the wall as a semi-infinite solid. But let us estimate a penetration depth. Following Comment No. 3, & - T-Tya From Tables A.lbg at 400 K, k = 59 W/m K, p = 7830 kg/m’, c = 487 kg K. To caleulate the absorbed radiant energy we take an absorptance of 0.9 for a heavily oxidized steel surface, and neglect re-radiation since the surface temperature is low compared to the source temperature; then A6T* = (0,9)(5.67x10 1000)" = 51.0x10° W/m? 4 ‘Then for T, = 500 K, 8, ~ (500-300)/(51.0x107/59) 0.23 m = 23 om Bot the firewall is only 5 mm thick! Next we see if a lumped thermal capacity model is appropriate. An effective heat transfer coefficient is he JAT,, = 51.0x107/(100-400) = 85 Wim" K. If the heat flow out the back face is negligible, the characteristic length is the wall thickness. For V/A Bi = bL/k = (85)(0.005)(59) = 0.007 A lumped thermal capacity model is adequate. Following Section 1.5.1, the energy conservation principle applied to the wall with negligible convective heat loss and re-radiation gives .5 om, ar a7. 4 ve Tega, cae Pecarane fame a pel Integrating, T-Ty = (q/pcLt, or t = (T-T,)peL/q, ‘Bence, t = (500-300)(7830)(487)(0.005)/51.0x10* = 74.8 s Comments: 1. The natural convection heat transfer coefficient is at most 5 Wim? K, cf. the effective radiation heat transfer coefficient of -85 Wim? K. Hence natural convection can be ignored in obtaining this estimate of the required time. 2. Strictly speaking q, should be denoted qu — see Eq. (6.54).Exercise 3-41 Given: Semi-infinite solid with a step-change in surface temperature. Required: Sotution for surface heat flux using the approximate integral method. Assumptions: 1. One-dimensional heat flow. 2. Constant properties. @) To determine the constants for T = A + Bx + Cx? three boundary conditions are required. T=T, atx=0 T=Tp at x=6 (since temperature changes do not penetrate further) AT /x=0 at x = 5 (10 give a smooth profile at x= 8) Hence, a To : ° 3 OT = -kCT-T) & fi <2 oF no kT) [ 3] jro= 3 rt) @ 3 z 2 aft 2 Of iy A scer,t,) & @ at J pccr-toyax= pects-To) J [: ;| a= 5 pol Ta) “Ge @ Equating Eqs. (2) and (3), a xd 3 PCTs To) “Ge 65 _ Gu hence, {2 = @ (iii) Integrating Eq. (4) and solving for 3 gives 5 = 2(30)”. Substituting back in Eq. (2), = 2 gat) = TO) oy oe Gan Comments: 1. This value of qs is (x/3)" = 1.023 times the value given by Eq, (3.59). 2, The integral method will be used to solve heat convection problems in Chapter 5. 3, Although approximate, the integral method often gives satisfactory results, and has ‘been widely used for engineering analysis.Exercise 3.42 Given: Water pipe buried to prevent freezing. Required: Necessary depth if'T, varies between —10°C and 12°C, and T.. Assumptions: 1. Simusoidal variation of ground surface temperature. 1°C. Eq, (3.63) applies, T-Te TC) = = oO" sinfat-x(0/20)*} To Examination of the sketch shows that we are interested in the depth at which the amplitude has decreased to °C externas Tt (@ = 2n/(24)(3600) = 7.2710" rad/s ‘The thermal diffusivity of soil varies significantly with type of soil and moisture content: from 2 Table A.3 a value of 0.4x10°° m?/s is typical x11 2107 2y04t10°)* Comments: 1. Proper formulation of "buried pipeline” problems is difficult. Ground surface temperature variations can be erratic due to the weather. The in-depth temperature profile is not steady and uniform but varies on an annual cycle, ‘The soil thermal diffusivity cannot be specified precisely. Thus the prudent engineer will apply an appropriate safety factor to the result obtained above. 2. Perform a sensitivity study on the result, ie., vary parameters such as 0. and T.. (the mean of T, must also be changed accordingly) to see their effect on the required depth.Exercise 3-43 Given: A thermocouple installed 1 mm below the surface of the cylinder wall of a 2-stroke I.C. engine operating at 2500 rpm. Required: Amplitude and phase difference of surface temperature. Assumptions: 1. The temperature variation is sinusoidal with time. 2. The cylinder wall can be treated as a semi-infinite wall. Eq, (3.63) applics. e705)" sing eot—x(@/200)"] Thermecounle (@=2n(2500/60) = 261.8 rad/s is 4 i 2) = atts 7 “ombustion [| | (2)(12.0x10*) os ‘oaner If T* is the amplitude of the temperature variation then at x = 1 mm, THT . *—To. _ e-a(ortay* — ¢ 2000309 - 3,678x107 T-To Peer teks -tete eet 3.678x10% — 3.678x10 Thus T;~To 9.4°C. ‘The phase difference is x(@/20)” = 3.303 rad = 189°. External environment, Comments: 1. Use CONDI to examine some instantaneous temperature profiles.Exercise 3-44 Given: Surface of a thick AISI 1010 steel plate exposed to a heat flux of 50,000 W/m”. Required: Plot of surface temperature versus time for 10's. Assumptions: 1. The firebrick wall is a perfect insulator so that all the electrical heat flows into the steel. 2, ‘The steel plate can be modeled as semi-infinite for the first 10's. From Table Ala, properties of AISI 1010 steel at 300 K include k = 64 W/m K, a= 18.8x10°% m2. Solution using CONDI ‘The required input in SI units is: 1, Boundary condition: 2 2. Plotting option: T, vst 3. trange for plot: 0, 10 4, Thermal conductivity = 64 5. Thermal diffusivity = 18.8x10°¢ 6. Ty = 300 7. qg = 50,000 COND! provides the following plot. Ts (10) Se & im wo a a0 66 750 7% 1000 Comments: 1. Since the thermal diffusivity and conductivity of firebrick are about 2% of ‘that of the steel, the heat loss into the brick will be comparatively small. 2. Use CONDI to give the temperature profile T(x) at time t= 10s. Hence estimate how ‘thick the plate should be for the semi-infinite solid model to be valid,Brercise 3-45 Given: Measured phase difference for a long rod with one end heated to give a sinusoidal ‘temperature variation. Required: Thermal conductivity of the rod. Assumptions: 1. Negligible heat exchange with insulation, From Eq. (3.63) the peak temperatures occur when sinfot-x(@/20)*] = 1 ha Lion. jLocm wa( Now te-t, = 15 min=900s ee Xp-%q =0.-0.1=0.6m © = 21/909 = 0.069 A wont ‘The thermal conductivity is then k= poo. = (8300)(470\3.22x10°) = 12.6 Win K Comments: 1. Since itis difficult to ensure that the sides of the rod are perfectly insulated, an altemative approach might be to have the rod transverse to a flow in a wind tunnel. ‘A fin-type analysis with a sinusoidal base temperature would then be required. 2, Notice that t, and ty must be measured accurately because errors in the tare magnified by taking a difference and then squaring.Exercise 3-46 Given: Locations where the soil is @) dry, and Gi) wet. Required: Depth below surface where the temperature variation is 10% of the surface variation, for both annual and diumal variation. Assumptions: 1. The components of the surface temperature variation are sinusoidal. @ Dry soil. Equation (3.63) applies, T-To Ti-To and we see that the amplitude decays as exp[-x(2@/a)“]. Hence, O1= eras x=—-Qofay*ind.1 For the annual variation @ = 2n/1 yr=2n/(365)(24)(3600)= 1.99x10~ rad/s. From Table A.3, = 0.35x10 ms. Hence x =-12(0.35x10)/1,99x107 Jin 0.1 = 4.3m For the diumal variation, @ = 2n/1 day = 2n/(24)(3600) = 7.27x10°> rad/s X= -12(0.35x10°)/7.27x10 >in 0.1 = 0.23 = e220 sini ot-x(/20)") (i) Wet soil. From Table A.3, or= 0.5x10~* m?/s. Proceeding as above, ‘Annual variation: Diumal variation: Comments: 1. We can sce why winemakers often use caves to store wine. A constant cool tempera- ‘ture environment is essential for the proper maturing of fine red wines, particularly those made from the cabemet-sauvignon grape. 2. Notice that the thermal diffusivity of soil depends far less on moisture content than does the thermal conductivity.Exercise 3-47 Given: A YSZ coated Inconel-750 component initially at 30°C, suddenly exposed to hot gases at 800°C with h, = 180 W/m? K. Required: Surface and interface tempertures as a function of time, finite solid. Assumptions: 1, Model as an semi 2. Assume quasi-steady conduction in the coating, ic., a linear temperature profiles. An overall heat transfer coefficient including the resistance of the coating is 11,8, 1 , 0.0003 3 Unt ws 56x10740.06x10 U=178 Wim? K (The thermal resistance of the coating is seen to be almost negligible.) Solution using COND1 ‘The required input is: ‘Surface boundary condition = 3 (convection) Plotting option = 2 (T, vst) ‘trange for plot = 0, try 100s k= 13.5 (from Table A.1b evaluated at ~400 K) = 3.35x10° T.=30 ‘The CONDI plot is shown below. Since 8/k « 1/h, the interface temperature can be taken to be the same as the surface temperature, 300 SSeS & vs am we sb 6m 67 1000 G10Exercise 3.48 Given: Annual variation of ground surface temperature. Required: Depth for maximum temperature in midwinter. Assumptions: 1. The annual component of the ground surface temperature is sinusoidal. For a periodic surface temperature variation, Eq. (3.63) applies, T-Ty e™20)" sinfext-x(@/20)'] = & P*sinfa-Bx] where 2 o- —* ___ ~ 20107 rads, B = (2. )” =|_208107_] - 0.2887 (365)(24)(3600) 20. (2)(1.2x10 ©) This equation gives a maximum surface temperature (summer solstice) at wt = (2n + L/2}r; for = 0, t=3 months = 7.884x10° s, Six months later (winter solstice) t = 9 months = 2.365x10" s. Extrema in the temperature profile can be found by setting 9T/0x = 0, bal = (1 -TpI-Be *sin(ct-fx) + ¢ Peos(ct-Bx\(-B)] or, tan[e@t-Bx] +1 = 0. Substituting for @ and B, tan[(2.0x10 “)(2.365x107) - 0.2887x] = -1 Solving, 0.2887x = 4.730-G/4)m, for the first maximum x =822m ‘The percent decay in the amplitude is then (1-e *)100) = [1-e ~9288%822)((199) = 90.7% Comments: 1. Note that tang = -1 for ¢ = (a - W/4)n; n = gives the first maximum value, 2. Use CONDI to plot T(x) profiles for various times, e.g., t = 1,2,3...6 months. ‘The plot for t = 6 months is shown below.Exercise 3-49 Given: Forearm (A) at 32°C suddenly contacting a slab (B) at 100°C. Required: Skin surface temperature if the slab is ()) aluminum, (ii) 18-8 stainless stee!, Gil) Pyrex glass, or Gv) Teflon. From Eq. (3.64) Ta-Ti ke [ou]*_[p08]* a ree (“ae Ty = 32°C, ka = 0.37 Win K, ot =O.1x10% m/s 7) @ Aluminum =210 Solving, T; =96.9°C. Gi) 18-8 Stainless Steel 00°C, ke = 17 Wim K, og = 4.20510 m*/s a7 [ orto |" _ 199 0.37 | 4.20x10%} = 916°C. Gi) Pyrex glass 100°C, kg = 1.09 W/m K, ag = 0.51x10 m/s Ye 1.09 | 01x10 ‘037 [osixio*| ~'? Solving, T, =70.5°C (iv) Teflon ‘Tp = 100°C, ky = 0.42 W/m K, og =0.18x10~ m/s 3201 [ees Page T-100 0.37 | 0.18x10% : Solving, T; = 632°C. Comments: 1. A temperature of ~48°C is commonly accepted as a threshold for bum injury. None of the above surfaces are safe to touch,Exercise 3-50 Given: Two samples of bologna meat initially at 8.2°C and 23.1°C brought into contact. Required: Temperature response in room temperature sample 6.3 mm from interface. Assumptions: 1. ‘The samples can be treated as semi-infinite solids for short enough times. 2. The heat conduction equation, Eq. (3.6) applies to transient heat conduction in bologna meat. Equation (3.64) gives the interface temperature. Since the thermal properties of both solids are the same, the interface temperature is simply the average of the two initial temperatures, and is constant, T, = O.5)(T,+T,) = (0.5)(23.1+8.2) = 15.7°C. ‘The temperature response is then given by Eq. (3.58) as T-T, 2 = erfe_* iTo Gay? ‘Ty = 23.1°C, T, = 15.7°C, x = 63x10" m, o = (0.8)(1230)(4660) = 1.40x107 ms, Hence, T = 23.1-7.4erfel8.42t 12] ” 7 1 Using Table B.4, q T | t(s} | 842r"? | erfen | TC a | L 2 10 | 2.66 o.oo | 23.1 a 50 | 119 [0.092 | 224 > 100 | 0842 fo.234 | 214] 8) 150 | 0.687 |o330 | 207 i A 200 | 0595 Jo339 | 201 7 i - 300 | 0.486 Jo.492 | 19.5 a ECE : ie 500 | 0376 joss | 187 ‘The graph shows the comparison. Comments: 1. The actual temperature response shows evidence of a "penetration time” for a “thermal wave" to reach the thermocouple. On the other hand, Fourier’s heat ‘conduction equation implies an infinite speed of propagation of the thermal wave to give a parabolic partial differential equation. This non-Fourier behavior is attributed to the nonhomogeneous inner structure of processed meat and the role played by the moisture content--but the physics is not well understood. See also: Kaminski, W., nonhomogeneous inner structure”, J. Heat Transfer, 112, 555-560 (1990). iyperbolic heat conduction equation for materials with aExercise 3-51 Given: Jet of air at 45.0°C impinging on a plastic plate coated with liquid crystals. Required: Convective heat transfer coefficient if time to color transition is 9.3 s. Assumptions: 1, Liquid crystals color-transition temperature is 38.0°C. 2. Conduction in the plate can be modeled as one-dimensional in a semi-infinite solid, 3. Constant heat transfer coefficient and thermal properties. 4, ‘Thermal resistance of paint layer negligible. The low thermal conductivity of the plate and gradual variation of the heat transfer coefficient underneath the jet suggests that conduction along the plate can be neglected. Also, since the time is short we can treat the plate as a semi-infinite solid. Then Eq. (3.61) applies, THTo Le efe _* _-e tomrtdoten ee] ox = (at)'? eto (4on)!? Gon ay T, = 38.0°C, Ty = 20.0°C, T, = 45°C, k = 0.24 Wim K, 0 = 0.12x10¢ m/s, t= 9.3 s. Since erfe is a transcendental fanction, Eq. (1) should be solved by iteration using Table B.4. Alternatively, CONDI, item 3, can be used by guessing h, and checking t for T, = 38.0°C. A few iterations gives h, = 406 W/m? K. Comments: 1. Also 18 COND we can check the penetration depth to see how thick the plate should be for the semi-infinite solid assumption to be valid. ‘The penetration depth is seen to be about 4 mm; thus, a plate thickness of 1 cm should be sufficient. 2. Use of the liquid crystal method is described by J. W. Baughn and X. Yan, “Local heat transfer measurements in square ducts with transverse ribs", J. Enhanced Heat Transfer, 202, 1-7 (1992). 3. A good error analysis of this technique is given by: B. Tan “Design and testing of an automated system using thermochromatic liquid ‘erystals to determine local heat transfer coefficients for an impinging jet,” M.S. Thesis, School of Engineering and Applied Science, University of California, Los AngelesExercise 3-52 Given: Proposed method for measuring convective heat transfer coefficients using the temperature response of a copper rod coated with a known Tyyp material Required: (j) Heat transfer coefficient if Ty = 25°C, T, = 300°C, Typ = 55.5°C, and the time to melting t = 400 s. Gi) Evaluation of the use of an AISI 302 stainless steel rod. Gi) Evaluation of the use of a Teflon rod. Assumptions: 1. One-dimensional conduction in a semi-infinite solid. 2, The heat transfer coefficient, and solid properties are constant (evaluated at 300 K). ‘Substituting x = 0 gives the time response of the surface temperature, h, 1-0 Werf 5 fan" o T, = 55.5°C, Ty = 25°C, T, = 300°C. From Table Ala for pure copper, k = 401 W/m K, cc = 117x10% m’/s, and 1=400s. Since erfe is a transcendental function, Eq. (1) should be solved by iteration using Table B.4. Alterna- tively, CONDI, item 3 can be used by guessing h, and checking t for T, = 55.5°C. A few iterations gives h, = 200 Wim? K. While using CONDI we should check the penetration depth to see how long the rod must be for the semi-infinite solid assumption to be valid: a value of 8, = 0.8 m is found! Hence a rod length of at least 1 m is necessary, and surely the use of copper was a poor choice. ii) For AISI 302 stainless steel, Table A.1a gives k = 15 Wm K and o = 3.88x10° m/s. Using CONDI, item 3, with h, = 200 W/m? K as a known input now, the time to melting is 16.25 s and 8, = 3 cm. Thus stainless steel is a more practical choice-the elapsed time is long enough for an accurate measurement, and the required rod Iength is reasonable. Giii) For Teflon, Table A.2 gives k= 0.35 W/m K, 0.= 0.15x10° m’/s. Again using CONDI with h, = 200 W/m? K as an input, the time to melting is 0.24 s and 5, = 0.8 mm. The small penetration is attractive for some applications, but the time might be too small to measure3-52 (continued) accurately--a finite time is required to initiate the gas flow. However, forced convection heat transfer coefficients depend on temperature only through property variations, and there is no good reason to use T, = 300°C; indeed it may be easier (o use a lower temperature, For T, = 80°C, t = 16.8 s and 8, = 6 mm. Comments: 1. Experimental procedures should be designed before construction and use. That is, engineering calculations should be made to predict the system behavior, and the system parameters chosen to give adequate results. 2. As another example of a design calculation, consider the effect of the coating. If the thickness and thermal conductivity of the melt material are estimated to be 0.5 mm and 0.5 W/m K, respectively, the thermal resistance of the coating is 0.5x10°%0.5 = 10° K/W. The convective resistance is 1/200 = 5x10° K/W; thus the coating resistance is 20% of the convective resistance. If high convective heat transfer coefficients are to be measured accurately, the coating must be very thin.Exercise 3-53 Given: Tire carcass 2 cm thick, initially at 20°C, exposed to steam at 150°C. Required: Time for centerplane to reach 130°C. Assumptions: 1. Model as a plane slab. 2. Thermal resistance of condensate film negligible. ‘The temperature response chart, Figure C.1a, can be used. 2 __@.01? | -E- = 1667s com LZ) seam e= = Doexlo* stear Sine Z/) st Be Tr1sot x,-150t TT ae 130-150 _ 20 _ 454 “ToT, 20-150 ~ 130 Using Figure C.la, Fo = 0.87 2b0.02m 1 = Fot, = (0.87)(1667) = 1450 s (24 min) Solution using COND2. ‘The required input is 1 (slab) Bi = 1000 (large) 1 (numerical output) Fo range and increment = 0.85, 0.95, 0.001 ETA range and increment = 0, 0,0 ‘COND? gives Fo = 0.858, t= 1430s (24 min) ‘Comments: 1. For Bi = 200, COND? gives Fo = 0.865, which is 0.8% different to the value used. The precise value of Bi is clearly not important, provided that it is large, 2. Since Fo > 0.2 the one term approximation can also be used for this problem.Exercise 3-54 Given: A'S om thick slab with an insulated back face; Tp = 700 K, T. ='300 K, h = 2000 ‘Whn? K. Required: Back face temperature after (i) 5 minutes, and (ii) 1 hour, for AISI 1010 steel and chrome brick. Estimate properties at T, = (700+300)/2 = 500 K, and use Tables A.1 and A.3, COND? can be used to give the desired results. Material - a Bi t Fo 8(n=0) T Wim K | m?/sx10° s K AISI 1010 54 13.3 1.85 | 300 | 1.60 | 0.199 | 380 ‘Steel 3600 | 19.2 | 0.000 | 300 Chrome 24 0.92 41.7 | 300 | 0.11 | 0.942 | 677 Brick 3600 | 1.32 | 0.0570 | 323 ‘Sample solution for AISI 1010 steel at t= 300 s using COND2: the required input is 1 (slab) Bi= 185 1 (aumerical value) Fo range and increment = 1.60, 1.60, 0 ETA range and increment = 0, 0,0 COND? gives 8 = 0.199 Comments: 1. Check the values of 0 using the temperature response chart, Fig. C.ta in Appendix C.Exercise 3-55 Given: A 1 cm diameter sphere initially at 320 K, suddenly immersed in air at 280 K with bh, = 100 Wim? K. Required: ‘Time required for sphere to lose 90% of its energy content for four materials. Evaluate all properties at 300 K using Tables A.1, A.2 and A.3, Use COND2 to obiain the solu- tions. The results are summarized below. ‘Material k a Big=hR/k | t.=R?/o | Fo(@=-0.9) | 1 ‘Wim K | m?/sx10* 8 8 Cork 0.043 0.16 11.63 156.3 0.244 38.1 Teflon 0.35 0.15 1.43 167 0.705 18 AISI 302 1S 3.88 0.0333, 644 23.0 148 Aluminum | 237 on 0.00211 0.257 365. 94 ‘The solution for aluminum can be checked using the lumped thermal capacity model Eq. (1.38), [x], TH, 7] Tote Hence Fo =In(0.1)/(-3)(0.00211) = 364 1=t,Fo = (0.257)364) =93 s Comments: 1. Notice the factor of 3 difference in the Biot number used to check if the lumped thermal capacity analysis is valid (based on a characteristic length of volume/area), and the Biot number used in the COND2 analysis (based on sphere radius). 2. Even though cork is a good insulator, it cools most rapidly: is this result contrary ‘to your physical intuition?Exercise 3-56 Given: $Aq008 2qn = 1 where cot = 2/Bi a Required: Formula for Ag, Eq. (3.69). ‘Multiply both sides of the expansion by cos 2,7 and integrate from 0 to 1 1 1 v+ focos?Aaman ae + fAmcos AqNCOS Aad + + = dandn a) 1 feos tandn= oe { SIMQ% Pa), Sioa +adN]* fosrnontr| 20ete) 20mm) Jo = SA q—Aa) SiO e) 2G Im)” 2A) Sif. Ag 008 Ay COS Dog Sit COS Ra +008 Ae 20ea—An) 2Qqtn) _2msin IegCOS Ia —An COS Am SDL Ay, Gia + using Eq. (3.67) Bi COS yy 608 y~Bi COS 608 Oe-Ae) n sinyncostyn]” 1, Si %ycO5 hy fneendn= [2 seen a ae from standard integral tables. Substituting in Eq. (1) 1, Sn Agc0s iy} _ sin hy Seaee dein Agact afb Da | Iq + NOR Ae sin 29005 =0Exercise 3-57 Given: 5 mm thick alloy steel rocket nozzle. ‘Required: Temperature response on the inside of the nozzle. ‘Assumptions: 1. Nozzle wall can be modeled as a slab of thickness 21. = 10 mm, 2. One-dimensional unsteady conduction. 3, Negligible heat loss from outer surface of nozzle. ‘We first calculate the time constant for the heating process: t =L?/o.= (0.005)?/12x10 = 2.15 ‘Next we calculate the Biot number Bi = hL/k = (6000)(0.005)/43 = 0.70 ‘Since Bi > 0.1, the lumped thermal capacity model should not be used. The maximum allowable temperature will first be attained at x = L (n= 1), and the desired value of @ is, 1300-2100 300-2100 Solution using COND2 ‘The required inputs are: Geometry = 1 (stab) Bi=0.70 ‘Output option = 2 (@ vs. n plot) Fo = (must guess and iterate) Ty range=0, 1 A few iterations will give Fo = 1.04 for 6 (1 = is t=1,Fo= (21M 1.04)=2.2s = 0.444. Thus the allowable duration of firing Comments: 1. Since Fo > 0.2 we could have used the one-term approximation Eqs. (3.75) and (3.76). ‘Then 8 = Gf QM); 8. = Arexp(-AFFo). From Table 3.5 for Bi = 0.70, A} = 0.5615 and A, = 1,092. Solving for @., 0, = @/fy (21) = 0.444/cos(0.5615)* = 0.606. Solving for Fo, 0.606 = 1.092exp(-0.5615Fo), Fo = 1.05, in close agreement with the result given by COND2. 2, The allowable duration of firing is very short. Liquid-fueled rocket motor nozzles are usually cooled by passing the fuel through tubes on the outside surface of the nozzle, before injection of the fuel into the motor.Exercise 3-58 Given: Air flowing through a bed packed with 6 cm diameter iron spheres; h. = 80 W/m? K. Required: ‘Time elapsed for spheres at inlet to lose 90% of their available energy. 16.7x10° m?/s, We first calculate the 50 W/m K, Assuming cast iron, Table A.1 gives Jumped thermal capacity Biot number, HARI) _ (8010.03/3) _ p16<0.1 i y % - Kk ‘Thus the lumped thermal capacity model is valid. Referring to Chapter. 1, Section 1.5, the time constant is 1, =K(R/3)/oh, = (50)(0.03/3)/(16.7x10-*(80) = 374 s From Eq, (1.38), 0.1. = exp(+4/374) 618. Comments. 1. Always first check to see if the lumped thermal capacity model is valid. 2, Note that the Biot number in the criterion Bi < 0.1 is based on the characteristic length V/A (= R/3 for a sphere). 3, ‘The temperature response chart or COND2 could be used for this problem: check to see that the same result is obtained.Exercise 3-59 Given: A 2m thick ceramic slab with a metal sheath exposed to a hot air stream; h, = 800 W/m? K, 1h, = 1600 Wim? K. Required: Time for center to reach 1200 K for To = 300K, T, = 150K. Assumptions: 1. The heat transfer coeflicient is constant over the slab surface. 2. Both sides of the slab are sheathed and exposed to the air stream. ‘The contact resistance and convection resistance are in series, (eet ieee U=533 Win? K An effective Biot number based on Uis UL _ 633X001) _ k 7 177 ~ TecTe _ 1200-1500 To-Te 300-1500 From Appendix C, Figure C.1a, Fo = 1.5 L? pci = (0.01)°(2600)(1150)/3.0 = 99.7 t= Fot, = (1.5)(09.7) = 150s Soh using COND2. ‘The required input is 1 (slab) Bi=1.777 1 (oumerical values) Fo range and increment = 1.4, 1.6, 0.001 ETA range and increment = 0,0, 0 =025 Fetal sneath COND? gives Fo = 1.423. Hence t= (1.423)(09.7) = 141.98 Comments: 1. Although less precise than the computer solution, the chart result is quite adequate for most engineering purposes.Exercise 3-60 Given: A 1 cm thick slab of hard rubber undergoing heating and cooling. Required: Duration of (i) heating stage, (ii) cooling stage. Assumptions: 1, Constant properties. 2. Large enough slab to neglect edge effects. From Table A.3: p = 1200 kg/m, ¢ = 2010 J/kg K, k = 0.15 Wim K, = 0.062x10° m*/s. (ji) Heating stage. Use Fig. C.1a with Bi = 1000 to approximate the large heat transfer coefficient for condensing steam L? _ (0.005) o008Y = 403 s & — 0,062x106 at = 2 = 012; hence, Fo = 0.95. ‘ot, = (0.95403) = 383 s (6.4 min) (@) Cooling stage. First check the lumped thermal capacity Biot number, hVIA) bel (8)(0.005) = 0.27 > 01 occas K ors ‘This value is somewhat too high to use the lumped thermal capacity model. Thus we again need to use Fig. C.la, but now we must approximate the initial temperature profile as uniform. For a suitable value we find the average temperature at the end of the heating stage. For Bi = 50, Fo = 0.95, Bi?Fo = 2370, and Fig. C.2a gives = 0.90. T-140 Hence, 8 = 1-0.90 = 0.10 = ;T = 175°C 35-25 1275-25 0.0976; hence, Fo = 9.0. Fo t, = (9.0)(403) = 3630 s (1 hour) ‘Comments: 1. The solution for part (ii) is not “exact” because the initial temperature profile is not uniform; however we are seeking a solution at a rather long time so that the exact shape of the profile will have little impact on the result, 2. Try the lumped thermal capacity model for the cooling stage-it agrees well,Exercise 3-61 Given: A 3 cm diameter rod of a composite material initially at 30°C. Required: Response for heating by condensing steam at 120°C, and cooling in 20°C air. Assumptions: 1. The rod is long enough to neglect end effects--1D unsteady conduction. 2. The condensing heat transfer coefficient is very large. Steam condensation coefficients are very large and thus the Biot number for the first stage will also be very large (the surface temperature of the rod will be very close to 120°C). Thus the lumped thermal capacitance model is not appropriate. For convenience we will use the temperature response chart form of the exact solution. Referring to Fig. C.1b we require 110-120 | -10 30-120 90 ou For Bi = 1000, Fig. C.1b gives Fo = 0.47. Since at = k/pe = (1.2)(1500)(1800) = 0.444x10° m’ls. t, = FoR %/a = (0.47)(0.015)/0.444x10 * = 238 s (3.97 min) For the cooling stage we first check the Jumped thermal capacitance Biot number, Bigyc = h,(V/AVk = h,(R/2Vk = (15)(0.015/2)/1.2 = 0.0938 Since Bi < 0.1 we can use the lumped thermal capacity method, But what initial temperature should we use? The surface is at 120°C, but the center is at 110°C. We will use the average temperature at the end of the heating process, which can be obtained from the fractional energy loss in Fig. C.2b. Using the curve for Bi = 50, Bi?Fo = (50)*(0.47) = 1175 and © = 0.93. Hence T-30 = 0,93(120-30); solving, T = 113.7°C ‘The lumped thermal capacity model time constant is obtained from Eq. (1.38) as pVchhA = p(R/2}efh, = (1500)(0.015/2)(I800V15 = 1350 s 30-20, +1350, 113.7-20 3 t= 3021s (50.3 min) Comments: 1. Check t, using COND? or the long-time one term solution.Exercise 3-62 Given: Slab L 0. Assumptions: 1, One dimensional conduction. 2. Constant thermal diffusivity Fourier’s equation Eq. (3.6) applies. grant ger: ot Ox? @ with t = 0: T = 100cos(x/2L) a t>0, x=0, Tx =0 (2b) t>0, x=L, Tx =0 (Qc) The solution of Eq. (1), subject to boundary condition (2b), is obtained from the analysis of Section 3.4.1 as TG) =e MAcosans $= oUL?, 9 =x or mo (AX sin An) Tha =e \-Adsind) = 0 from Eq. (2c) which requires sink = 0, or 4, = n,n =0,1,1,.... Thus TEA) = Ae * cos nm, @) = ‘The constants A, are determined from the initial condition, Tg = & Aecosnan = 100cos xn/2 First calculate Ag: for n = 0 1 1 {Jdo208 0.cos 0am = {100 cos mny/2.c0s0 dn a a3-62 (continued) Ag = (100)(2/n)[sin mn/2Jy, = 200/ ‘The remaining constants A,, n = 1,2,.. are given by 1 A, = (2)(100) {cos mn/2.cos nen dy 3 1 = (200,412) f 2 nm jn “El 3 1 = +n) © x +cos(1-n) 7 (202 feos 1) En teos(1-n) afm 2 (+) 200/92 9+ & r3) OS Substituting in Eq. (3), 200/, , 2 5 axe, 8 wz 2am-+5. Fa f Bettas = 1 in(1+n) = + in(1-n) (200) Dr sin(l +n) =n singly (a-Dx TEn) = 6S cos 4mmy + Comments: 1. As t > 22, -> 2 and T > 200/n. 2, Notice how Ao is evaluated as a special case, before evaluating A, for m-Exercise 3-63 Given: Convective cooling of a plane slab at long times. Required: Comparison of rate of cooling using the lumped thermal capacity and léng time approximate solutions Assumptions: 1, Constant properties and heat transfer coefficient. In Eq, (1.38) for the lumped thermal capacity temperature response, Tis the average temperature, § = te Lg thver T, For a slab of width 21, V/A = L; substituting Bi = h,L/K, Fo = aW/L? gives © =e F, and, d@/dFo = -Bie FY - -Bi6 w From Eqs. (3.73), (3.75) and (3.77), the one term approximate solution for the average temperature is 8 = AB, and dOldFo = -16 @ where Mis given in Table 3.5. The table below compares these two results for 8 = 0.1. Bi (AB/EFo), p_ (68/EFO), tum 0.02 0.002 0.001989 Ol 0.01 0.009678 10 o1 0.07401 10.0 10 0.2402 = 2 0.2467 Comments: 1. Clearly the suggestion has no merit. The LTC model is seen to be a good approximation only for Bi < 0.1, as stated in the text 2, The fallacy in the suggestion is that the claim of a nearly uniform temperature at long times is not valid-uniform with respect to what? For the claim to be valid the ratio (T-T,)/(T,~T,) must be nearly unity. Study Fig. 3.22 to see that this, ratio tends to infinity as Bi —> =. 3. Of course, the Limit Bi —> co corresponds to a prescribed surface temperature.Exercise 3-64 Given: An 8 cm x 8 cm square cross-section bar of AISI 316 stainless steel, initially at 620°C, ‘cooled by air at 30°C. Required: ‘Time for center temperature to drop to 130°C if the heat transfer coefficient is 400 Wim? K. Assumptions: 1. A Jong bar so that the conduction is two dimensional. Evaluate properties at an average center temperature of (1/2)(620+130) = 375°C ~ 650 K. From ‘Table A.1, p= 8238 kg/m’, k = 19 Wm? K, c= 557 J/kg K. The lumped thermal capacity Biot number is Bi =hL/c = (400)(0.02)/(19) = 0.42 > 0.1; (L= WA) ‘The lumped thermal capacity model is not valid. Thus we use the product solution, item 1 of Table 3.6, withL =4 em 0=PED.POD Sac = Chay Tor, TeTe To-Te | | To-Te ( ber 2Laub Zao + Cosi =O412; Bi = hL/k = (400)(0.04)(19) = 0.84 Using Figure C.la: Fo= 1.5 = ot/L? a=k/pe = 19/(8238)(557) = 4.1410 m/s t= FoL?/o= (1.5)(0.04)°/(4,14x10~) = 579 s (9.7 min) Comments: 1. ‘The properties should be evaluated at a temperature somewhat lower than 375°C to account for the temperature profile across the bar. If lower bound of (620+30)/72 = 325°C is used the resulting change in t proves to be negligible. 2. Check the value for Fo using COND2, or the long-time approximate solution.Exercise 3-65 Given: AS x 10 x 20 em clay brick heated in an oven. Required: Time for center to reach 1500°C if h, = 100 Wim? K. Assumptions: 1. Radiation heating can be ignored. Table 3.6, item 6 applies with L; = 0.025 m, 12 = 0.05 m,Ls = rica = PODPGYDPG,t) TexTe ] [Te-Te 1500-1700, S00 TINO ~ 0.1667 = 018.28 [i ToT. ‘500-1700 “01667 = Gu16ea8is hela _ (100\(0.025) _ 47 k 17 : Similarly Foz = 1.40x10~ t; Fos = 0.350x10~ t, Bi = 2.94; Big = 5.88 ‘To use COND2, iteration is required: the procedure should be organized by constructing a table. ‘Since conduction across the $ cm width dominates, a first guess for time t is obtained by choos- ing values of Fo, that give values of 0.1 somewhat higher than the desired value for ®,.. The time tis calculated from Fo. and is then used to calculate the corresponding values of Fo2, and Fos. 1, = 0.025 m 1, =0.05m 1,=01m 818.2823 Fo 8 bs | Fo ba Fos 8s 140 0.299 2500} 035 0.736 | 0.0875 0985 | 0.217 150 0.271 2679 | 0375 0.711 | 0.0937 0.980 | 0.189 1.60 0.246 2857 | 040 0.687 | 0.100 0.975 | 0.165 1.70 0.224 3036 | 0.425 0.664 | 0.1063 0970} 0.144 Interpolating 0, = 0.1667 when t = 2840 s (47 min). Comments: 1. As will be understood after studying Chapter 4, the heat transfer coefficient b. ‘will not be constant over such a wide temperature range, nor will it be the same on all sides of the brick. In addition, radiation heat transfer will play a role. Hence a value of h, = 100 W/m? K should be viewed as a nominal value, and the above result should be regarded as a rough engineering estimate.Exercise 3-66 Giver: A 2 cm thick slab of resin filled cloth undergoing transient heating and cooling. Required: Duration of @) heating stage, (i) cooling stage. @ Heating stage. Use COND2 with Bi = 50010 approximate a negligible contact resistance at the press surface, T? _ pot? _ (1700)(1200)0.017 __ a ok 06 i 3408 85-100 20-100 0.1875 0.1875 at n = 0, COND? gives Fo = 0.779; t= t,Fo = (340)(0.779) = 265 s Gi) Cooling stage. We cannot use COND? for this stage since the “initial” temperature is not uniform as required for the analysis of Section 3.4.3. However, let us check the Biot number for possible lumped thermal capacity model solution. b(V/A) _ bb _ (70.00) Bi k k 0.6 =0.0117 Since the Biot number is small the temperature profile will be approximately uniform during the ‘cooling stage. For an appropriate initial temperature we find the average temperature at the end of the heating stage. For Bi = 500, Fo = 0.779, COND2 gives ® = 0.88, Hence 6 = 1-0.88 = 0.12. TY 1 To-T. — 20-100 =0.12; solving, T 4°c ‘Using Eq. (1.38), te = po(V/A)/h. = (1700)(1200)(0.01/7) = 2910 s Tt 40-20 ToT. 904-20 =0.284 =e; t= 1.258 t= tt, = (2910)(1.258) = 3660 s (61 min) Comments: 1. An exact solution for the second stage requires using the temperature profile at the end of the first stage as the initial condition for determining the constants A, in Eq. (3.68). Since the work involved is considerable, a numerical solution procedure ‘would be more appropriate (see Section 3.5),Exercise 3-67 Given: An 8 cm diameter, 12 cm high can of beer at 300 K is placed in a refrigerator at 277 K. Required: An estimate of the beer temperature after 6 hours if h, = 5 W/m? K. ‘Assumptions: 1. The beer properties can be taken to be those of water at an estimated average temperature of 290 K. Free convection inside the can of beer will tend to mix the beer; however as a worst case we will assume no mixing and check the Biot number for a lumped thermal capacity analysis. From Table A.8: k = 0.591 Wim K, p = 999 kg/m3, c= 4186 J/kg K. Using Eq, (1.40) Bi=b,(WA)/K v ()(0.087/4)(0.12) +. =0.0150 A 2n(0.087/4)+n(0.08)(0.12) 1127. Bi = (5(0.0150/0.591 ‘This value is a litle higher than the accepted maximum of 0.1 for the lumped thermal capacity ‘model to apply; but the convection mixing will increase the effective k value and reduce the Biot ‘umber, $0 the model will be adequate for our purpose. Using Eq. (1.38), t, = po V/A)ih = (999}(4186)(0.0150Y5 = 1.255x108 s t= tite = (6)(B600V/1.255x104 = 1.721 17 = 0.179 TT, 1277. "3 0179 = To 277} solving T= 281.1 K. Comments: 1. This is a lower bound for the given heat transfer coeflicient. An upper bound could be obtained by assuming no mixing and using COND2, and Eq. (3.822)Exercise 3-68 Given: A 6.cm sphere buried at a depth of 30 cm in a large box of soil. Required: Is it practical to use the steady state solution, Table 3.2, item 7 to determine kyo? ‘The major question conceming the practicality of the proposed method of determining soil con- ductivity is the time required to approach a steady state. For this purpose we need to estimate a characteristic time for this conduction process. Inrespective of the exact details of the geometry the Fourier number defined in Section 3.4.1 gives the appropriate time scaling, t.,.2 Fox bike g ‘Take the box to be a cube with 2 m sides, soL ~ 1m, Estimating a= 04x10 m?/s for the soil from Table A.3, 1)? /0.4x10 -5x10° s (694 hours ~ 30 days) fe Clearly, the heat flux will approach its steady value 100 slowly for the method to be practical Comment ‘A transient experiment in winich the sphere is heated rapidly to 30°C and main- (oy tained constant by a thermostat may be more suitable, Carslaw and Jaeger (see the Chapter 3 Bibliography) give the following expression for the temperature at radius r for conduction from a sphere of radius R into infinite surrounds, peel Sey SP a T.-To + ony Ifa temperature sensor is placed at, say, rR = 6 cm, then Eq, (1) can be used to solve for the unknown value of a. @Exercise 3-69 Given: A 4kg beef roast, initially at 5°C placed in a 150°C oven. Required: (j) Time for center to reach 70°C, and (ii) relative importance of heat required for vaporization. Assumptions. 1. The roast is spherical. 2, The heat transfer coefficient is constant at 12 W/m? K. 3, Properties as for water. @ Evaluate properties at 310 K. From Table A.8, k = 0.628 Wim K, p = 993 kg/m3,, 174 Jikg K. Hence a= k/pc = 0.152x10% m?*/s, First we calculate the time constant for the process, tg =R?/0= (0.10)°/0.152x10 = 65,800 s (18.3h) ‘Next we calculate the lumped capacity model Biot number Bi =h(R/3)/k = (12)(0.103/0.628 = 0.64 ‘Since Bi >> 0.1, the lumped capacity method cannot be used. The desired value of @ is 70-150, 5150 7052 Solution using COND2. ‘The required inputs are: Geometry = 3 (sphere) Bi=191 ‘Output option = 2 (0 vs. n plot) Fo=must guess and iterate nrange =0, 1 {A few iterations will give Fo = 0.243 for @, = 0.552. The time for the center to reach 70°C is 1=1.Fo = (65,800)(0.243) = 16000 (4.44h) Gi) For water hg = 2.26x10° J/kg (Table A.12a). Hence the enthalpy of vaporization is Quay = Whig = (0.4)(2.2610%) = 904 KT Ifthe complete roast were heated to 150°C, the change in sensible enthalpy would be Q= cW(150-To) = (4174)(4)(150-5) = 2421 kT From COND2 with Fo = 0.243 we obtain & = 0.637, thus the actual sensible enthalpy change is, Quenitie = ©Q = (0.637)(2421) = 1542 KI Comments: 1. The enthalpy of vaporization is a large percentage (59%) of the sensible enthalpy gain. Thus our estimation of cooking time is too short, 2. Clearly one cannot ignore evaporative heat loss in this type of problem. 3. Notaall the liquid evaporates from the roast: some drains as juice into the pan,Exercise 3-70 Given: A telescope mirror blank in the form of a disk placed in an oven for stress relieving, Required: Time required for minimum temperature to reach 400°C. Assumptions: 1. The beat transfer coefficient is constant over the disk surface. A product solution must be used. Table 3.6, Item 4 applies O= PRC. fe cjiter oF 0.05 = 0.18.2 Fora slab of half thickness L = 0.025 m, ct _ (0.51x10)t i, = Bek 020.025) _ 9, cm 3 B= = 0275 LP © 16x10t; Bi; t 1.09 For an infinite cylinder of radius R = 0.125 m, R 264x101; Bip = 12) i Z = 1376 ot _ 051x101 _ R @.1257 Fo, = ‘To.use COND?, iteration is required: the procedure is organized by constructing a table. Since conduction across the disk dominates, a first guess for time should be obtained by choosing a value of t which gives 0,1 Somewhat greater than the desired value of 8. = 0.05. Timet | Slab,L=0.025m | Cylinder, R=0.125 m s Fo; 8a Fo: G2 | Oe102 Lisx10* | 9.384 0.0983 | 0375 0596 | 0.0586 Lis 9.628 0.0925] 0385 0585 | 0.0541 1.20 9.792 0.0887 | 0392 0577 | 0.0512 123 10.00 0.0842 | 0401 0567 | 0.0477 Interpolating, t = 1.21x10* s (202 min) Comments: 1. ‘The assumption of a constant value of h, is not to0 good: the above result should be regarded as an engineering estimate, rather than a precise result.Exercise 3-71 Given: A stainless steel billet at 20°C placed in an oil bath at 300°C. Required: Center temperature after 500 s by three methods: (i) lumped thermal capacity model, Gi) an infinite cylinder model, and (ti) a finite length cylinder model. Assumptions: 1. The heat transfer coefficient is a constant with time, and over the billet surface. Evaluate k and c for 18-8 (AISI 302) stainless steel at a guessed average temperature of 400 K: from Tables A.1b and A.1c,k = 17 W/m K, c= 512 J/kg K. From Table A.la, p = 8055 kg/m? ‘Then ot = k/pe = 4.12x10% m/s, @ —_Lamped thermal capacity model: Eq. (1.38) applies A= 2¢m/4)(0.15)? + 2(0.15)(0.3) = 0.1767 m2; V = (x/4)(0.15)°(0.3) = 0.00530 m? pV _ (512)(8055)(0.00530) _ A 400(0.1767) an THe ek, F300 sonst. vs eee; DO + solving T= 244°C Tot. ° * 20-300 nee Gi) Infinite cylinder model. Use COND2 Fo = Ot = G12ITYSO) _ 9.565; i= BR - 4O0O075) _ 1 76 R? 0.075)" k an COND? gives 0, = 30. = 0.552; solving, T. = 145°C. 20-300 (ii) Finite cylinder model. Use COND? and the product solution item 4 of Table 3.6. Infinite cylinder; Fo, and Bi, and Q.1 as above: @.1 = 0.552. BIRIOVEO _ 90916; Big = AOMOID - 3.53; 9.=0.987 .15Y ay Slab: Fo: = , = (T-300)(20-300) = (0.987)(0.552); solving T, = 147°C is much greater than 0.1, the error in the lumped model result is to be expected. 2. The infinite and finite cylinder models give almost the same result for this problem.Exercise 3-72 Given: Plywood, 1 cm thick, sandwiched between plates at 105°C. Required: Time for center to reach 95°C from 25°C. Assumptions: 1. Use properties of white pine perpendicular to grain. For white pine perpendicular to the grain, Table A.3 gives k= 0.10 W/m K, a= 0.071x10% m?/s at 300 K. ‘Since data at other temperatures are not given, these values must suffice. LE + t =1?/er=(0.005)"/(0.071x10) = 352 s TT. _ 95-105 TT, ~ 25-105 0.125 ‘COND? can be used for this problem. Since the plate temperatures are specified we use a large Biot number to simulate the situation. Using Bi = 250, COND2 gives Fo = 0.948 for 6(0) = 0.125. t= Fo = (352)(0.948) = 334 s (5.6 min) C4 ‘Comments: 1. Use COND2 to check how large a Biot number is required for the result to be independent of Bi.Exercise 3-73 Given: A 5x10x20 cm clay brick heated in an oven. Required: ‘Time for the center to reach 1500°C if h, = 200 W/m? K_ ‘Assumptions: 1. Radiation heating can be ignored. Item 6 of Table 3.6 applies, with L; Ortoce =POUP(YP(Z.1) | To-Te | | To-Te ‘Ly sob 2 sab .025 m, Ly = 0.05 m,L5 = T. ToT. 1500-1700 © 500-1700 = 0,167 = 8.18:283 Foy = St = 35H «5 60.104; Biy = PEL = COMOMD . 04 (0.025)? k 17 i Similarly Fop = 1.40x10~t; Fos = 0.350x10~t Bi = 5.88; Bis = 11.76 oo To use COND2, iteration is required: the procedure should be organized by constructing a table. Since conduction across the Ly width dominates, a first guess for time t should be obtained by choosing values of Fo, which give values of @.1 somewhat higher than the desired value for 0... ‘The time t is calculated from Fo, and is then used to calculate the corresponding values of For and Fo}, 1, = 0.025 m 1,=005m | 13=0.10m Fo, 8a ts | Foo Oa Fos Bs | 010.203 1.0 0.295 1786 | 025 0.788 | 0.0625 0.995 | 0.231 1.1 0257 1964 | 0.275 0.754 | 0.0688 0.992 | 0.192 12 0223 2143 | 03 0.722 | 0075 0.988 | 0.159 13 0.194 2321 | 0.325 0.691 | 0.0813 0.984 | 0.132 Interpolating, 8, = 0.1667 when t= 2100 s (35 min). Comments: 1. Chapter 4 will show that h, will not be constant over such a wide temperature range, nor will it be the same on all sides of the brick, Also, radiation ‘heat transfer will play a role. Hence a value of h, = 200 W/m? K should be viewed as a nominal value, and the above result regarded as an engineering estimate.Exercise 3-74 Comments: An egg initially at 5°C immersed in boiling water. Required: Center temperature after @) 4 minutes, and (ji) 7 minutes. Assumptions: 1. The egg can be modeled as a4 cm diameter sphere of water. 2. ‘The outside heat transfer coefficient is constant at 1200 W/m? K. @ Guess an average temperature of 300 K: Table A.8 gives k= 0.611 Wim K, o& (= v/Pr)= 0.147x10~ m?Js for water. COND2 will be used. 4, =R%/ar= (0.02)"(0.147x10) = 2721s Fo =the = (4)(60/(2721) = 0.0882 Bi =h,Ryk = (1200)(0.02/0.611 = 39.3 Tr To-Te COND? gives 0 = 801 (T-100) = 0.801(5~100), hence T = 23.9°C Gi) Fo=(7)(60)(2721) = 0.154 COND? gives @ = 0.465; T= 55.8°C Comments: 1. After studying Chapter 4 the student should tackle the problem of estimating a suitable value for the heat transfer coefficient,Exercise 3-75 Given: A large rectangular safe with 10 om thick asbestos insulation exposed to a fire. Required: Time for temperature of contents to increase from 20°C to 150°C. Assumptions: 1. Interior of safe is isothermal. 2. One-dimensional conduction through insulation, 3. Insulation outer surface temperature is constant. ‘We are told that the combined effect of convection and radiation from the fire is to raise the out- side of the safe to 800°C. Thus we can use COND2, or the temperature response charts, to solve the problem of transient conduction in the insulation by letting Bi —> e, which gives T, = T.. For slab of half-thickness L = 0.1 m, centered on the inside surface of the insulation we calculate o,= Tete _ 150-800 ° ToTe 20-800 ‘and find the corresponding Fourier number for Bi — », Solution using COND2. ‘The required inputs are: Geometry = 1 (slab) Bi=500 ‘Output option = 2 (@ vs. 7 plot) Fo = must guess and iterate ‘range =0,0.1 ‘A few iterations will give Fo = 0.168 for @ = 0.833 at = 0. ‘The time constant for the heating process is te =L?/0.= (0.10)? /0.4x10-) = 25,000 s Hence the time required for the inside of the safe to reach 500°C is t= Fot, = (016825000) = 4200 s (70 min) 1833 ‘Comments: 1. This time is long enough for the fire to be extinguished. 2. Use the temperature response chart to check this result. 3. Since Fo < 0.2,the long time approximation cannot be used.Exercise 3-76 Given: A cylindrical, 2 cm diameter, AISI 302 stainless steel pin, initially at 20°C suddenly exposed to saturated steam at 100°C. Required: Temperature at r= 0, x = 0.04 m after (1) 10 s, and (i) 100s. Assumptions: 1, The heat transfer coefficient for condensing steam is very large, hence Bi — «. ‘Since Bi —> =, the lumped thermal capacity model cannot be used. Item 4 of Table 3.6 applies, @ = P(O.04,1)CO,1) = 0:82 ‘The thermal diffusivity is evaluated at a guessed average temperature of 350 K: using the data in Table A.1, k= 16 Wim K, c= 496 I/kg K, p = 8055 kg/m®. Hence, a. pc = (16)/(8055)(496) = 4.00x10~ m/s. @ t=10s ‘The two Fourier numbers are Lt _ 00X10 )09) _ 9 916 vy 0.05)" 7 cat _ 4.00x10-)(10) Infinite cylinder of 1.0 cm radius: Fo = 2 = = 0.400 elinder of 1.0m radius: Fo, = oon? ‘Then 8; Iq-o and 67 In-p can be obtained using COND2; for Bi = 500 6; =0.741; @ TT. = = 618) = (0. = 0.11 ToT 8182 = (0.741)(0.160) = 0.119 Solving for T, T= 90.5°C. Gi) t= 100s Fo; = (0.016)(10)= 0.16 Fop = (0.400)(10) = 4.00 ‘Again using COND2 0; = 0.278, 0) =0 = 0,0, =0; T= 100°C Comments: 1. Condensation heat transfer coefficients are calculated in Chapter 7.Exercise 3-77 Given: An AISI 302 stainless steel pin, 2 cm square, initially at 20°C suddenly exposed to saturated steam at 100°C Required: Temperature at x= 0.04 m, y= =O after () t= 10s, and Gi) t= 100. Assumptions: 1. The heat transfer coefficient for condensing steam is very large, Bi—> =. Item 6 of Table 3.6 applies 0 =P(0.04,0PO,0P(O,1) = 618285 ‘The thermal diffusivity is evaluated at a guessed average temperature of 350 K. Using Table A.1, k= 16 W/m K, = 496 Ikg K, p = 8055 kg/m?; hence 0 = 400x106 ms. @ t=10s ‘The three Fourier numbers are = St _ 4.00110 9110) _ 916 vv (0.05) Fo = Fos = G.00x10}410) =0.400 oon ‘Using COND? with Bi = 500, 8, (7)=0.08) = 0.741 82(n=0) = Os(n=t TT. 00 To-T. 20-100 Solving for T, T = 86.6°C. Gi) t= 1008 Fo, = (0.016)(10) = 0.16 Fop = Fo3 = (0.410) = 4 Using COND2, : Fo, ra! 010205 = (0.741)(0.476)(0.476) = 0.168 Comments: 1, Condensation heat transfer coefficients are calculated in Chapter 7.Exercise 3-78 Given: Glass cylinder initially at 30°C placed in a fumace at 800°C. Required: A graph of the center temperature as a function of time, Assumptions: 1. "The heat transfer coeficient is constant over the surface of the cylinder. be 25em—4 To=30°C ota 7. = 800°C 20cm 10 Wan? K. 83x10 m?/s (Table A.3) k= 1.38 Wim K (Table A.3) Item 4 of Table 3.6 applies, and the center temperature 6, is given by, TeTe To-Te where 8,1 and Q_2 can be obtained using COND2. The Biot numbers are BL A= (20\0.10Y1.38 % = 0182 for Gq) and gp. Then 0, = 0.1022 and T, = @-(To~Te)+Te. The results are tabulated and graphed below. t | Fo,=ow/L? | Fox=0vR? | O21 2 | 08182 | Te s ec 1x10? | 8.3x10% | 531x103 1 1 1 30.0 Sx10? | 415x107 | 2.67x107 1 1 1 30.0 ixio? | 83x10? | 531x107 | 0995 | 0.997 | 0.992 36.2 2x10 0.166 0.106 0.957 | 0.955 0914 96.2 4x10 0332 0.212 0.835 | 0.781 0.652 | 298 8x1 | 0.664 0.425 0.610 | 0475 | 0.290 | 577 1x10* 083 0.531 0521 | 0369 0.192 | 652 2x108 1.66 1.062 0.235 | 0.103 | 0.0242 | 781 4x108 3.32 2.124 0.05, | 0.006 | 3xi0* | 800 ix10° 83 531 axto* | 4xio® | 1.6x10° | 800 Comments: 1. The property values evaluated at Te. C 300K might not be appropriate. ¢00 for this large temperature range. Bs 400 300 200| 100Exercise 3-79 Given: A large water storage tank of diameter 13.5 m and height 10 m. Required: If heaters are switched off, time for water temperature to drop from 22°C fo 16°C. Assumptions: 1. Ambient conditions are constant and the surrounding air is still. 2. Negligible heat loss through the base of the tank. 3, Natural convection inside the tank gives a nearly uniform water temperature. kn —_4, ‘As the tank walls are cooled, natural convection will be induced and tend to give a more uniform ‘water temperature. In order to obtain a lower mf] — bound on the time we will first assume a uniform water temperature and use the lumped ‘capacity model. From Eq. (1.38) with h, replaced by U, the time constant is t, = pVc/UA. Since the thermal resistance of the insulation is relatively large we can ignore the water side resistance, and approximate estimates for the outside heat transfer coefficients will suffice, Taking h,. = 3 Wim? K and h,. = 5 Wim? K Peete 0076, 1 UK Bach, 005 3S .52+0.125; U = 0.608 Wim? K VIA = RR7H/(ER?+2nRH) = (6.75) 10V/(6.75 +2x10) = 2.52 m From Table A.8 for water at 290 K, k = 0.591 W/m K, p = 999 kg/m’, c = 4186 J/kg K. Hence, 4, = (999)(2.52)(4186/(0.608) = 1.73x107 s (200 days!) 1-1, 16-5 _ g-nzuo’, 7.53x10° s (87 di or 5 3x106 s (87 days) ‘This is a very long time. Before proceeding to refine the answer, the engineer should ascertain ‘whether the customer is interested in a more accurate answer. It could well be that itis sufficient to know that the time is longer than some value, say 1 week or 1 month. If a more precise estimate is required, we should next obtain an upper bound estimate by assuming that there is negligible convection; then the temperature response charts or COND2 can be used. Consider a cylinder of water with 2L = 20 m, R = 6.75 m. ‘The Biot numbers are Biz, = UL/k = (0,608)(100.591 = 10.3, Big = UR/k = (0.608)(6.75/0.591 = 6.94 ‘The customer is probably interested in when the volume averaged temperature has reached 16°C,3-79 (continued) that is, when the fractional energy loss is 22-16 _ 6 ® a= = 0353 17 An iterative procedure is required to implement the product rule for a finite cylinder (item 4 of Table 3.6, and Eq.3.81a)). To use Figs. C2a and C2b we need Bi?Fo = (U/k}"at = U*vkpe. Bi?Fo = (0.608)°t/(0.591}(999)(4186) = 1.50x10“"t and b = @, +, -©, 0, tl] BFo 107 1S 007 022 0.27 2x10’ 3.00.12 038 0.45 14x10? 210.10 0.30 0.37 13x10’ 195 0.09 0.28 0.34 Hence t = 1.3x107 s (150 days) We should expect the true result to be closer to the lower bound of 87 days. Comments: 1. Since these times are so long, the specified constant ambient conditions are unrealistic. 2. At this point the heat transfer engineer should seek more information about the problem and ask why the time estimate is required. It could be that these times are Jong enough for the problem, as posed, to be irrelevant.Exercise 3-80 Given: A 10 cm Inconel-X alloy cube, initially at 810°C, quenched in a coolant at 40°C. Required: (i) Time for center temperature to undergo 95% of its decrease to equilibriiim, (i) Time required for a 95% fractional energy loss. Assumptions: 1. A constant and uniform convective heat transfer coefficient of 800 W/m? K. 2. Inconel properties are taken as constant at the average of the initial and fluid temperatures. (Inconel properties will be evaluated at the average of Ty and T,, T, = (1/2)(810+40) = 425°C = 700 K. From Table 1, p = 8510 kg/m?, k = 18.7 Wim K, ¢ = 535 kg K. To check if the Jumped thermal capacity model is valid we evaluate the Biot number from Eq. (1.40), Bipye = ALCV/AYK = (800)((0.1)°(6)0.1)7V/18.7 = 0.713 > 0.1 Since 0.713 is much larger than 0.1 we need to use the exact solution (or the solution for long times because we are investigating the end of the cooling process). We will use the temperature response charts in Appendix C. We use the product solution given as item 6 of Table 3.6, 0 = Px,)P,NPC1) ‘We require 8, = 0.05. For a cube, the three slab solutions are identical, thus we require P(A) = T,-T/(Ty-T,) = 0.05! = 0.368 Bi = h,L/K = (800)(0.05)/(18.7) = 2.14, From Fig. C.1a, Fo = 0.9. ‘The corresponding time is t= FoL4/a, For the Inconel-X, 0 = k/pe = (18.7)/(8510)535) = 4.11x10* m’/s. Hence, t = 0,9)(0.05)°/4.11x10 = $47 s (9.1 min). (i) From Eq. 6.826), with , = @, = @, ® = &, +0,(1-b,) +,(1-8,) 0.95 = 30, -30} +0; Solving (using a standard calculator routine), , = 0.6316. Then, from Fig. C.2a, BitFo = = 3.50(2.14)? = 0.76 t= FoL ar = (0.76)(0.05)"/4.11x10% = 462 s (7.7 min) Comments: 1. More accurate solutions can be obtained using COND2, or the Jong times one- term solution. However, our assumptions of a constant and uniform heat transfer coefficient and constant Inconel properties imply that our answer will not be too accurate, no matter how precisely we solve the model problem. 2. The time in part (ii) is shorter than in part (i) because of the temperature distribution across the block. If the Bi number were small, the temperature would be almost uniform across the block and the two answers would be the same.Exercise 3-81 Given: A convectively cooled slab, thickness 2L. Required: The interior heat transfer coefficient for conduction out of the slab. Assumptions: 1. Fo > 0.2, $0 long time (single tem) solution is appropriate, By definition h= o ‘We first use the first term of Eq. (3.70) to obtain 3T/Ax, ee i aL e Artsindjcos, [TL L To-Te are _ 2Aasin?&y TL ° yasinycosty 2) ‘Then from Eq. (3.71), re __2sin?hy 4 = =e fe ___ “A TTe = ToT) = 8°" 5 Gish com) =) and from Eq, (3.70) 2sindycoshy Ty =CTo-Tee Fe T.-Te = To-TeJe Ty+sinh cok, tre 2sin?Ay-2A, sind coshy ‘Combining, (T-T)AT Te) = ToT ee TQ +5inkcoshy) 8) ‘Substituting Eqs. (2) and (3) into Eq, (1) gives ‘ik Msin?ay L sin? Ay—ysinthy cos 2nsin?Ay @ and Nog, = qs __ Ma Sin hy Ay sindy cosy For Bi =», Table 3.5 gives 2y = (2.467)! = n/2, and Nuz. = 17/2. = 4.934, Comments: 1. The interior heat transfer coefficient is useful when modeling large systems of compo- nents when the thermal response of each component needs to be modeled as simply as possible. 2. Prepare a graph of Nu versus Bi, First show that Eq. (4) can beExercise 3-82 Given: A convectively cooled cylinder, radius R Required: The interior heat transfer coefficient for conduction out of the cylinder. Assumptions: 1. Fo > 0.2, so long time (single term) solution is appropriate. By definition h Retaining one term in Eq. (3.72) gives hy T4e= ToT [» [¥] (0) Hence, using Table B.3, ia dy oF [pe = —Clo-TedAre™ [i}.a0 ® Also, from Eq. (3.73) TT, = ToT. XI-®) = (To-Te)Ae By From Eq. (1) THM = (ToT A PIA) ‘Combining, Tr, = FTW.) = Ty-T.)Ae Bho) ay Fl ne KR [OD Tt, Brot) hi Paha) @) a= BORD , Zohar ek” Bylo) ‘using Tables 3.4, 3.5 and B.3, 3304) Bi—>, Ay = 2.405; By =2 a + Jou) =0 Nu=)j=5.784 Comments: 1. The interior heat transfer coefficient is useful when modeling large systems of compo- ‘nents when the thermal response of each component needs to be modeled as simply as possible, 2. Prepare a graph of Nu versus Bi, First show that Eq. (3) can beExercise 3-83 Given: A convectively cooled sphere, radius R. Required: The interior heat transfer coefficient for conduction out of the sphere. Assumptions: 1. Fo > 0.2, so ong time (single term) solution is appropriate. « a Re TH, Retaining one term in Eq, (3.72) gives By definition h= ‘i 1 TAT, = (To-TeA eH Oy AR) te ‘Thus ho af Flan = MoT )Ae Sloe -ajfe Mcoshy—sindy = (To-Te)Are aR @ ‘From Eq. (3.73) TT, = (To-TL-) = (ToTe)Are YB, @ From Eq. (1) 7 TAT. = ToT) Combining, (sinky-2ycosy) RQ4By-sinhy) WOR) _ 2Giny~Ayooshy) ® k A4By—siny, From Tables 3.4 and 3.5, Nu= forBi>—. 24 =m B= Comments: 1. The interior heat transfer coefficient is useful when modeling large systems of compo- ‘nents when the thermal response of each component needs to be modeled as simply as 2. Prepare a graph of Nu versus Bi. First show that Eq. (4) can beExercise 3-84 Given: A regenerative heat exchanger matrix consisting of parallel plastic plates 6 mm (= 21.) thick. Required: Overall heat transfer coefficient as a function of plate spacing. Assumptions: 1. Fo > 0.2 over most of cycle so that the interior heat transfer coefficient is con- stant, ‘The overall heat transfer coefficient is obtained from ceed UA BARA ‘The interior heat transfer coefficient is obtained from Exercise 3-81 as Ay agsin? Ay ML Sanaa sinhycost 2 where k, = 0.15 Wim K, L = 0.003 m, and 2, Bi) is given in Table 3.5. The convective heat transfer coefficient is he = 4k;/b = (40.1 = 0.4/ and the Biot number is | O) Bi = heL/k, = (0.4/0)(0.003)(0.15) = 0.008/b Using Eqs. (1) through (4) the following results are obtained. ’ | Bi 7 UY] GyGh.) m wnt K wit K | Wim? K ‘G00r | & 400 1293 | 977 BA 0.002 | 4 200 1332 | 800 15 0.004 | 2 100 1380 | 580 0.72 0.006 | 1.33 66.7 140.5 45.2 OAT 0.008 | 1 50 142.4 37.0 0.35 0.01 08 40 143.6 31.3 0.28 Comments: 1. Notice that h; is relatively insensitive to the Biot number. 2. U increases as the plate spacing decreases; a small plate spacing is desirable for this type of matrix.Exercise 3-85 Given: Temperature time response of the center of a grapefruit immersed in water at °C. Required: Thermal diffusivity. Assumptions: 1. A large heat transfer coefficient giving Bi «. 2. ‘The thermal diffusivity is constant. ‘The simplest procedure to determine ot from the given data is to process each data point, using the temperature response chart Figure Clo, or COND2, to find the Fourier number corresponding to the dimensionless center temperature @, = (T.—T.)/(To-Te). Then since Fo = at/R?, = FoR? _ (0.034)Fo _ 1.93x10*Fo ~ t — 60t [min] {min} t | tm | @ | Fo @ min | °C mpsxi0 10 | 196 | 098 | 0045 | 87 15 | 180 | 090 | 0066 | 85 20 | 15.8 | 0.79 | 0.086 a3 25 | 130} 065 |o10!] 85 30 | 108 | 054 | 0.131 84 35 40 84 | 042 | 0.158 87 72 | 036 | 0.174 84 45 | 58} 029/0196| 84 ‘The value of ocobtained at t= 10 min should be discarded since 4T-/at is small for t™ 15 ‘Thus we conclude that &.=8.4x10"* m?/s. Comments: 1. At 10°C the thermal diffusivity of water is 13.7x10-* m?/s, Thus o: for the grape- fruit is 8.4/13.7 = 61% of the value for water at 10°C.Exercise 3-86 Given: Slab of thickness L with face at x= 0 insulated. Required: (j) Temperature response to constant heat flux q, imposed on the face at x=; and (ii) temperature response if a flux —q, is subsequently imposed at t= t) Assumptions: 1. Constant thermal diffusivity. 2. One-dimensional conduction. @ Let0=T-Tp Y tan B02 ® * and = @(x,0)=0 0 t % ‘Since the third condition is nonhomogeneous, we choose 06%. = WOK.t) + 60%) + 2) Wg ey =-40)— woere Shea Sos and (sO) = 03) — 010) with homogeneous boundary conditions y,(0,t)=0; ye(L,t)=0 ‘The general solution for (x,t) can be obtained by the method of separation of variables as WO%st) = Ao+ E Age cosBax: Ba =nA/L, = ‘Substituting Eqs. (2) and (3) into Eq, (1) and rearranging gives a £8. © ob, aconstant, with 9,(0)=05 $(L)=a4Mk dx’ dt ‘The general solutions of Eqs. (5) are o=F bx? +cx+d, @=abt+e Using the boundary conditions gives c= 0, 20) = (ogy fk t+ 908) = (Qu/2KL x74 Because the solution for (x,t) depends on a(t) and (x), we may arbitrarily choose d and @ @) @ ©3-86 (continued) ah. Bort TT i qk 6? Gi) By the principle of superposition for a linear differential equation, and using part @i), we have T-Tp = 0 = O(%,1)-O(x,t-+11) Sh [ata ote) 2 5 am k 6? ci we = L ve Berea Cie El cog BEE ae eh Pe J= L and is the temperature response for t> t; when q, = 0 for t> ti. ‘Comments: 1. The use of the principle of superposition in this manner allows one to obtain the temperature response for an arbitrary surface heat flux variation, q(t). ‘See, for example, G.E. Myers, Analytical Methods in Heat Conduction, pp. 153-163, Geniumn Publishing Corporation, Schenectady, NY, 1987.Exercise 3-87 Given: Slab of thickness L, face at x = 0 insulated, energy E deposited on face at x=, at time t= 0. Required: (i) Temperature response T(x,t), and (ii) time required for backface to attain half its maximum temperature rise. Assumptions: 1. One-dimensional conduction. ) Let 0(x,t) be the temperature response for a unit uniform heat fux obtained by setting q, = 1 in the result of Exercise 3-86, part (i) L | 3a? | ot 2 ane G ete - SES i If abeat fux of q, is imposed from time t= 0 tot = 0 + At, then following part (i) of Exercise 3- 86, the resulting temperature response is TE,D-To = 95 {0(%,t}-O,t-AD} a, ae Oasis) for qdt=E E _¢ 2061) =e OED for at +0 Gi) From part @) EL| a , 20 3 The lat Ee cos 2 at t L ‘The maximum temperature rise at x = 0 is as t ~~, since the model does not allow for heat loss from the front face, ‘Trnax—To = Eoy/KL; hence ty, is obtained from ‘Taking the first term on the right hand side gives 1 oo at LF ™, solving, ty ~ 1.39 L?/ee. Comments: 1. The practical use of this result is demonstrated in Exercise 3-88.Exercise 3-88 Given: ‘Temperabure-time response of back face of a laminate heated by a laser pulse. Required: Estimate of thermal diffavisity. Assumptions: 1. Back face well insulated. ‘From Example 3-87 the time for the back-face temperature rise to equal half its maximum rise is y= 1397 Po From the emf-time trace the maximum emf increase is =90 mV. The time for the. 10 90/2.= 45 mV is approximately (28~18)1.0x10 = 1.0x10 s. For a 1 mm thick laminate, (1,39)(0.001)?/(1.0x107) = 170 Solving, o = 14x10 ms. to increase Comments: 1. Although simple, this method is of limited accuracy. The above result is probably ‘g00d to within 5%; using the original emf-time trace, the error could probably be re- duced to 2-3%. 2, The literature contains data for cz obtained using this method given to three decimal place accuracy. Such results should be viewed with skepticism. 3. However, for engineering purposes a few percent error is quite acceptable. There will usually be other larger sources of error in the complete design calculation.Exercise 3-89 Given: Square plate with edge temperature prescribed (Example 3.12). Required: Computer program to calculate temperature distribution using a 5x5 mesh. Assumptions: 1. Temperatures are constant across the thickness of the plate to give a two- dimensional problem, : so 0 0 0 o ‘The Figure shows the mesh, and the prescribed tempera- 5 tures along the edges. The total number of nodes is 5x5. 4 100 L ° ‘The temperatures at interior nodes are eee i 3100} ° Ton = 4g Tmart*Tartn Tne t+Te tn) 4 2 100 v7 ° ‘The exact solution is given by Eq. (3.30) as : fF FEEL gin B® sinh BE acd et pgsinh BED a a THY)=Ts Results using Gauss-Seidel iteration with e= 1x10“, and a flow chart of the program, are shown. below. Node | Temperature, °C | Exact Value | Percent Error Tt 42.857 43.203 0.801 Tt 52.678 54.054 2.546 Ty 18.750 18.203 3.005 Ts, 25.000 25.000 0 Ts 7.143 6.797 5.090 Ts 9.821 9.542 2.924Exercise 3-90 Given: Square plate with edge temperatures prescribed (Example 3.12). Required: (i) A computer program to calculate temperature distribution using a NxN mesh Gi) The effect of Non accuracy. Assumptions: 1. Steady, two-dimensional conduction. ‘The Figure shows the mesh NxN. The prescribed tem- peratures along the boundaries are Ty,.1 = 50, Tyyyv = 50, ‘Tyn = 100; n# LN, others Ty,q = 0. The temperatures at interior nodes are, Try = CYT *Taei 2 Tmo-1*Te-t)- Results using ‘Gauss-Seidel iteration with € = 1x10“ are shown for nodes T; through Ts (see Example 3.12 for locations). Also shown is a flow chart. 12 5x5 9x9 17x17 Node | Calculated | Percent | Calculated | Percent | Calculated | Percent | Exact Values | Error | Values | Emor | Values | Enor | Values Ty 42857 | 0801 | 43108 | 0227 | 43.176 | 0.0625 | 43203 Tt 92.678 | 2546 | 53.607 | 0827 | 53.930 | 0.229 | 54.054 Ts 18.750 | 3.005 | 18382 | 0983 | 18249 | 0253 | 18.203 Ts 25.000 | 0 25.000 | 0 24.997 | 0.012 | 25.000 Ts 7143 | 5.090 6.894 | 1.427 6821 | 0353 | 6.797 Ts 9.821 | 2.924 9.627 | 0.891 9562 | 0210 | 9542 Comments: 1. We expect the accuracy toe oForer as? Cora) as the mesh size is halved, the percent ertoris reduced by a factor of 4,Exercise 3-91 Given: A thin rectangular plate 0< x | ssco1 | 58292 } 58.205 T, | 45.628 | 45310 | 45224 T;, | 36672 | 36348 36.262 T, | 73494 | 78.455 78.445 Tio | 60.822 | 60.622 60.565 T | 47595 | 47309 | 47228 T2 | 38303 | 37985 37.897 Comments: 1. Check that the accuracy is of order Ax?--as the mesh size is halved, the percent error is reduced by a factor of 4. 2. Asan additional exercise, calculate the heat flow across the edge at 100°C using Eq. (3.88), and compare it to the heat flow into the fuid calculated using Newton's law of cooling. Investigate the effect of mesh size on the accuracy of this energy balance.Exercise 3-93 Given: A 2 mm thick straight rectangular stainless steel fin, 2 em long. Base temperature 100°C, T, = 20°C, h, = 300 Wim? K. Required: (i) Finite difference formulation for one-dimensional steady conduction, i) Gauss-Seidel iteration with Ax -0,5cm to obtain temperatures for five iterations. Assumptions: 1. One-dimensional conduction. 2. Negligible tip heat loss. ‘An energy balance on an interior finite control volume requires that Qeony = BART, -T.) xem, x wane Substituting in Eq. (1) and dividing by kA/Ax, h Ax? qT) Tas Ty Let B? = hO/KA,; then, solving for Ty Tyga *T yey *B7AK-T, T, 2+B7Ax? i m=2,3,4. ) For node 5, a negligible heat loss implies dT/dx = 0 and symmetry of the temperature profile about x = L. Hence we can simply put T,,,; = Tj. = T, in Eq. (2) to obtain 27, +B°Ax7T, Reet @) 2+B7Ax? B? = hO/KA, = (300)(2)/(15)(0.002) = 2.0x10* BPAx? = (2.0x104)(0.005)" = 0.5 ‘The nodal equations are3-93 (continued) T, = 100°C 1 {100 +T, +(0.5)(20)} = 0.4(110+T,) 25 T, = 0.4(1,+T, +10) Ty = O.4(T;+T, +10) T, = 04@T, +10) As an initial guess choose T; = 70°C, T; = 50°C, Ty = 40°C, T, = 35°C. Then using Gauss-Seidel iteration, Ty = 0.4(110450) = 64°C T, = 04664440410) = 45.6°C T, = 0.4(45.6435+10) = 36.24°C T, = 0.4(2x36.24+10) = 32.99°C T, = 0.4(110445.6) = 62.24°C Ty = 0.4(62.24436.24410) = 43.39°C T, = 0.4(43.39432,99410) = 34.55°C T, = 0.4(2x34.55+10) = 31.64°C T, = 0.4(110443.39) = 61.36°C 0.4(61.36+34.55+10) 0.4(42.36+31.64+10) T, = 0.4(2x33.60410) = 30.88°C T, = 0.4(110442.36) = 60.94°C 0.4(41.82+30.88+10) = 33. 0.4(2x33.08+10) = 30.46°C T, = OA(L10+41.82) = 60.73°C = 152°C 0.4(41.52+30.46+10) = 32.79°C Ts = 0.4(2x32.79+10) = 30.23°C Comments: 1. Using Eq, (2.36) the tip temperature is T = 204(100-20)/cosh2.828 = 29.43°C. 2. A few more iterations are required to get an appropriately converged numerical solution.Exercise 3-94 Given: A 2 mm thick, straight rectangular stainless steel fin, 2.cm long, Base temperature 100°C, T, = 20°C, h, = 300 Wim? K. Required: Computer program to solve for temperature distribution using Gauss-Seidel iteration; effect of node spacing on solution accuracy. Assumptions: 1. One-dimensional, steady conduction. 2. Negligible tip heat loss. ‘An energy balance on an interior finite control volume ton requires that 7 : Qh = Qoony +O O sytem ee ; a ae —_t—__L_ Qy = 7 > 5 Qoomy = B,OAXCT,,-T,) 7 Substituting in Eq. (1) and dividing by kA/Ax, 2 (Ty-T.) “Tyas Ty Let B= h,O/KA,: then, solving for T,, gives Tyrer Tine *B°AX7P, T= ets m=2,3,..Mel @ 2+BAx? For node M, a negligible heat loss implies dT/dx = 0 and symmetry of the temperature profile. about x L. Hence, we can simply put Ty. = Tp.1 = Tyg: in Eq. (2) to obtain Ty *B°AR7T, «Fe BFax? 8) B? = h,O/KA, = (300)(2)/(15)(0.002) = 2.010 m? Also, T, = 100°C.3-94 (continued) Its also of interest to calculate the heat loss, l Q = fa,r-7)eax ; = 2h(T,-T,AW2) + (Ty-T JAK +... + Ty y TAX + (Ty T(AX/2)1 From the analytical solution FINI gives Ty = 29.4°C, Q = 337.4 W/m. The table and graph show the effect of node spacing on heat flow and the temperature distribution. Also given are the number of iterations for a convergence criterion of 0.001 K. Ax,m 0.005 0.002 0.001 0.005 0.0002 Q,Wim 357.2 3404 3379 337.3 337.2 Iterations 247. 407.1404 6993 0 0.005, 0.01 0.015 0.02 x(m) Comments: 1. The above solution was obtained using MATLAB. 2. A node spacing of 0.002 m gives less than 1% error in Q.Exercise 3.95 Given: A 2 mm thick, 4 cm long composite material fin, base temperature 400 K, losing heat by radiation to surroundings at 360 K. Required: (j) Finite difference formulation of the steady, one-dimensional heat conduction problem, (Gi) Four iterations of a solution with node spacing of 1 em. Assumptions: 1. Steady one-dimensional conduction. 2, Radiation can be calculated using a radiation heat transfer coefficient for ‘exchange between a small gray object in large, nearly black surroundings. 3. Negligible heat loss from the fin tip. An energy balance on the finite control volume requires w Qh = Qa Qe MO ~. jy =k am Tt oe cea Qua * SFE [TAT VIP (Ty, “Ty OAK Substituting in Bq. (1) and dividing by kAJAx, (Fag-Ty) Tener Tn OL(T AT V2 Ax? ator, = 81a T IF KA, Let 7 = 40e/kA,, and rearrange: 2 Tat Tm PAX TT TPT, 2+PAx [7,472 T, >» m=2,3,4 ® Notice that T,, appears on the right side of this equation as well: it will be treated as a "source" term and evaluated at the previous iteration during Gauss-Seidel iteration. For node 5, the assumption of negligible heat loss implies dT/dx = 0, with syrametry of the temperature profile about x= L. Hence we can simply put Tp. = Tp, = Tz in Eg, (2) to obtain3-95 (continued) 27+ Ax UT +T,2PT,, HP ORT TP Ts ° f = doe OIA, = (4\5.67x10 *)(0.85)(2)/(4)(0.002) = 4.820x10$ PAX7[T, AT, V2F = (4.820x10 )(0.01)(T,/2+180)> = 4.820x10 (T,,/2+180)° T, = 400 K, and the nodal equations are 400+T,+4.820x10 “°(T,/2+180)*(360) 2+4,820x10 *(T/2+180)* +T.,+4.820x10 (7/2 +180)°(360) 2+4.820x10 °(T,/2+180)° Ty¥T5+4.820x10 %T/2+180)°360) 4 _ 2 y+4.820x10 2(T.y2+180)°360) 2+4.820x10 *(T,/2+180)° : 2+4.820x10 °(T/2+180)° ‘As an initial guess choose T, = 388 K, T; = 380 K, T, = 374 K, T; = 370 K. Then 400+380+4.820x10 °(388/2+180)7(360) 2+4.820x10 °(388/2+180)* = 386.64 K = 386.64+374+4.820x10 °(380/2+180)°360) _ 37g 11 K 2+4.820x10 °(380/2+180)° 378.11+370+4,820x10 °(374/2+180)°(360) 2+4,820x10 °(374/2+180)° = 372.56 K (2)(372.56)=4.820x10 “°(370/2+180)*(360) 2+4.820x10 °(374/2+180) = 371.24 Comments: 1. A hand calculation for this problem is time consuming: use of a spreadsheet or writing a computer program is indicated (see Exercise 3-96)Exercise 3-96 Given: A 2.mm thick, 4 cm long composite material fin, base temperature 400 K, losing heat by radiation to surroundings at 360 K. Required: Computer program to solve for temperature distribution using a finite difference ‘method. Effect of mesh size on accuracy. Assumptions: 1. Steady one-dimen: 2. Negligible heat Joss from the fin tip. 3. Radiation calculated for a small gray object in large, nearly black al conduction, surroundings. foe An energy balance on an interior finite control volume ieee fi requires sot: Wy = Qaat Qe o ; TaTat OW ee ee ee 2 7 i Oe = Qua = CCPAKT STA) = oe PAX(T2+T2)(T,,+T,.XT aT a) Substituting in Eq. (1) and dividing by KAJAx oe Ax (T2+T (7, +T,) Ta*Taa = TA (TT) Tet Fin Let 7° = oeO/KA, and rearrange, Tyne Tet PACT ATT AT IT, a, = Tat Tan PANT TNT TT eo MeL : 2+PAx(T2+T AT, *T,) Notice that T,, appears on the right side of this equation as well: it will be treated as a "source" term and evaluated at the previous iteration during Gauss-Seidel iteration. For node M, the assumption of negligible heat loss implies €T/dx = 0, with symmetry of the temperature profile about x=. Hence we can simply put Ty, = Trae = Tye in Eq. (2) to obtain3-96 (continued) My PACT AT IT AT Ty QP ATT Tag T ye) T °@ Also, T, = 100°C. P= GeOMA, = (5.67x10*(0.85)(2)/(4)(0.002) = 1.205x10 ‘The heat flow is also of interest, and is given by L Q = fh,cr1,)9¢x 3 = eof(T}'-Ts)(An/2) + (Tf Tax +... + Tey -TAAx + TTA (Ax/2)] ‘The table and graph show the effect of node spacing on the heat flow and temperature distribution Also given are the number of iterations for a convergence criterion of 0.001 K. dx,m 0.010.005 0.002 0.001 Q,Wim 16.00 15.60 1549 15.48 Werations = 28894271351 Temp (K) ° om oman: Comments: 1. A node spacing of 0.005 m gives a less than 1% error in Q. 2. The above solution was obtained using MATLAB.Exercise 3-97 Given: tem 5 of Table 3.2 with afb =2.0 Required: A computer program to calculate the shape factor. | Assumptions: 1. Steady-state conduction, lee 2. Constant thermal conductivity. Y | Because of geometric symmetry, only 1/4 of the area need be considered. The temperatures of outer wall, Ty and inner wall, Ty2 are constant. Choose Ty1 = 100°C, Tw2 = 0°C, and Ax= Ay. There are M node points in the x-direction and N(=M) node points in the y-direction. ‘Temperatures at interior grid points using Eq. (3.85) are Tan = Taant*Tar ta Tae tTt ; o ‘Temperatures along line B-C are obtained from item 3 of Table 3.7 with q, = 0 (due to symmetry) M-1 Tan =F aris tTonintTannV2; 2¢ms Mt o ‘Similarly, the temperatures along line E-F are 1 N-1 Tan= 5 Tana + Tite Tine 21; 2¢ns SE ® After the temperatures at nodes are obtained, the heat flow can be evaluated as @ where Tyy is the average temperature along line G-H-I and A/Ax ~4ML, Q= KML (TaeTea) ‘Then the shape factor is calculated as © From item 5 of Table 3.2, the shape factor is3-97 (continued) s. nL. ~0.93in(afo)-0.0502 = 10.570L Results obtained using Gauss-Seidel iteration with a convergence criterion of € = 1x10™, and a flow chart are shown. Nodes (MxM) | S.a | (Sca-SVS xl 13.443 | 0.272 21x21 12.032 | 0.138 31x31 11.531 | 0.0909 4x4 11.267 | 0.0659 51x51 11.102 | 0.0503, 101x101 | 10.793 | 0.0211 nite ona Tan I et bundy ‘calculate Ta aiong fine Sing me oxo ‘Comments: 1. A rather fine mesh of 51x51 is necessary to be within ~5% of the exact solution, 2. When possible it is advisable to perform an energy balance on the numerical solution to check its validity. In this exercise such a check requires evaluation of the heat flow across B-A-F and comparing to the value obtained for G-H-I. 3. Asan altemative to Eq. (4), use Eq. (3.88) to calculate the heat flows, and compare results.Exercise 3-98 Given: Two-dimensional steady conduction. Required: Derivation of finite difference formulas for boundary conditions given as Itéms 3,4,5in Table 3.7. Assumptions: 1. No intemal heat generation. 2. Constant thermal conductivity. @ Item 3. Prescribed heat fux q, This result can be obtained directly from Eq. (3.88); solving for To, A net 1 erat) 4 2 ‘ ae 2k : \ 2 Gi) Ttem 4, oe! comer in contact with a uid, ‘An energy balance on the finite control volume shown in the figure requires Net heat conduction , heat convection _ “iy {into volume into volume Glow =e EE Ae 7 7 aectt 2 Te-To Ax ay 2 ( Galo =neeTe—Tay lay Oyo = hee Tay ‘Substituting in the energy balance and taking Ax = Ay, ~E tots Ta)] + AAT Te) =0 Solving for To, Fat BiTe me Ba ‘14Bi i k iii) Item 5: Interior comer in contact with a flui Net conduction , heat convection _ ‘into volume * into volume To-Tr Ay Ola =-k To= so alan =k“ ay To-Ts Ax Otay =k tS3-98 (continued) TrTo ay Quo = he(TeTa) Glan =k Ax Oyo =ne(Te=Ta) ‘Subtituting in the energy balance and taking Ax = Ay. He |F Clo Ty Chy To) + 5 Lo Ta)-(TaTo)] + HeANCTeTo) =O Solving for Tp, Te#Ts + CsTsBiT, =‘ a = 34Bi hAxExercise 3-99 Given: ‘Two-dimensional steady conduction. Required: Derivation of finite approximation for an interior node near a curved non-isothermal surface (Item 6, Table 3.7). Assumptions: 1. No intemal heat generation. 2, Constant properties. ‘An energy balance applied to the control volume shown in the figure requires that Net conduction _ g into volume T.-To aAx 2 TiTo {14 ax [2 12 cao +k Dividing by (142)(140) of and solving for, ee alae D(l+b) eaExercise 3-100 Giver: A straight rectangular fin. Required: 1, A computer program to solve for temperature distribution and fin efficiency. 2. Comparison with the analytical formula given in Exercise 3-13. Assumptions: 1. Two-dimensional steady-state conduction. 2. Insulated fin tip. i Choose a base temperature Tg = 100°C and an ambient temperture T, = 20°C. Use a mesh with Ax = Ay as 2 shown. There are M node points inthe x-direction and 123 n ‘Nnnode points in the y-direction. At interior nodes, Eq. (3.85) is used as Tan = 2 Caan TortaTeet Tata) © At convective boundaries, Eq. (3.86) is used Taa= ort MoztTotitTmi*BiTe], 2¢msM-1 2 1 1 . Tan = pj Tana +5 To tntTaeinBiTe] , 2SmSM-1 @) where mesh Biot number Bi =h,Ax/k. At an insulated boundary, Bi Tua= 7 Mets +t CrenitTyni)l, 28nsN-A ® Following Exercise 3-78, item 4, the temperatures at nodes (M,1) and (M,N) are derived to be Twa Spr i reas*Tua)+ > on 1 1 1 Tax = Ta [s Chetty an © After the temperature distribution is obtained, the average fin-side surface temperature is ‘evaluated using the trapezoidal rule as ey Mt Te eq [2 THT + Z Tos Hence fin efficiency can be evaluated as3-100 (continued DL Te—Te) _ Ty-Te Net ohLctp—T) ~ Ta~Te ‘The fin efficiency from the analytical formula given in Exercise 3-13 is &p an 0.809 Mesa = FLAT ‘Sample results are given in the Tabie (e = 1.0x10~) T.) Nodes (MN) | nea | 1-H) Tlexact 13x5 0.802 0.86 19x7 0.804 0.62 25x9 0.805 049 31x11 0.805 0.49 37x13 0.806, 0.37 Comments: 1. Referring to Comment 1 after Exercise 3-13, the heat flow in this fin is ( almost 1-D since Bi =h,k = (4)(0.05)/1 = 0.2 is relatively small. Rework the exercise with h, = 100 Wim? K so as to have a large 2-D effect.Exercise 3-101 Given: Two-dimensional steady conduction. Required; Finite difference approximation for a convective boundary with intemal heat genera- tion. Assumptions: 1. Constant properties. ‘An energy balance applied to the control volume shown in the figure requires that, Net heat conduction, heat convection , intemal heat _ into volume —* across boundary * generation = kia) 2B (acl) Dividing by k and solving for T;,., 1 Ax 6 Trae Tag +3 Tint Tine) #BiTe + EO anes © 24Bi where Bi=h,Ax/k.Exercise 3-102 Given: A 4 cm square plate with edge temperatures of 60°C, 40°C, 20°C and O°C. Required: Use of Gauss-Seidel iteration to solve finite difference equations on a1 cm square nodal mesh, and give result after Ist and 2nd iterations. Assumptions: 1 conduction problem. ‘Temperatures are uniform across the plate to give a two-dimensional 2. Thermal conductivity is constant. The figure shows the specified temperatures along the edges, and the nodes. From Eq, (3.85) the interior nodal temperatures are given by: V/A) Tp+T,+40460) WANT +Ty4T 5460) 14) (T,+T 6420460) 1A) +47 440) V4)(Ty+T AT #T9) ANT #1545420) 1A) T #74040) For an initial guess, assume approximate linear variations in the x- and y-directions, as shown. ‘Then the first iterate is (/4)B7.5432.5440460) = 42.50°C (1/4)(42.5+30+30+60) = 40.63°C (1/4)(40.63+27.5+20+60) = 37.03°C (1/4)(42.50+30+25440) = 34.38°C (1/4)(40.63434.38427 5422.5) = 31.25°C (1/4)(37.03+31.25420+20) = 27.07°C (1/4)(34.38422.540440) = 24.22°C (1/4)(31.25424.2242040) = 18.88°C ‘Ty! = (1/4X27.07+18.88+0420) = 16.49°C Comments: 1. iteration, ‘The second iterate is = (1/4)(40.63434,38+40+60) 1/4)(43.75+37.03431,25+60) = 43. 1/4)(43.01427.07+20460) = 37.52°C 1/4)(43.75431.25+24.22440) = 34.80°C TT? = (1/4)(43.01+34.80427.07+18.88) = 30.94°C Te = (14\37.52+30.94416 49420) = 26.24°C 1/4)(34.80+18.88+0440) = 23.42°C 1/4)(30.98423.42+16.49+0) = 17.71°C To! = (1/4)(26.24417.71+40420) = 15,99°C Notice how the latest available temperature is always used in Gauss-Seidel 2. The exact result for the center temperature is the average of the boundary temperatures, T, = (1/4)(60+40+2040) = 30°C: observe how successive iterates approach this value. 3. The comer nodes can be assigned temperatures that are an average of the adjacent edge temperatures, e.g. at the top left comer T = (1/2}(60+40) = 50°C. Show that this choice ensures a zero net heat flow through the outside edges of the comer finite control volume. However, these comer temperatures are not used in the solution procedure, and do not affect the interior node temperatures.Exercise 3-103 Given: A 4 em square, 3 mm thick plate with edge temperatures of 60°C, 40°C, 20°C and 0°C. Required: (j) ‘Temperature distribution. (i) Effect of mesh size on calculated heat flow out of the O°C edge. Assumptions: 1. Side surfaces of plate well insulated to give two-dimensional heat flow. 2. Constant thermal conductivity, k = 200 W/m K. For a square mesh, Eq. (3.85) gives the temperatures of the interior nodes, Ty = Tenet Tineta Tat Tmt) ‘The edge temperatures (excluding the comers) are Tray = 60°C, m=2,...MoL Ta = 0°C, m=2,...MeL Thy = AOC, BZN Tray = 20°C, n=2,...NeL “The comer temperatures are taken as averages of the adjacent edge nodes, Ty, = (0.5)(0440) = 20°C; Ty = 50°C, Ty. = 10°C, Tay = 40°C | For a 10 mm mesh size and a convergence criterion of 10 °C the following temperature distribution is obtained with 20 iterations. 50.00 60.00 60.00 60.00 40.00 40.00 44.29 4286 37.11 20.00 40.00 3429 30.00 25.71 20.00 40.00 22.86 17.14 15.71 20.00 30.00 0.00 0.00 0.00 10.00 The heat flow out the bottom edge is obtained using Eq. (3.88). For a plate thickness t, i Q = CAOWADY fPTp Tash nT) mt where the contribution from the comer control volumes has been ignored. The table shows the effect of mesh size on T (x = 1 cm, y =3 cm), Q, and the number of iterations required to obtain 10% K convergence in the temperature field.3-103 (continued) Comments: 1 2. Mesh size, mm 10 5 2 1 T (om, 3 em), K 4429 44484455 44,55 Heat flow, W 4243 58.12 79.05 -94.92 Number of iterations 20 16 408 1410 The temperature field is seen to be converging to grid-size independent values. On the other hand, the heat flow increases with decreasing grid size. behavior is due to an ill-posed problem. A temperature discontinuity at a comer implies an infinite heat flow—which the numerical solution approaches as the grid size is reduced. (See the discussion following Eq. (3.30) in the text.) Thus, the calculation of Q is physically meaningless, ‘The comer temperatures do not enter into the calculation of the temperature field. In the case of a square mesh we assign the comer the average of the two adjacent surface temperatures: this choice ensures a zero net heat flow into the comer volume. The comer temperatures are used in the calculation of Q, but, as mentioned above, this calculation is physically meaningless.Exercise 3-104 Given: An 8 em by 4 om plate with prescribed edge temperatures. Required: Use of Gauss-Seidel iteration on a 2 cm x 1 cm mesh to solve for the steady temperature distribution. Assumptions: 1. Plate faces well insulated--two dimensional conduction. 2. Constant thermal conductivity. ‘The figure shows the prescribed temperatures along the edges. Equation (3.84), gives the interior nodal temperatures. For a geometric mesh factor of B (away) = 4, 20+4)T,,, = T, iT, mac *T mast) = ONT y*Te ng) ANOT ag 1 Tp) For an initial guess assume approximate linear variations as shown. Then, taking advantage of the symmetry the first iterate is T, = (1/10)(20+30) + (4/10(40+20) = 29.0 Tz = (1/10)(29.0429.0) + (4/10)(40420) (1/10)(20+20) + (4/10\29.0+10) = 19. (1/10)(19.6+19.6) + (4/10)(29.8+10) = 19.84 (1/10)(20+10) + (4/10)(19.640) = 10.84 Te = (1/10)(10.84+10.84) + (4/10)(19.8440) = 10.10 ‘The second iterate is T, = (110)(20+29.8) + (4/10)(40+19.6) = 28.82 Ty = (1N0)(28.82+28.82) + (4/10)(40+19.84) T, = (1/10)(20+19.84) + (4/10)(28,82+10.84) T, = (1/10)(19.85419.85) + (4/10)(29.70+10.10) T, = (1/10)(20+10.10) + (4/10)(19.8540) = 10.95 Te = (110)(10.95+10.95) + (4/10)(19.89+0) = 10.15 Comments: 1. Notice how the latest available values are used in Gauss-Seidel iteration. 2. The comer temperatures are not required to calculate the interior temperatures. In the case of a square mesh we may assign the comer the average of the two adjacent surface temperatures: this choice ensures a zero net heat flow into the ‘comer element. An energy balance on the comer volumes will show that comer temperatures of 24°C and 16°C are appropriate for this Aw/Ay = 2 grid,Exercise 3-105 Given: An 8 cm by 4 cm plate, 1 cm thick, with prescribed edge temperatures. Required: Computer program based on Gauss-Seidel iteration to solve for temperature’ distribu- tion. Effect of mesh size on calculated heat flow through the top edge. Assumptions: 1. ‘Two-dimensional, steady conduction, 2. Constant thermal conductivity. For a rectangular mesh with B = 4, Eq. (3.84) gives the temperatures of the intetior nodes as 7 Tygsy = DMO T ye *F pet 9) AO) Ty gt Tyan et) ‘The edge temperatures (excluding the comers) are Tran = 40°C, m=2,..Me1 Ta = 20°C, m=2,...M-L Ty = O°C, 1=2,..4Ne oe Than = 20°C, n=2,...NE mt The boundary conditions imply temperature discontinuities at the comers. In order to choose appropriate comer temperatures we will derive a finite difference approximation for heat flow through the comers. Using the notation of Table 3.7, an energy balance on the control volume at the top left comer gives, Ax TiTo Ay _, ToT Ax foe uot ax 2 ay 2 AS $5. For Away = 2: q,.Ay/2-qgAn/2 = k(2.5Ty-2Ty-0.5T\V2 As for a square mesh, we choose Ty to give a zero net heat flow at the comer volumes (q,,Ay/2 = q,yAx/2). ‘Then 2.5Ty = 2T,40.5T, = (2)0)4(0.5)(40); To = 8°C = Ty Similarly, T,, = 4°C, Ty, = 20°C, Tyyy = 24°C y fe 5, Aa we yi T i | 1 For a2 cm by 1 cm mesh the following temperature distribution is obtained with a convergence criterion of 0.005 K. The boundary conditions imply temperature discontinuities at the comers.3-105 (continued) 8 40 40 40 24 0 «2785 3285 3196 20 0 2140-2717 2670 0 = 1887 -23.06 2298 20 4 20 20 20 20 ‘To attempt to calculate the heat flow through the top edge, we rederive Eq. (3.88) for Ax = 2Ay to obtain, Ml Qo = DO aAxt = ae [2ST a y-2Tyy-a 0-25 Tyr T med] where we have ignored the contribution from the corner volumes. The table shows the effect of mesh size on the center temperature and the heat flow. Mesh size, cm Center temperature, °C Heat flow, W 2x1 20.17 1334 1x05 27.63 -186.4 0.25x0.125 21.79 -292.3 Comments: 1. The center temperature is clearly converging with decrease in mesh size. 2. The heat flow is increasing rapidly as the mesh size decreases, The reason is mentioned in the discussion following Eq. (3.30) in the text. The temperature discontinuity at the comers implies an infinite heat flow and is physically unrealistic. As the mesh size is decreased, Q will increase without bound. Notice. that the assigned comer temperatures do not enter into the calculation of the. interior temperatures, but do affect the heat flow. However, as stated above, the heat flow calculation is not physically meaningful.Exercise 3-106 Given: Long, 4 cm square bar (k = 2 Wim K) with surface conditions as shown. Required: ‘Temperatures after the first and second iterations using Gauss-Seidel iteration on a 1 em square mesh, ‘Assumptions: 1. ‘The bar is long enough to ensure two-dimensional conduction. 2, The thermal conductivity is constant. The figure shows the prescribed surface condi- tions and the nodes. Symmetry allows us to consider just nine nodes. From Eq. (3.85) the interior nodal temperatures are T, = (1/4\(100+T,+T AT) T, = (14 \10042T,+T Ty = (MAYT AT AT AT) Ty = (14)(T 427 4T,) Ty = (AT HT A Ty40) Ty = (14) Te 2T +0) Table 3.7, item 2, gives the finite difference approximation for convection at a plane surface. ‘The mesh Biot number is Bi = h.Awk = (100)(0.01)/2 = 0.5 T, = (/2.5)(T,+0.5)(100+T.)H0.5)(0)] = 04T,+0.27,420 Ty = (1/25) Ts#0.5)(T,4T,)+(0.5(0)] = 0.4T540.2(T,4T;) T, = (1/2.5)[Ty+(0.5)(T,+0)+(0.5)(0)] = O4Ty+0.2T, For an initial guess use the temperatures shown in the figure. Then the first iterate is T= (0.4)(60) +(0.2)(30) +20 = 50°C Tz = (0.25)(100+50+75440) = 66.25°C Ty = (0.25)(100+2x66.25+50) = 70.63°C Ty = (0.4)(40)+(0.2)(50+10) = 28.0°C Ts = (0.25)(66.25 +28.0+50+20) = 41.06°C (0.25)(70.63 +2x41.06+25) = 44.44°C (0.4)(20) + (0.2)(28.0) = 13.6°C Ty = (0.25)(41.06+13.6+25) = 19.92°C3-106 (continued) Ty = (0.25)(44.44+2x19.92) = 21.07°C ‘The second iteration is TP = (0.4)(66.25) +(0.2)(28.0) +20 = 52.1°C TP = (0.25)(100+52.1+70,63+41.06) = 65.95°C T} = (0.25)(100+2x65,95+44.44) = 69.08°C T} = (0.4)(41.06) +(0.2(52.1+13.6) = 29.56°C T? = (0.25)(65.95 +29,56+44.44+19.92) = 39.97°C T} = (0.25)(69.08 +2x39.97+21.07) = 42.52°C T} = (0.4)(19.92) +(0.2)(29.56) = 13.88°C Te. = (0.25)(39.97+13.88+21.07) = 18.73°C TP = (0.25)(45.52+2x18.73) = 20.75°C Comments: 1. Do not agonize over the initial guess~just use something reasonable, No matter hhow bad the initial guess, the Gauss-Seidel method will converge to the proper solution. 2. Notice how the latest available temperatures are always used in Gauss-Seidel iteration.Exercise 3-107 Given: Long, 4 cm square bar with surface conditions as shown. Required: Computer program to solve for steady temperature distribution, and effect of mesh size on heat flow into the side at 100°C. Assumptions: 1. Steady, two-dimensional conduction. 2. Constant thermal conductivity of 2 Wim K. For a square mesh, Eq. (3.85) gives the temperatures of the interior nodes, +f, Tayi tl, ‘mast ‘mn-Tn-ta) ‘The surface temperatures on the top and bottom sides are Try = 100°C; Ty, = °C at Ta100 mt Tao ‘Table 3.7, item 2, gives the finite difference approximation for convection from a plane surface. The fluid temperature , = 0°C, and the mesh Biot number is, Bi = h.Awk = (100)4x/2 = SOAK For a 5 mm mesh size and a convergence criterion of 10° °C (89 iterations) the resulting temperature distribution is (for the left half) 100.000 100.000 100.000 100.000 67.060 76.458 = «80.573 -82.391 48.853 58.199 63.445. 66.077 36.383 44.040 48.932 51.596 26.795 32.646 36.648 38.939 18.906 23.103 (26.079 (27.832 12079 14.781 16.734 17.903 5.889 7.211 8175 8.757 0.000 0.000 = 0.000 0.000 ‘The temperature distribution is shown in the graph. 100.000 82.915 66.876 52.437 39.680 28.409 18.291 8.951 0.000es 3-107 (continued) ta ts L ‘ax2) To calculate the heat flow through the top we must derive appropriate formulas for the comer nodes. Using the notation of Table 3.7, the net heat flow into the top left comer volume is x Dito Ay yy Ta7To Ax , Ande iatPAy, bet DY 49,5 = 0 Solving for a, noting that Ax = Ay, and T; = Tp since the top is isothermal, 1) 4 (1,-1,); (@, positive) Ax 8 dana = Tay “Tyy-) +50AT) xT.) From Eq. (3.88), for m = 2,...M-1: 1 : am wAX = 22a Tan-t 7 Taw *Tnet, | (4, positive) axe Ax Quan = Gant dans * day Se i the heat flow per unit length ad Also, div = damn by symmetry. The effect of mesh size on Q,,, is given below. Ax(mm) 200 SB Quop (Wim) 460 3924 387.6 385.5 Comments: 1. The heat flow is seen to be converging to a grid-size independent value.Exercise 3-108 Given: Long, 8 em square bar with surface conditions as shown. Required: (i) Finite difference approximation at top comers. (i) Temperature after the first and second iterations using the Gauss-Seidel’ method. Assumptions: 1, Steady, two-dimensional conduction. 2. Constant thermal conductivity of 0.08 W/m K. | ( Consider the left top comer as shown, An energy balance on the finite control volume requires that the net heat flow into the volume equal zero. TyTy Ax ay 2 HTT: Qo = 44 2 ‘Substituting in the energy balance and taking Ax = Ay, k, ax Ax Ket) he GOT AS #0 : 1 Ax hAx Solving, T, = saloon, E | Bi= E (20)(0.02)((0.08) = 5; 1/(2+Bi) = (1/7) qb = (80)(0.02)(0.08) = 20°C Initial guesses for the node temperatures are shown in the figure. The first iterate is: Ty! = (1/7)[(60+30}420] = 15.71 Using Table 3.7, item 3, Ty! = (1/2)(45)H(1/4)(15.71480)410 = 56.43°C3-108 (continued) Ty! = (1/2)(60)+(1/4\(56.43456.43)+10 = 68.21°C Using Table 3.7, item 2 for the side surface nodes, and item 1 for the interior nodes, = (W/D[45+(12)(15.71420)] = 8.98°C Ts. = (1/4)[56.43 +8.98+30+60} = 38.85°C Tg. = (1/4)[68.21 +38.85+38.85+40] = 46.48°C Ty = (1/7)[304(1/2)(8.98+10)] = 570°C ‘Ty = (1/4)[38.85+5.70+40+15] = 24.89°C Ty. = (1/4)[46.48+24.89+24.89+20] = 29.06°C Ty = (UN[15+(1/2)(5.70+0)] = 2.55°C Ty = (1/4)[24.89+2.55+0+20] = 11.86°C Ty) = (1/4)[29.06+11.86+11.86+0] = 13.19°C A second iteration gives (IT)[(56.43+8.98)+20] = 12.20°C = (1/2)(38.85)+(1/4)(12.20+68.21)+10 = 49.53°C = (1/2)(46.48)+(1/4)(49.53+49.53)+10 = 58,00°C “2 = (1/7)[38.85+(1/2)(12.20+5.70)] = 6.83°C T? = (1/4)[49.53+6.83+46.48+-24.89] = 31.93°C T? = (1/4)[58.00+31.93+31.93+29.06] = 37.73°C T} = (1/1){24.89+(1/2)(6.83 42.55)] = 4.23°C = (1/4)[31.93+4.23+29,06+11.86] = 19.27°C Te = (1/4)[37.73+19.27+19.27+13.19] = 22.36°C Ti = (/DI11.86+(1/2)¢4.23+0)] = 2.00°C Ty, = (1/4)[19.27+2.00+13.1940] = 8.61°C Ti = (1/4)[22.36+8.61+8.61+0] = 9.90°C Comments: 1. The temperatures are falling as the iteration proceeds, indicating that the initial guesses were high-but this is of no real concern. 2. Notice that the latest available values are used in Gauss-Seidel iteration,Exercise 3-109 Given: A long, 8 om square bar with surface conditions as shown. Required: Computer program to solve for the temperature distribution with h, = 20 W/m? K, and heat flow out bottom side for mesh sizes of 1 cm, 0.5 em and 0.2 om. Assumptions: 1. Steady, two-dimensional heat conduction. 2. Constant thermal conductivity. For a square mesh, Eq. (3.85) gives the temperature of |" T the interior nodes, -— tis ---4 wn = Naa Taig Fant Ta) ! ‘The surface temperatures on the bottom are \ = OC, m=1,..M. et wal Table 3.7, item 3 gives the finite difference approximation for an imposed heat flux; thus on the top. q.Ax/2k = 80Ax/(2)(0.08) = 500Ax 1 He Thay = 5 Tones Toaty * Tuy) “500A, m=2,..M-1 (On the sides, item 2 is used. = h,Awk = 20Ax/0,08 = 250Ax The - Fy ama 2a feat Thal 1=2suN-l 1 Tye = =e sN-1 Ma 22500x Orang Ea Tals 9° ‘An energy balance on the top left comer control volume gives T, Toy, 47 ye-2)* IN” Sa5pas — Feaspag Ton * Tins) + 1000 Ax] i 1 Similarly, Ty. = Sa Tew +Tyy-1) +1000 Ax] ‘The heat flow at the bottom is obtained using Eq. (3.88) and an energy balance for the comer control volumes, For the left comer volume.3-109 (continued) 4 k Ax Que = FA a-Ty he MeN) = 0.04, 2 Similarly, Qge = 0.04T 2 Mel Q = 0041, 2 +42) 008 YS (2 -Tya 5 la Tm % ‘The tables show the effect of node spacing on heat flow, and the temperature distribution for Ax 1.05 cm. ‘The convergence criterion was 0.001 K. 777 gq 7 7 a 7a] a a a] 7 5 77] 2.00} 0.60 '6,00} 0.00] 0.00] 0.00] 0.00) 0.00] 00] Tro 0.00) ‘.00} ‘o.09] 2.09] 0.09) 2.09) B09) 0.10} 0.23 0.35] 046] 0.56] 062] 0.68) 0.71] 0.73] 7 0.65] 0.62)" 0.55] 0.46] 0.35] 023) ‘0:%0} 02%) 0.47 ori] 093] _t.11 127] 1.38) 1.45] _1.47| 745 39 127] 11] 0.93] ort 0.47] 021 0.32) 0.72 4.09|_143]_1.72| 1.98] 2.13] 223] 227] 223 2.13] 1.95] 772) 1.43 7.09 072] 0.39} 0.44] 0:99] 151 197] 237] 270] 2.04 3.08] 3.13] 3.09 2.70] 2.37] 137] 151 ‘0.89] 44] (058) 1.30 197} 258] 311] 3.53] a4) 4.03] 4.10] 4.03 384] 3.53] 3.14 2.58] 197 730] ‘06 0.74] 1.65] 251] 329] 304) <43| 487] 6.11] 6.19] EM 487] <9] 3.04] 3.23] 254 165 074] (0.82) 2.06] ‘312| 4.06] 4.90] 6.56] 6.05] 6.34] 6.44] 34] 6.05] 3.35] 4.99] 4.08] 3.12] 2.06] 0.83] ‘of 1.14] 2.54] 3.85] 5.02| 602] ~6.83| 7.41] _7.77| 7.09] 777 74 683] 6.03} 6.02] 3.85] 24] ERT) To] 4.40] _3.14 47i] 6.13] 734] 831] 9.0%] 0.43] a.s0] 3.43] 2.04 at 734] 5] ari 3a 740] tif 1711320] ‘S74] 7.46] 8.91] 10005| 10.88] 11.28] 11.56] 71.39] 70.89] 70105] 3a 7.46] 374) 3.80] 77 12] 2.09] 4.66] 7,00] 9.06] 10.77| 12.12] 13.06] 13.66) 13.09] 13.66] 73.08] 12.19] 7077] 2.06] 7.09) 4.65] 2.09) 13[ 2.57] 6.71 8.85] 11.00] 13.00] 14.6] 15.67] 16:53] 16.55] 16.33) 15.87] 14.56] 73.09] T1.00] 3.55) B71 257] 14] 3.20] 7.09] 7050] 13.37| 15.49] 17.46| 16.70] 19.44] 1.69] 79.44 18.70] 179] 75.69] 13.37] 70.59) Tor] 320) 75| 4.05] 690} 12.99] 16.31] 18.91] 2087| 22.24] 23.04] 23.31 23.04 "22.24 20.87 7894 16.31 12.99] 2.80) 406] 16] 5:37] 11.47] 16.25] 19.95] 22.79] 24.89| 26.33] 27-18] 27.46] 27.79] 26.33] 24.89 2279) 718.55] 7625] 7.47 637| 17| 7.92] 15.35 20.60) 24.47) 27 40] 2685] 31.03) 31.89] 92.16 34 31.03] 29.5] 27.40] 24.47 20.55] 15.36] 722] Q. Wim Iterations Comments: 1. 0.01 0.005 0.001 0.612 0.626 0.648 96 3214410 ‘This solution was obtained using MATLAB. ‘Consult a text on numerical methods for methods used to determine how a finite difference solution converges to the exact solution as time step and mesh size are decreased. Note that the above solutions have converged to within 0.001 K of the solution of the difference equations, not the differential heat conduction equation.Exercise 3-110 Given: A 3 cm square bar with surface temperatures as shown, Required: (i) Gauss-Seidel iteration to obtain the steady state temperature distribution on a 1 em square mesh. (ii) For k= 1 Wim K determine the heat flows across each surface and check energy balance. Assumptions: 1, End effects negligible to give a 2-D problem. 2. Conductivity is constant. (@) Symmetry requires us to consider only two unknown nodal temperatures, designated T, and T,,. Initial guesses, are T,° = 30°C, T;° = 20°C (by rough interpolation). Both are interior nodes and the finite difference Bee eee, approximation is given by item 1 of Table 3.7. Then T, = (1/4)(60+30+T, +T,) t 1 ' t ° T, = (AT, +30+T, +0) / ‘Of course, itis easy to solve these equations directly for T, and T;, But, to illustrate the use of ‘Gauss-Seidel iteration and its convergence, we will use the iterative method. T= (1/4)[60 +30 +30 +20} = 35.00°C Ty = (V/A)[35 +30-+20+0) = 21.25°C T} = (W/4)[60+30+35 +21.25] = 36.56°C Ty = (V/4)[36.56 +30+21.25 +0] = 21.95°C T) = (1/4)[60+30+36.56 +21.95] = 37.11°C T} = (1/4371 +30+21.95 +0] = 22.27°C Ty = (1/4)(60+30+37.11 +2227] = 37.34°C Ty = (1/4)[37.34 +30 +22.27 +0] = 22.40°C Ty = (1/4)[60 +30 +37.34+22.40] = 37.44°C T) = (1/4)[37.44 +30+22.40 +0] = 22.46°C3-110 (continued) Tf = (1/4)[60 +30 +37.44 +22.46] = 37.48°C TE = (1/4)[37.48 +30 +22.46 +0] = 22.48°C = (1/4)[60 +30 +37.48 +22.48] = 37.49°C ‘Ty = (1/4)[37.49 +30 +22.48 +0] = 22.49°C = (1/4)[60 +30 +37.49 +22.49] = 37.50°C TS = (1/4)[37.50+30+22.49 +0) = 22.50°C T) = (1/4)[60+30+37.50+22.50] = 37.50°C Ty = (1/4)[37.50 +30+22.50+0] = 22.50°C and the solution has converged to 0.01°C. (@) First we will derive a finite difference approximation for heat flow through the comers. Using the notation of ‘Table 3.7, the net energy flow into the top left comer volume is ay TT Ay, -k 0 ey oF ax ‘Taking Ax = Ay and solving Gea Gey = WAX)L2T,~(T,4T,)] = .0/0.011(2)(45)-(60+30)] =0 ‘We see that our practice of assigning the comer temperature equal to the average of adjacent surface node temperatures results in a zero net heat flow into the comer control volume. Thus, for an energy balance we need consider only the remai 3.88), two nodes on each surface. Using Eq, Quop = (2)(1.0/0.01)[(2)(60) - 37.50 - 1/2(45+60)] = 6000 Wim Queuom = (2)(1.0/0.01)[(2)(0) -22.50 - 1/2(15+0)] = -6000 Wh Qiidas = (2)(1.0/0.01){(2)(30) - 37.50 - 1/2(30+45) + (2)(30) -22.5 - 1/2(30+15)] = 0 Wim3-110 (continued) Comments: 1 Although we have a converged solution and an exact energy balance, our result is not accurate—there are too few nodes to accurately calculate the temperature distribution. How do we determine the true heat flow through each side? Our methodology does not allow us to calculate the heat flows into the comer control volumes—it only tells us that qo. = yy. At first sight we might expect to obtain an accurate enough answer by reducing Ax until the contribution to Q from the comer volumes is negligible. However, in this problem the discontinuity in surface ‘temperatures at the comers implies an infinite heat flow. (See the discussion following Eq, (3.30).) Reduction of Ax results in a continuously increasing value of Q. Although a physically valid temperature distribution can be obtained, there is no physically meaningful heat flow.Exercise 3-111 Given: A 3 cm square plate with specified edge temperatures of 30°C, 20°C, 10°C and 0°C. Required: Gauss-Seidel iteration to solve finite difference equations on a 1 cm square nodal mesh. Assumptions: 1, Side faces perfectly insulated to give a uniform temperature across the plate thickness. 2. Constant thermal conductivity. ‘The figure shows the edge temperatures and the nodes. y From Eq, (3.85) the interior node temperatures are given t by: 7 T, = (1/4)(30+20+T,+T,) = (1/4)(50+T,+T,) T, = (1/4)(30+T, +104T,) = (1/4)(40+T,+T,) T, = (1/4)(T,+20*T,+0) = (1/4)(20+T, +T,) 7 T, = (/4)(T,#T,+10+0) = (1/4)(10+T, +7) Tnitial guesses are based on approximate linear interpola- 2» tion and are shown in the figure. The iterations give (1/4)(50+15+10) = 18.75°C Ty = (1/4)(40+18.75+5) = 15.94°C. 10 Ta: = (1/4)(20+18.75+5) = 10.94°C Ty = (1/4)(10+15.94+10.94) = 9.22°C T} = (1/4Y50+15.94+10.94) = 19.22°C T} = (1/4)(40+19.22+9.22) = 17.11°C TP = (1/4)(20+19.22+9.22) = 12.11°C Ty = (/4)\10H17.11612.11) = 981°C (1/4)(50+17.11+12.11) = 19.81°C (1/4)(40+19.81+9.81) = 17.40°C = (1/4)(20+19.81+9.81) = 12.41°C T? = (1/4)(10+17.40+12.41) = 9.95°C3-111 (continued) Ty = (1/4)(50+17.40+12.41) = 19.95°C = (1/4)(40+19.95+9.95) = 17.48°C (1/4)(20+19.9549.95) = 12.47°C Ty = (1/4)(10+17.48+12.47) = 9.99°C Ty = (1/4)(50+17.48+12.47) = 19.99°C = (1/4)(40+19.9949,99) = 17.50°C ‘Ty. = (1/4)(20+19,99+9.99) = 12.50°C (1/4\(10+17.50+12.50) = 10,00°C Comments: 1. A few further iterations will show that the converged solution is T, = 20°C, Ty = ITS°C, T, = 12.5°C, Ty = 10°C (exactly). Notice that (1/4)(T,+T4T,4T,) = 15°C, which equals the average of the edge temperatures, 1/4(30+20+10+0) = 15°C, 2. Although our solution has converged, it is not accurate because the mesh is too coarse. However, the midpoint temperature of 15°C is exact. Indeed, it can be shown analytically that the midpoint temperature is equal to the average of the edge temperatures (see Figure 3.7). 3. Notice how the latest available values are used in Gauss-Seidel iteration. 4, Comer temperatures were given values equal to averages of adjacent surface nodes. The resulting heat flow out of the comer volume is zero (check by pesforming an analysis similar to that used to derive Eq. (3.88). However, the comer temperatures are not used to calculate the interior temperatures.Exercise 3-112 Given: A 3.cm square plate with specified edge temperatures of 30°C, 20°C, 10°C, 0°C. Required: Solution of finite difference equations on a 1 cm mesh using a direct method. Assumptions: 1. Sides are perfectly insulated to give a uniform temperature across the plate thickness. 2. Constant thermal conductivity. ‘The figure shows the edge temperatures and the nodes. From Eq. (3.85) the interior node temperatures are 47, = 30+20+T,+T, 4T, = 30+T,+10+T, 4T, = T,+20+T,+0 4, = T,*T, +100 Rearranging in the standard form for a set of linear equations, 47,-T,-T, = 50 1,47, -T, = 40 1, +47,-T, = 20 -T,-T,+4T, = 10 (i) Hand-held calculators have standard routines for sol 1g simultaneous linear equations for a limited number of unknowns (e.g. a Texas Instruments TI-68 can handle up to S unknowns). Using such a calculator the result is T, = 20°C; T; = 17.5°C; Ts = 125°C; T, = 10.0°C. (ii) In matrix form the equation system is 4-1-1 0|/T{ [50 [A][T] = [B] 14 0 -1| [To 40 aed He The solution is -10 4 -1)[t,] ~ |20 o -1-1 4) Ir, 10 7 = ABI where the inverse matrix [A]! can be obtained using a standard computer subroutine. Comments: 1. Direct methods are simple to use when the number of equations (nodes) are few. However, for the finite difference approximation to accurately represent the differential equation, many nodes must be used. Then direct methods can have large roundoff errors and iterative methods are preferred.Exercise 3-113 Given: Long 3 cm square bar with surface conditions as shown. Required: (j) Finite difference approximation at bottom comer. Gi) Temperatures for three Gauss-Seidel iterations. ii) Energy balance after third iteration. Assumptions: 1. End effects negligible and surface conditions constant to give steady, two- dimensional conduction. (@ Consider the left bottom comer as shown, An energy balance on the finite control volume requires that the net inflow of heat equal zero. To Ts Qulave * ~. Oily = HTT Ys Oh ‘Substituting in the energy balance and taking Ax = Ay, hAX(T,-T,)-1k(Te-T,) -K(T-T.)} = 0 Solving for T;, (5#T) +Bi 2B i) The figure shows the initial guesses. Also Bi 1/3. Using Table 3. h,Awlk = (200)(0.01)/(2) = 1.0; 1/(2+Bi) = ems 1, 2 and 3, together with the result of part (i) we proceed. T} = (1B)1404(1/2)(50+30)+(1)(0)] = 26.67°C T = (1/4)[50+26.67 +40+30] = 36.67°C 30 50 T= (1/3)[30+(1/2)(26.67+10)+(1)(0)] = 16.11°C fem t 1 T} = (1/4)[36.67+16.11+30+20] = 25.70°C ' ' 1 eo tyes! Td = (1)[16.11+20+(1)(0)} = i 204°C fee. res Te. = (1/2)(25-70)+(1/4)(12.04+20) = 20.86°C ! t 48 __-. Te = (18)[36.67+(1/2)(50+16.11) 41] = 23.24°C. T] = (1/4)[50+23.24+36.67+25.70] = 33.90°C Ie,=10 em rea3-113 (continued) T? = (1/3)[25.70+(1/2)(23.24+12.04)+(1(0)] = 14.45°C T2 = (1/4)133.90+14.45+25.70420.86] = 23.73°C T} = (1/3)[14.45+20.864(1)(0)} = 11.77°C 72 = (1/2)(23.73)+(1/4)(11.77+20.86) = 20.02°C TP = (1/3)133.90+(1/2)(50+14.45) +(1)(0)] = 22.04°C T} = (W/4)[50+22.04+33.90+23,73] = 32.42°C T} = (1/3){23.73+(12)(22.04+11.77)(1)@)] = 13.55°C T) = (1/4)[32.42+13.55+23.73+20.02] = 22.43°C T= (1/3)[13.55+20.02+(1}(O)] = 11.19°C 12 = (1/2)(22.43)+(1/4)(11.19+20.02) = 19.02°C Gi) Energy balance. First we need to derive a finite h difference approximation for the hea flow through the 1, eee eee top comers. Using the notation of Table 3.7, the energy | flow into the volume at the top left comer is | Dito Ay 4 ToT AX 7 ny AV -g SE = 0 my ol ax 2 Dysart 2 I For Ax = AY, “Ggy = Kites )-BiT,]; Bi = a ‘The heat flow into the two top comers is then Q = -2q,,Ax/2 = (2.0)f(2+1)(50)-(50+22.04)-(1)(0)] = 155.9 Wim For the central portion of the top edge. Eq. (3.88) gives Q = (2)(0.01)(2/0.01)[(2(50)-32.42-(1/2)(50+50)] = 70.3 Wim Quon = 155.9470.3 = 226.2 Win ‘The convection out the sides is obtained using Newton's law of cooling, = (2(200)(0.01)[(1/2)(50-0) +(22.04-0) +(13.55-0)+(1/2)(11.19-0)] = 264.7 Wim Qeond~Qeony = 226.4-264.7 = -38.3 Wim Comments: 1. "The energy balance indicates that many more iterations are required to obtain a satisfactorily converged solution, 2. Notice that the fatest available values are used in Gauss-Seidel iteration.Exercise 3-114 Given: Steady two-dimensional conduction with internal heat generation. Required: Finite difference approximation for the surface heat flux. Assumptions: 1. Constant thermal conductivity ‘The energy balance on a finite control volume of unit, depth adjacent to the boundary is Net conduction heat flux across _ heat generation into the volume ~ face at x=0 ~ within volume =0 te E I eee gl Oy = a g Qube = a.A¥ Q, = Q)"(Ax/2)Ay (for unit depth) ‘Substituting in the energy balance and taking Ax = Ay, aa Oa 1 Tag Oya Tye) Biat K 2k Solving for a,, 2 k 1 20, Tap 1 4a 10 Tan o Comments: 1. If a known surface heat flux q, is imposed, Eq. (1) should be solved for T,,, to give the required nodal equation (c.f. Table 3.7, item 3).Exercise 3.115 Given: A long 4 m square bar with volumetric heating Q.” = 3x10° W/m’, k = 15 Wim K, and surface conditions as shown. Required: () Finite difference approximations for interior and surface nodes. (ii) Temperatures after first and second iterations using Gauss-Seidel iteration on a 1 om square mesh. Assumptions: 1. Steady, two-dimensional conduction. 2. The thermal conductivity is constant. (@ Interior node: Set B = 1 for a square mesh in Eq. (3.84), Tra = UA Tgat9* Testa Tmo Tmag* Qy Ax2M) Side surface nodes: Include a volumetric heating source term into the derivation of Eq. (3.86). ‘The volumetric fon 4 heating for a volume Ax by Ax/2 is Q.”Ax2/2 per unit oe ; depth, Hence iste fa | Ag? QAR? we I 2k I a4 Solving for T,.. 1 1 Ax? = ry ty. y Typ) BIT, + 2 (T, ,-T,) +1500Ax(T, y-T,) , where T, = 0°C. on?3-116 (continued) ‘The tables show the effect of node spacing on the heat flow into the fluid, and the temperature distribution for Ax = 0.005 m. The convergence criterion is 0.001 K. x, m 0.01 0.005 0.002 Quon Wim 4004 3830-3730 erations 29 100509 70,00] 0.00] 0.00 0.00] 0.00] 0-00] 0.00] 0.00] 0.00] 70,79| 15.05| 17.78] 19.34] 19.81] 19.31] 17-78| 15.05] 10.79] 78.86] 26.63| 31.76] 34.67| 35.61] 34.67] 31.76] 26.63] 18.86) 25.24) 35.85| 42.95| 46.99] 48.30] 46.99] 42.95] 35.85] 25.24) 30.64) 43.59] 52.19] 67.05] 58.61] 57.05] 52.19] 43.59| 30.64] '35.77| 60.67| 60.19] 65.39] 67.04] 65.39] 60.19] 50.67| 35.77] 44,88] 68.13] 67.50| 72.30] 73.78] 72.30| 67.60] 58.13] 41.88] 52.36] 67.47| 74.37| 77.53] 78.46| 77.53] 74.37] 67 47| 52.36] 80.00] 80.00] 0.00] 80.00{ 80.00] 80.00} 80.00] 80.00] 80.00} (elalaolol= lolol] Also shown is the temperature profile on the symmetry axis for 3x10° < Q\” Wim? < 3x10", calculated with Ax = 0.005 m. 350 300 250 200 Temp (C) 150 5 Node A iteb #866 tO ed I OvEHeS Greed —e-Crete8 + Gre367| Comments: 1. The calculations were made using MATLAB.Exereise 3-117 Given: One-dimensional unsteady heat conduction. Required: Crank-Nicholson form of the nodal equation for an interior node. Assumptions: 1. Constant thermal properties. 2. No internal heat generation. Following the derivations in Section 3.5.2, ee o AU = pe(Ax-1-1)(TE'-Ta) Flatt etsy) (Geto ]at=-« 2 rapa : Fla TiadCTigh TE] (ts) t= 2 apt ‘Substituting in Eq. (1), dividing by At and rearranging, 1-2Fo ARF yi 142F aan [Bee tehetiaerith] + Comments: 1, For a given mesh size, the Crank-Nicolson method gives more accurate results than either the explicit or implicit methods, Although oscillations can occur, they never become unstable.Exercise 3-118 Given: Slab of resin, convectively heated on one face, as described in Example 3.13. Required: Effect of mesh size and time step on the accuracy of the solution. Assumptions: 1, As for Example 3.13. ‘The following table compares the surface temperature response for various combinations of Ax and At, with the exact solution. Axjem 2 1 os 1 1 1 Bi og 04 02 04 04 04 Fomax 0227 0357 0416 0.357 0357 0.357 Fo 0250.25 0.250.125 0.0625 0.03125 Atmin 347 0.867 0217 0433 0217 0.108 tumin T.°C Exact 0.00 20.00 20.00 20.00 20.00 2000 20.00 20.00 347 52.00 4687 4645 46.29 4597 45.81 46.32 693 5520 5378 5343 5347 $331 53.23 53.30 1040 59.52 58.10 5780 57.87 57.76 $7.70 $7.68 1387 6235 61.22 60.96 61.04 60.95 60.91 60.85 1733 64.63 6335 6342 6350 6343 63.39 63.32 34.67 7158 71.02 7088 7095 7091 70.89 70.82 52.00 7555 75.17 7507 75.13 75.11 75.10 75.02 69.33 78.47 78.17 78.09 78.14 7813 78.12 78.05 Comments: 1. Unless an accurate solution is required at very short times, the choice of Ax ‘has a more significant effect on accuracy than does At.Exercise 3-119 Given: Slab immersed in water with nucleate boiling on surface, as described in Example 3.14. Required: ‘Temperature profiles for (i) he = 280(TTq)” W/m? K, and Gi) he = 600(TTyay)!* ‘Wim? K. Assumptions: 1. Edge effects negligible to give a one-dimensional problem. ‘The problem set-up follows Example 3.14, which also gives a suitable flow diagram for a computer program. The Newton J]? iteration is implemented as follows: 2FO(T}1+BiT ay )4Th 14+2Fo+2FOBi F fe faa Tet Ax BAX(T{—Ty Bi= = Sent at where h, = B(T{?-T,4)" Hence 2Fo(T Es +BAxk NTT) Toa] Thl= : 142Fo+2Fol BANA) T} Tay)" aT 1400 "Toa Toa WT} cd Tyan) = 2FOBAx/k. We define aryl eei Tae Tet) pa gget 2D +00} Tea qa) Fer) = TY" - where TPT) = ced (TY Tyo”, CPi) = al +b Tae) Te TH ah a age Hence, Sep = dncTi a) = fi r ah i+. a Sree Taner Ta = i _2F_ | fait or! ff ‘Then using Newton’s method, TY"! = T} --3-119 (continued) ‘Sample results are shown below. @ he = 280(T—Tya,)? Wan? K Case 1. x=0.001 m,At= 1s T 6 0 0s it 10 16 15 2 20 500.00 500.00 390.64 469.80 387.21 449.48 384.65 423.70 382.90 407.24 380.55 391.31 378.86__ 383.85 500.00 491.62 480.51 453.73 42751 400.27 387.97 500,00 497.52 492.47 471.50 440.85 406.20 390.68 500.00 498.73 495.44 471.25 445.49 408.27 391.62 Case 2. Ax= 0.0005 m, At= 1s Node T 6 Location x,cm__0 05 Ti 10 16 a 2.0 500.00 500.00 390.60 469.88 387.23 449.50 384.65 423.70 382.90 407.26 38055 391.33 378.87 383.88 1 é 0 Os 500.00 491.67 480.55 453.76 42754 400.31 388.01 it 10 500.00 497.55 492.51 47154 440.89 406.24 390.73 16 15 500.00 498.75 495.48 471.29 445.53 408.31 391.67 2i 2.0 300.00 500.60 388.86 466.24 386.67 445.87 384.51 42239 382.83 406.68, 380.47 390,88 378.77 _ 383.55, 500.00 491.94 479.67 452.51 426.61 399.55 387.49 500.00 498.21 493.14 471.03 439.79 405.28 390.07 500.00 499.30 496.39 47747 444.40 407.28 390.97 Case 4: Ax=0,001 m, At=0.25 s ‘Node T 6 Location x,cm___0 Os i 10 16 2 t@)=0.0 500.00 500.00 1.0 388.28 463.70 20 386.44 443.89 50 384.44 421.78 10.0 382.80 406.40 200 380.43 390.66 30.0 _378.73__383.40 500.00 492.23 479.05 451.88 426.15 399.18 387.24 500.00 498.67 493.55 470.75 439.24 404.81 389.76 300.00 499.62 497.02 471.09 443.81 406.77 390.643-119 (continued) GD he = 200(T—Tays)!* Win? KR ‘Node i 6 ir 16 2 Locationx,cm___0 0s 10 15 20 t@)=0.0 500.00 — 500.00 500.00 500.00 500.00 1.0 38955 469.50 491.54 49750 498.72 2.0 386.12 448.99 48032 492.40 495.40 5.0 383.61 422.99 453.30 471.23 477.03 10.0 381.91 40643 426.87 44031 444.99 20.0 379.65 390.46 399.46 405.42 407.50 30.0 378.05 383.04 387.15 389.85 390.80 Comments: 1. The two different h, expressions give approximately the same temperature res- ponse, 2, Reducing 4x and At gives little change in the results, showing that the choice of ‘x= 0.001 m and At= I s is adequate for this problem.Exercise 3-120 Given: One-dimensional unsteady conduction. Required: Implicit form of the finite difference approximation for a convective boundary condition (Item 2, Table 3.8). Assumptions: 1. No intemal heat generation. 2. Constant properties. An energy balance applied to the control volume shown in the figure requites that during the time step At, Increase in intemal net conduction 4 convection across ‘energy within volume into volume boundary : ait ie po ea UCTS =Th) = k(-1) —$>——— At + he (1-11 Ta? ae Divide by pcAx/2, spt, <2 OAL erst iy 49 BA. en Te'-To age T TY) +2 = K ee Te") = 2Fo(T}*!—Tf'!) + 2BiFors!—TH") Solving for Tif", 2rocri* eBiTi)eT, 142Fo+2FOBi Comments: 1. Notice the way that the conduction into the volume is written: the term (Ty~To) ensures that the heat conduction is into the volume if T, > To, ie. the term is positive. ryExercise 3-121 Given: One-dimensional unsteady conduction. Required: Explicit and implicit forms of the finite difference formula at a boundary with known heat flux (Item 3, Table 3.8). Assumptions: 1. No intemal heat generation. 2, Constant properties. An energy balance on the control volume shown in the figure requires that during the time step At, Increase in intemal _ net conduction , _heat flow energy within volume = into volume * across boundary (@ Explicit. T pe (THT) =k) Este tas Divide by pcAx/2, iet_y GAL oni _ opi oat Ax THN? age OTD? age Solving for TH, iA Ty! <2Fo [ri + + (-2FoyT} Fora stable solution the coefficient of T}, must not be negative, that is, Fo < 1/2, as before. Gi) Implicit, $8 coyerptny enc De ary (1-1) S"=75) = k(1-1), ix Dividing by pcAx/2 and solving for Tit, Ate geet)Exercise 3-122 Given: Two-dimensional unsteady conduction. Required: Explicit and implicit forms of the finite difference approximation at a plane surface with a convection boundary condition (Item 2, Table 3.9). Assumptions: 1. No internal heat generation, 2. Constant properties. ‘An energy balance applied to the control volume shown in the figure requires that during the time step At, w ol Z Ho Increase in intemal _ net conduction, convection across a energy within volume ~ into volume boundary Hae @ ea TH 9% ax ery!) =xaxy) aae[at | BA * ala i723 TT pee nanettpat Dividing by po(Ax)?/2 hoAx 1s +2 eae S ‘THT = oF nae oT wo T}) + x oQw To) +2 we Th) Solving for Tif? Th! = Fo@T\+T)+Th+2BiT!)+(14Fo-2BiFo)Th, where Fo = oAt/(Ax)?, Bi=h,Ax/k. Fora stable solution the coefficient of Tj, must not be nega- tive, that is, Fo2+Bi) < 1/2 GH Implicit, r ister! Ax tani a egaery eae ar [48] ETE pe (Ax TET) = ke) I at tk] Et] ae ax ,) 8" hist +k 2 Ha a At+ h.Ax(Te TH )At Dividing by pe(Ax)"/2, and solving for Ti" FoRT i + ry 4Ti42BiT 4 Th 1 8 T2FOQsBi)Exercise 3-123 Given: ‘Two-dimensional unsteady conduction. Required: Explicit and implicit forms of the finite difference approximation at a plane surface with known heat flux (Item 3, Table 3.9). Assumptions: 1. No intemal heat generation 2. Constant properties. ‘An energy balance applied to the control volume shown. in the figure requires that during the time step At, Increase in 5 i = Ret conduction , heat flux ‘neal every = “into volume * across boundary @ Explicit. pe AE 9 (ax 1TH =79) = k(x) FB aver k a 4| a Ate gi@x-DAt Divicing by ee ns Th ie 7 i-th + oat GAx (xy k Solving for Ti", TS! = FoQT | +Th+Th+2q)Ax/k)+(1-4Fo)T}, where Fo = oAt/(Ax)°. Fora stable solution the coefficient T must not be negative, ie., Fo< 1/4. Gi) Implicit. ih ie ht) + ae aes (i-T) +2 rey a att eae pe ax ayer}! 0}y = k(Ax-1) 7 TH ae 1_yist Ax | TESTO seg gta off | De AEF ae) Ar Dividing by pc(Ax)?/2 and solving for TS oT it + Tit! 4 Tit 429i! Ax) +3 ‘te ae 14+4FoExercise 3-124 Given: Two-dimensional unsteady conduction. Required: Explicit and implicit forms of the finite difference approximation for exterior comer with a convection boundary condition (Item 4, Table 3.9). Assumptions: 1. No intemal heat generation. 2. Constant propertics. ‘An energy balance applied to the control volume shown in the figure requires that during the time step At, n Increase in fon _ convection a intemal energy = Met Conduction , across Te within volume ™#0 volume "Boundary Te (@ Explicit. 2 ; qT pe [+] aera =] a Ax). ThTo 5. +k 2 a) re ‘At+ 2h, —To)At Dividing by po(Ax?/4 ey AX At ' top) 22 AL cpi_piy 42 OAL gir 4 Beak oat TY'-Th (ax? (Ti-Th) + ane (TT) +4“ (an? Solving for Ti}? ‘Tit = 2FOCT| +Ti+2BiT})+(14Fo4FOBI)T) where Fo = cAt(Ax)’, Bi =h,Ax/k. Fora stable solution the coefficient of T}, must not be nega- tive, that is, Fo(1+Bi) < 1/4. Dividing by pe(Ax)?/2, and solving for Ts? appt = EO TE 2piTg aT} (ee 144F0+4FoBiExercise 3-125 Given: Two-dimensional unsteady conduction. Required: Explicit and implicit forms of the finite difference approximation for interior comer with a convection boundary condition (Item 5, Table 3.9). Assumptions: 1. No intemal heat generation, 2. Constant properties. ‘An energy balance applied to the control volume shown, in the figure requires that during the time step At, Increase in fon , convection internal energy = Metconduction 4 across within volume oundary @ Explicit. 2 expert) = ee oy a t+ k(x) + k(Ax\(1) = us 7 At+k |) ae Ph van [8] tbat = by tc yay = 2 ae t+ 4 ae athy+$ a iT} +2 GM cathe $M OM cts Solving for Tij!, Ti? = = ocr] +27}+27T)+T)+2BiT}) + (14 Fo~44/3) FOBT}, 3 where Fo = aAv(Ax)’, Bi=h,Ax/k. Fora stable solution the coeflicient of T} must not be nega- tive, that s, FOB) < >. Gi) Implicit. ri mH a pe 2 (anni Th) = (s Jou ) = 1/2(1+0.5684) = 0.3188 ‘Thus Fo = 0.1 will certainly satisfy the stability criterion (see Comment No. 1). The corresponding time step is At = FoAx*/a. = (0.1)(0.01)°/0.75x10% = 13.33 s.3-128 (continued) TE = 600°C = 2)(0.1)(873) + [1-(2K0.1) (873) = 873 K 873 K. Thus substituting in Eqs. (1), (2) and (4), Ty! = (0.1)(8734873) - [1-(2\(0.1)1(873) = 873 K, (0.1)(8734873) - [1-(2\(0.1)(873) = 873 K (0.1)(873+873) - [1-(2)(0-1)(873) = 873 K (2)(0.1)[873+(0.5684)(273)} + [1-(2)(0.1)-(2)(0.1)(0.5684)](873) = 804.79 K hl = 50+(5,67x10%)(0.91}27374804.79")(273+804.79) = 90.16 Wim? K Bi! = (90.16)(0.01)/1.75 = 0.5152 Ty? = 873 K T?=873K 12 =873K ‘Ty? = (0.1)(873+804.79) + [1-(2)(0.1)(873) = 866.18 K Ty? = (2)(0.1)[873+(0.5152)(273)] + (1-2)(0.1)-@2)(0.1)(0.5152)}804.79 = 763.63 K Comments: 1 Use of a larger time step, say corresponding to Fo = 0.3, in principle will give a stable solution. However, due to the rapid variation in Bi, there will be severe oscillations in the solution for the first few time steps. Notice how the choice of mesh size and Fourier number determines the apparent speed at which the temperature change propagates into the slab. In this case, the centerline temperature cannot change until 5x13.33 = 66.7 s have elapsed. The differential heat conduction equation is a parabolic one, implying an infinite speed of propagation of the temperature response, However, in some situations the speed of propagation is not infinite, and the parabolic form of the heat conduction equation does not apply. See, for example, Exercise 3-50.Exereise 3-129 Given: Slab of 8 cm thick soda-lime glass at 300°C cooled by air at 30°C with h, = 44 W/m? K. Required: Numerical solution for the temperature response using the explicit method on a 1 em mesh with Fo = 0.3—-results for five time steps. Assumptions: 1. Bdge effects are negligible and h, is constant over the slab surfaces. 2. kand a are constant. ‘Symmetry allows us to consider the half slab as shown. From Table A.3, soda-lime glass properties at 300 K are 1, K = 0.88 Win K, a = 0.44x10 mis. ‘The mesh Biot number is Ax _ 44y “k 08 Bi ‘The appropriate stability criterion is Eq. (3.94), ost _-_!__ wo33 WB) ” B05 Thus the given value of Fo = 0.3 satisfies the stability criterion. The corresponding time step is 2 2 at = Fodx? _ (03/0000 «6g 5 (4 min) a 0.44x10 © Equation (3.90) for the interior nodes T,, T,, T; is, TA" = [1-@Q)03)T, +03T py +Te.,) = O4T,,+0.3(T Ty) and Eq. (3.93) for the surface temperature Ty is Te" = (2)(0.3)(T,' +0.5)(30)] +11 -(2)(0.3) -2(0.3)(0.5) IT For the centerplane temperature T, we recognize that the centerplane is an adiabatic surface and use Bq. (3.93) with Bi 0, = (2)0.3)T; = 0.6(8;'+15)+0.1TE | +{1-@)0.3)IT,3-129 (continued) = 0.6T; +0.4T, ‘The initial condition is ‘T,? = 300°C; thus the temperatures at the first time step are 1.6)(300+15)4(0.1)(300) = 219°C .4)(300)+(0.3)(300+300) = 300°C. 1.4)(300)+{0.3)(300+300) = 300°C 1.4)(300)+(0.3)(300+300) = 300°C .6)(300)+(0.4)(300) = 300°C }.6)(300415)+(0.1)(219) = 210.9°C .4)(300)+(0.3)(210.9+300) = 273.27°C .4)(300)+(0.3)(3004300) = 300°C .4)(300)+{0.3)(300+300) = 300°C .6)(300)+(0.4(300) = 300°C .6)(273.27+15)+(0.1)(210.9) = 194.05°C .4)(273.27)-+(0.3)(210.9+300) = 262.58°C .4)(300)+(0.3)(273.274300) = 291.98°C .4)(300)+(0.3)(300+300) = 300°C .6)(300)+-(0.4)(300) = 300°C At the fourth time step, Tot = (0.6)(262.58+15)+(0.1)194.05) = 185.95°C (0.4Y(262,58)+(0.3)(194.054291.98) = 250.84°C .4)(291.98)4(0.3)(262.58+300 (0.4)(300)+(0.3)(291.98+300) fe! = (0.6)(300)+(0.4)(300) = 300°C At the fifth time step, ‘Ty; = (0.6(250.84415)+(0.1)(185.95) = 178.10°C (0.4)(250.84)4(0.3)(185.95+285.57) = 241.79°C (0.4)(285.57)+(0.3)(250.84+297.59) = 278.76°C. (0.4)(297.59)+(0.3)(285.57+300) = 294.71°C TF = (0.6)(297.59)4(0.4)(300) = 298.55°C Comments: 1. Notice that the stability criteria for a convective boundary condition is always ‘more restrictive than the criterion for interior nodes (Bi > 0). 2. Notice how the choice of mesh size and Fourier number determines the apparent speed at which temperature changes propagate into the slab. In this case, the centerfine temperature cannot change until (5)(68.2) = 341.02 s have elapsed. ‘What happens if a smaller mesh size is used?Exercise 3-130 Given: Slab of 8 cm thick soda-lime glass at 300°C cooled by air at 30°C with h, = 44 W/m? K. Required: Numerical solution for the temperature response using the explicit method for 5, 10, 20 and 40 nodes (at 20 min = 1200s). Assumptions: 1, Bdge effects are negligible and h, is constant over the slab surfaces. 2. kand x are constant eee errr eed 4 dus ‘Symmetry allows us to consider the half slab as shown. 1 From Table A.3, soda-lime glass properties at 300 K are 1 k = 0.88 Wim K, 0 = 0.44x10% m4s. Equation (3.90) sen I 2 for the interior nodes Ty, TyynTyg i i a eu i 1 Ty = Fo(Ta.1*Ta) +(1-2FOT i and Eq. (3.93) for the surface temperature is 2 Son Tj’) = 2Fo(T;' +BiT,) + (1-2Fo-2FoBA)T, ewe For the centerplane temperature Ty, we recognize that the centerplane is an adiabatic surface and put Bi = 0 in Eq. 3.93), = 2FOT}) 4 + (1-2Fo)T yy The mesh Biot number is Bi = is Bq. (3.94), b,Awk = (44)Ax/0.88 = SOx. The appropriate stability criterion ea 1 * 20B) —-2+50Ax) ‘The corresponding time step is = FoAx? __ FoAx? At = Foi a 0.44x10-* = 2.273x10°FoAx? ‘The initial temperature is Ty, = 300°C. COND? calculation: Fo=at/L? =(0.44x10~)20)(60)/(0.04)* Bi=h, L/k = (44)(0.04)/(0.88) = 2.00 T,=133.58'C Ty =246.21°C3-130 (continued) With 40 nodes, Ax = 0.04/39 = 1.0256 10°m For stability Fo < 1/2(1+(50)1.025%10"))= 0.4756 Hence choose the following time-steps and 0 Nodes 5 10 20 40 Mesh size mm. 10 4.4444 2.1052 1.0256 at s 100 20 4 1 Fo 0.44000 0.44545, 0.39701 0.41826 Tw °c 246.90 246.09 246.19 246.20 Error “cl 0.69 0.12 0.02 0.01 To c 137.14 133.26 133.51 133.57 Error °c 3.56 0.32 -0.07 -0.01 Comments: 1. Twenty nodes gives an error of less than 0.1°C, which is surely adequate.Exercise 3-131 Given: Large slab of 3 cm thick Teflon initially at 20°C, g,,4 = 3500 W/m? on one side, convection cooling to air at 10°C on the other. Required: Numerical solution procedure for temperatures using the explicit method on a 1 cm mesh with Fo = 0.2. Assumptions: 1. Bdge effects are negligible. 2. Radiation heat transfer from both surfaces to surroundings is negligible. 3. Constant k and a T,= 10 35. WimeK From Table A.2, Teflon properties at 300 K are k = Biot number is 35 Wim K, a = 0.15x10% m’/s. The mesh h, Bi = 2A® ~ 5X01) «1.09 K 035 ‘The appropriate stability criterion is Eq. (3.94), ipo 2B) 2-1 ‘The stability criterion for the interior nodes and the radiation heated surface node is Fo < 0.5 (Table 3.8). ‘Thus the given value of Fo = 0.20 is satisfactory. The corresponding time step is = Fo(axy? _ (0.200.017 At a 0.15x10€ = 1333 Table 3.8, item 3, gives the finite difference approximation for T, as Te" = (2)(0.2)IT;' + 3500}(0.01)/0.35] +1 -(2)(0.2)]T3-131 (continued) = (0.4)(T,' +100) +0.6T, Item 1 gives for the interior nodes T, and T,, Ti! = 0.2)(T)+T,) +067, = .2MT; +T;)+0.6T, Item 2 gives for the surface node T,, Ty” = 2YO.2)ITy + (1)AO)] + - 200.2) -2N0.2)(1 OTs = (0.4)(T;+10)+0.2T;, ‘The initial condition is T,? = 20°C. The temperatures after the first time step are ‘After the second time sep: Te = (0.4)(20+100)+(0.6)(60) = 84.0°C. Ty? = (0.2)(60420)+(0.6)(20) = 28.0°C 19.2" 15.2°C (0.2)(20+16)+(0.6)(20) (0.4)(20410)+(0.2)(16) ‘After the third time step: Ty? = (0.4)(28+100)+(0.6)(84) = 101.60°C 0.2)(84419.2)+(0.6)(28) (0.2)(28+15.2)+(0.6)(19.2) (0.4)(19.2+10)+(0.2)(15.2) (0.4)(37.44+100)+(0.6)(101.60) = 115.94°C (0.2)(101.60+20,16)+(0.6)(37.44) = 46.82°C (0.2)(37.44+14.72}+(0.6)(20.16) = 22.53°C (0.4)(20.16+10)+(0.2)(14.72) = 15.01°C Comments: 1. This calculation procedure is easily implemented on a spreadsheet, or by a simple computer program. 2. Notice that the stability criterion for a convective boundary condition is always ‘more restrictive than the criterion for interior nodes (Bi > 0).Exereise 3-132 Given: A 3 cm thick AISI 302 black oxidized stainless steel plate suddenly exposed to a solar radiation flux of 900 W/m? K. Required: (i) Explicit finite difference approximation for the surface node, (ii) Temperatures after first four time steps on a 1 cm mesh, Assumptions: 1. Bdge effects are negligible and surface conditions are uniform to give one- dimensional conduction, 2, Reradiation to the surroundings can be approximately treated by using a radiation heat transfer coefficient (@ We parallel the derivation of Eq. (3.93). The energy balance on the surface node now requires that during the time step At, Increase in internal energy = net conduction into volume + solar radiation absorbed + convection and radiation from surroundings. x=0 Convection and radiation from the surroundings can be accounted for by a combined heat transfer coefficient h = h,+h,, where b, = 402T,°, because both the air and surroundings are at the same temperature. ‘Then Titi pe SKaa.nfty-Ty) = kC.1), At +Guaye(1-D)At +h D(TE Ty) at ‘Rearranging and solving for Ty", GaotaAX ‘+BiT, + K } (1-2Fo -2FoBi)T,. a) where Bi = hAx/k = (h,+h,)Ax/k. For stability, the coefficient of Ty' must not be negative, Fos soapy nme at Ba. 3.98). (ii) Taking T,, = 300 K, h, « 6e, = (6)(0.75) = 4 W/m? K, and h = h.+h, = 6+4 = 10 Wim? K. From Table A.1, k= 15 W/m? K for 302 stainless steel. Hence Bi = hawk = (10\(0.01Y(15) = 0.00667 For stability, Fo < 1/2(1+0.00667) = 0.497 > 0.4, as required, Substituting in Eq. (1)3-132 (continued) Ty" = 2)(04)IT}' + (0.0067)(20) +(900)(0.89)(0.01)/15} + [1-2)(0.4)-(2)(0.4)(0.00667)]T, Ty” = O.8(T;' +0.668]+0.195T, @ Using Table 3.8, item 1 for the interior nodes and item 4 for the adiabatic node, Ti? = o4ctj+T,) +027; @ 1 = 04)(T+T;) +027, @ 1" = og, +027; © With T,° = 20°C for i = 0,1,2,3, the temperatures after each time step are: To} = 0.8)20+0.668)+(0. 195120) = 2043°C 1.20% (0.4)(20+20)+(0.2)(20) = (0.4)(20+20)+{0.2)(20) = oe (0.8)(20)+(0.2)(20) = 20°C Seas] Saas = 900 Wt (0.8)(20+0.668)+(0.195)(20.43) = 20, (0.4)(20.43420)+(0.2)(20) = 20.17°C (0.4)(20+20)+(0.2)(20) = 20°C T3? = 0.8\(20}+(0.2)(20) = 20°C (0.8)(20.1740.668)+(0.195)(20.52) = 20.67°C (0.4)(20.52+20)+0.2)(20.1 (0.4)(20.17420)+(0.2)(20) = 20.07°C 4 (0.8)(20)+(0.2)(20) = 20°C (0.8)(20.2440.668)+(0.195)(20.67) = 20.76°C (0.4)(20.67420.07)0.2)20.24) = 20.34°C (0.4)(20.24420)4(0.2)(20.07) (0.8)(20.07)+(0.2)(20) = 20.06°C ‘Aaicbatic T i i t = 20.28 ten! | ¢ { t 20.11°C ih ‘The physical time step corresponding to 2 mesh Fourier number of 0.4 is At = FoAxfct = (0,4)(0.01)73.88x10 = 10.31 s (with 0: from Table Ala). Notice that the temperature at node 3 does not change until 4x10.31 = 41.2 s have elapsed. This finite speed of propagation of the temperature response is an artifact of the finite difference method: clearly the time increates with increasing number of node points and increasing Fourier number. In fact, the original heat conduction equation is a parabolic one, implying an infinite speed of propagation of the temperature response. However, in some situations the speed of propagation may not be infinite and the Fourier’s heat conduction equation is not valid. See, for example, Exercise 3-50.Exercise 3-133 Given: An electrically heated type 316 stainless steel bar, 4 mm thick, cooled by impinging water on each side. Required: (i) Explicit finite difference approximations for surface and interior nodes. i) Solation for temperature after first four time steps using Fo = 0.3, Assumptions: 1. Conditions do not vary along or across the bar giving a one-dimensional problem. 2. Constant thermal properties and PR heating. ‘T,2 2080, n= 6500 Win @ _ Equation (3.89) becomes AU = Qt + Q, At and adds the term Q,” (1.1)AxAt to the energy balance used to derive Eq. (3.90), which for node 1 becomes Ti" = Fo(Ty'+T,) +(1-2Fo)T,' +FoQ."Ax 7k () In modifying Bq, (3.93) the term Q)” (1.1)(Aw/2)At is added to the energy balance to obtain for node 0 To” = 2Fo(T,'+BiT,') +(1-2Fo-2FoBi)T,' +FoQ, 7k @ Since node 2 is at a plane of symmetry we can set Bi = 0 in Eq. (2), Tj"! = 2FoT;'+(1-2Fo)T; +FoQ.” Ax rk @ (ii) From Table A.1a for 316 stainless steel at 300 K, k = 13 W/m K, @ = 3.37x10% m/s. The mesh Biot number is Bi = h,Awk = (6500)(0.001)/(13) = 0.5 ‘The most severe stability criterion is associated with the convective boundary condition, Eq. (2), 1 1 oe Ta = 0333 }.3, as required. °< aca WO eee ‘The nodal equations become To" = (2)(0.3)[Ty+(0.5)(20)] + [1-(2)(0.3)-(2)(0.3)(0.5)ITy = (0.3)(8.67x107}(0.001)°(13) Ty" = 0.6\(T;'+10) +0.17, +2.00 @3-133 (continued) Ty" = (0.3)(Ty+T,) +047; (5) Ty" = 0.67/'+0.4T;+2.00 © With T,° = 20°C, the temperatures after the first four time steps are (0.6)(20+10)+(0.1)(20)+2.00 = 22.00°C 1, = (0.3)(20+20)+(0.4)(20)+2.00 = 22.00°C Ty! = (0.6)(20}+(0.4)(20)+2.00 = 22.00°C ‘To? = (0.6)(22.00+10)+(0.1)(22.00)+2.00 = 23.40°C T)? = (0.322.00+22,00)+(0.4)(22.00)+2.00 = 24.00°C ‘Ty? = (0.6)(22.00)+(0.4)(22.00)+2.00 = 24.00°C Ty} = (0.6)(24.00+10)+(0.1(23.40)+2.00 = 24.74°C T3 = 0.3)(23.40+24.00)+(0.4)(24.00)+2.00 = 25.82°C Ty? = (0.6)(24.00)+(0.4)(24,00}+2.00 = 26.00°C ‘Tot = (0.6)(25.82410)+(0.1)(24.74)4+2.00 = 25.97°C T,* = (0.3X24.74+26.00)+(0.4\(25.82}42.00 = 27.55°C ‘Ty? = (0.6)(25.82)4(0.4)(26.00)+2.00 = 27.89°C Comments: 1. The physical time step corresponding to a mesh Fourier number of 0.3 is At = FoAx’/ot = (0.3)(0.001)7/3.37x10% = 0.0890 s. 2. This calculation procedure is easily implemented in a spreadsheet or by writing ‘a computer program, 3. Notice that the effect of the convective boundary condit mn reaches the center node only after four time steps (0.356 s). The choice of mesh size and Fourier ‘number determine the apparent speed at which thermal effects are propagated into the bar. In physical reality the heat conduction equation is parabolic, which implies an infinite speed of propagation.Exercise 3-134 Given: Slab of 4 om thick glass at 400°C to be cooled by air at 20°C with h, = 30 Wim? K. Required: Numerical procedure for temperature response using the explicit method oa a 1 em mesh—results for five time steps. Assumptions: 1, Edge effects are negligible, and h, is constant over the slab surface. 2, Kk and o: are constant Symmetry allows us to consider the half-slab as shown. ‘The mesh Biot number is Bi = h.Awk = (30)(0.01)/1.5 = 02 ‘The appropriate stability criterion is Eq. (3.94), Fo < 1/2(14Bi) = 1/2(140.2) = 0.417 Thus the given value of Fo = 0.4 satisfies the stability criterion. ‘The corresponding time step is At = FoAx/or = (0.4)(0.01)7/0.8x10 © = 50s (~1 min) Equation (3.93) for the surface node Ty is Ty" = 2Fo(T,'+BiT) +(1-2Fo-2FoBDTy = (2)(OAN(T,'+(0.2)(20)) +1 -2)(0.4)-2)(0.4V(0.2))T = 0.8(T,'+4)+0.04Ty w Equation (3.90) for the interior node T, is TH = Fo(Tg eT) +(1-2Fo)T} = .4)(Ty+T2) +(1-(2)0.4)) Ty = (0.47) 4T,) +027} @ ‘And for the centerplane node T, we recognize that the centerplane is an adiabatic surface and use Eq. (3.93) with Bi = 0, Ty" = 2FoT|'+(1-2Fo)Ty = VOT; +(2-(0.2(0.4))T;3-134 (continued) = 08 +0.2 @ ‘The initial condition is T,,° = 400°C; thus the temperatures at the first four time steps are Ty} = (0.840044) +(0.04)(400) = 339.20°C (0.4)(400+-400)+(0.2)(400) = 400°C uM = (0.8)(40044)-+(0.04)(339.20) = 336.77°C (0.4)(339.20+400)+(0.2)(400) (0.8)(400)4(0.2(400) = 400°C (04)(336.77+400)+(0.2\375.68) = 369.84°C (0.8)(375.68)4+(0.2)(400) = 380.54°C 11.-76°C (0.4)(317.214+380.54)+(0.21369.84) = 353.07°C (0.8)(369.84)+(0.2)(380.54) = 371.98°C To§ = (08)(353.0744)+(0.04)(311.76) = 298.13°C (0.4)(31 1.76+371.98)+(0.2)(353.07) = 344.11°C. (0.8)(353.07)+{0.2)(371.98) = 356.85°C Comments: 1. Notice that the stability criterion for a convective boundary condition is always ‘more restrictive than the criterion for intetior nodes (Bi > 0). 2. Notice how the choice of mesh size and Fourier number determines the apparent speed at which temperature changes propagate into the slab. In this case, the centerline temperature cannot change until (3)(50) = 150 s has elapsed. What happens if a smaller mesh size is used?Exercise 3-135 Given: Slab of 4 cm thick glass at 400°C to be cooled by air at 20°C with h, = 30 W/m? K. Required: Computer program to obtain temperature response using the explicit methéd. Bffect ‘of mesh size on temperature profile at t= 1000 s. Assumptions: 1, age effects are negligible and h, is constant over the slab surfaces. 2. K and ot are constant. ‘Symmetry allows us to consider the half-slab as shown. ‘Equation (3.90) for the interior nodes is Tye = FOT Ty) “(1-2 O)T a Equation (3.93) for the surface node is Ty”) = 2Fo(T, | 0" i Bi = h,Awk = G0)(0.01V/1.5 = 0.2 Using Table 3.8, the finite difference formulas are = i = (0.8/1.96)(Ty"'+4) +(1/1.96)Ty @ poten her , Tit = meeliaale “YS oa/ayte ets) +a yr; @ 2FoT {141 ia i = Se + OSL OT MAT} 8 The initial condition is T,,° = 400°C. Using Gauss-Seidel iteration, Ty! = 0.8/1.96)(400+4)+(1/1.96)(400) = 368.98°C (0.4/1.8)(368.98+400)+(1/1.8)(400) = 393.11°C Ty! = (0.8/1.8)(93.11)+1/1.8)(400) = 396.94°C Ty! = (0.8/1,96393.11+4)#(1/1.96)(400) = 366.17°C X 1 = (0.4/1.8)(366.17+396.94)+(1/1.8X(400) = 391.80°C = (0.8/1.8)(391.80)+(1/1.8)(400) = 396.36°C (0.8/1,96)(391.80+4)+(1/1.96)(400) = 365.63°C (O.A/1.8)(365.63+396.36)+(1/1.8)(400) = 391.55°C (0.8/1.8)(391.55)+(1/1.8)(400) = 396.24°C (0.8/1.96)(391.55+4)+(1/1.96)(400) = 365.53°C (0.4/1.8)(365.53+396.24)+(1/1.8)(400) = 391.50°C (0.8/1.8)(391.50)+(1/1.8)(400) = 396.22°C To! = (0.8/1.96)(391,5044)4(1/.96)(400) = 365.51°C Ty = (ON 8)(365.514396.22)4(1/1.8)(400) = 391 49°C T,} = @.8/1.8)(391.49)+(1/1.8)(400) = 396.22°C Comments: 1. We see that the temperatures have nearly converged after only five iterations. 2. Notice that we are iterating on T,,'*!; T,,' remain constant (in this case at 400°C). 3. Notice how latest available values are used in Ganss-Scidel iteration. 4. We did not bother to use an index to denote the iteration level since we are already using two indices~m and i.Exercise 3-137 Given: Slab of 4 cm thick glass at 400°C to be cooled by air at 20°C with h, = 30 W/m? K. Required: Numerical solution for the temperature response using the implicit miethod to investigate the effect of size of mesh and time step. Assumptions: 1. Bdge effects are negligible and h, 2. Constant k and ct, is constant over the slab surfaces. T= 20°C h, ="30 Wimek ‘Symmetry allows us to consider the half slab as shown. Using Table 3.8 the required finite difference: approximations are __ 2Fo(t;" Bits") +7} 1 1-2F0-2FoBi a Fo(Tcs*Tnet) Ten {ie a eer teeaeHSteet = 2,3,..,.M-1 (2) T2Fo : ms Q 2FoT ye +Ty eae @ T+2Fo ‘As a nominal case, choose M = 21, Ax = 0,001 m, and At = 10s. Then Bi = h,Axhk = (30)(0.001)/1.5 = 0.02; T, = 20°C. Fo = aAUAx? = (0.8x10°(10)(0.001)? = 8.0 ‘The initial condition is T,,° = 400°C. COND? calculations: Fo = ot/L.? = (0.8x10}(1000)(0.02)? = 2.0 Bi, = hL/k = (30)0.02)/1.5 = 0.4 (n1=0) = 0.52338; Ty, = 218.88°C3-137 (continued) en: 43395; Ty, = 184.90°C For the nominal case (M = 21, Ax = 1 mm, At = 10 s) the following temperature distribution is obtained with a convergence criterion of 10 K. xmm 0 5 10 15 20 T.°C 18531 19997 210.68 217.15 219.38 ‘The effects of mesh size and time step are shown below. Time step Nodes T, — Emor, Tyg Exror s "c C eeeeeeen °c 30 6 22138 2.50 18691 2.01 50 1 2213324518691 2.01 50 2122132244 18692 2.02 20 1 21988 1.00 185.71 O81 10 1.21939 OSI 185.31 O4T 10 2 =: 2193805018531 0.41 1 31 -218.94 0.06 184.95 0.05 a 41 21894 0.06 184.95 0.05 Comments: 1. Notice that the error depends primarily on size of time step, not number of nodes. ‘An optimal calculation will have a small time step, but not an excessive number of nodes. 2. Notice that absolute errors are given in the table. Do not calculate % errors since these depend on whether degrees Celsius or Kelvins are used. In fact, 9% errors are always meaningless for temperatures (but not for temperature differences).Exercise 3-138 Given: Slab of 4 cm thick glass at 400°C to be cooled by air at 20°C with h, = 30 W/m? K. Required: Numerical procedure for temperature response using the implicit method oh a 1. cm mesh: solution for first time step using a direct method. Assumptions: 1. Bage effects are negligible and h, is constant over the slab surface. 2. k and o& are constant. Symmetry allow us to consider the half slab as shown, ‘The mesh Biot number is Bi = hAxlk = (30)(0.01V1.5 = 02 Using Table 3.8, the finite difference approximations are | figs ou eerie n= do wimex Ty pint, FOC BITE Ty N04 NT}""+0.2)20))+Ty : ene ee sane a int T+2Fo+2FOBi TOA QO: qt. Fog Te "Ty! _ O4)(Ty" «Ty )Ty : 12Fo 1+Q)0.4) 2FoT "+7, (2)(04)T]*4T . 12Fo ~ 142,04) Rearranging in the standard form for a set of linear equations: T)*-0.4082T, = 1.632740.5102Ty 0.22227)" +7,'-0.2222T,*' = 0.55567; 0.44447! Ty! = 0.59967; or in matrix form Ts 1,6327+0.5102T,. 1 0.4082 0.2222 1 0.2222) |r"! = 0.sss6T;' 04d 1 rst 0.5556T At time t= 0, Ta? = 400°C; thus, the solution for the first time step is obtained from3-138 (continued) 1 ~0.4082 To} foos.71 2222 1 0.2222] |i] = |222.24 “0dddg 1 292.24 TA}(T) = BB] The solution is (T') = (Ay B) where the inverse matrix [AJ' can be obtained using a standard computer subroutine, Altematively, hand-held calculators have standard routines for solving simultaneous linear equations for a limited number of unknowns (e.g. a Texas Instruments TI-68 can handle up to 5 unknowns). Using such a calculator, ‘To! = 365.52°C; T,! = 391.50°C; T,' = 396,22°C, Comments: 1. Direct methods are simple to use when the number of equations (nodes) are few. However, for the finite difference approximation to accurately represent the differential equation, many nodes must be used. Then direct methods may have {>} large round-off errors, and iterative methods are preferredExercise 3-139 Given: A large slab of 3 om thick Teflon, q,,y = 3500 W/m? on one side, insulated on the other. Required: Numerical solution procedure for temperatures using the implicit method oa 1 em mesh. Assumptions: 1. Bdge effects are negligible. 2. Reradiation and convection from the exposed side are negligible. 3. Constant k and 0. From Table A.2, Teflon properties at 300 K are k = 0.35 e 6 i — ‘Wim K, o = 0.15x10° mi/s, Using Table 3.8 the he required finite difference approximations are 2Fo(T;"'+q,Axlk) +Ty a 1+2Fo poten yen! get = Foe et3 et a 1+2Fo a _ FO} "4Ty")4Ty Te 3) a TsiFo @) 2FoT;"'sTy pit _ Pots '+Ts . es 12Fo ‘The given value of Fourier number is 0.4, thus jt _ 2)(A)[T}"" +(3500)(0.01)/0.35]+Ty, je i = (0.8/1.8)(T," +100) +(1/1.8)T, (5) é TOO OMAHA HOOKUTe THT = O4/1.8)(Tg +7) +1 8)T, © Ti = OA. 8yTf eT) +0/1.8)T! o ia int iat Ty" = O8/1.8)T} Ng); ® ‘The initial condition is T,,° = 20°C. Using Gauss-Seidel iteration,3-139 (continued) Ty! = (0.4/1.8)(29.88420)+(1/1.8)(20) = 22.19°C (0.8/1.8)(22.19)4(1/1.8)(20) = 20.97°C (0.8/1.8)(29.88+100)+(1/1 8)(20) = 68.84°C (0.4/1.8)(68.84+22,19)4(1/1.8)(20) = 31.34°C (0.4/1.8)(31 34+20.97)+(1/1.8)(20) = 22.74°C (0.8/1.8)(22.74)+(1/1.8)(20) = 21.22°C (0-8/1.8)(31.34+100)+(1/1.8)(20) = 69.48°C (0.4/1.8)(69.48+22.74)+(1/1.8)(20) = 31.61°C (0.4/1.8)(31.61421.22)+(1/1.8)(20) = 22.85°C (0.8/1.8)(22.85)+(1/1.8)(20) = 21.27°C (0.8/1.8(31.61+100)+(1/1.8)(20) = 69.60°C (0.4/1.8)(69.60+22.85)+(1/1.8)(20) = 31.66°C (0.4/1.8\(31.66421.27}+(1/1.8(20) = 22.87°C (0.8/1.8(22.87}+(1/1.8)(20) = 21.27°C (0.8/1.8)(31.66+100)+(1/1.8)(20) = 69.63°C (0.4/1 8)(69.63422.87)+(1/1.8)(20) = 31.66°C (0.4/1.8)(31.66421.27)+(1/1.8)(20) = 22.87°C (0.8/1.8)(22.87}+(1/1.8)(20) = 21.28°C and the iteration process has nearly converged. To obtain a direct solution write Eqs. (5) through (8) in the usual form of a set of simultaneous linear equations, With T° = 20°C, Ate 0.44447, = 55.556 -0.22229T5! 47, -0.222207,! = Wut -0.22222T;' 47! 0.222227, = 11.111 ~0.44444T,! 47; = a ‘Solving using a standard routine on a hand-held calculator, Ty! = 69.63°C; T,' = 36.67°C; T,' = 22.88°C; T,! = 21.28°C. Comments: 1. The Gauss-Seidel iteration had indeed converged to 0.01°C. 2. Notice how T,,' is not updated in the Gauss-Seidel iteration process. 3. We did not bother to use am index to denote iteration level-there are two indices, already.Exercise 3-140 Given: A large slab of 3 cm thick Teflon, gas = 3500 W/m? on one side, insulated on the other. Required: Numerical solution for temperature response using the implicit method to investigate the effect of size of mesh and time step, Assumptions: 1. Edge effects are negligible. 2. Reradiation and convection from the exposed side is negligible. 3. Constant k and to--2 ee 3500 wim? Fr TT From Table A.2, Teflon properties at 300 K are k = 0.35 Wim K, @ = 0.15x10° m/s, Using Table 3.8 the required finite difference approximations are iat te eee Re +qAwk)+Ty qi = Fea Ty a 1+2Fo Fo(Tes Tne *T, . =2,3,....M-1 (2) = m=2,3,..M ® Ty 8 As a nominal case choose M = 16, Ax = 0.2 cm and At= 2s, Then g,An/k = (3500)(0.0020.35 = 20°C Fo = aAVAx = (0.15x10%\2)1(0.002)" = 0.0753-140 (continued) MATLAB was used to solve the difference equations. The table below shows the effect of size of mesh and time step on the surface temperature at t = 100 s. ‘The convergence criterion was 0.001 K. At, s Ax, m 0.005 0.0025 0.001 Comments: 1. 2. Ty °C 5807 5807 5803.58.07 6227 6237 6243 «62.47 6287 62.72 6291 63.15 ‘A mesh size of 0.005 m (7 nodes) is too large to give a satisfactory result Consult a text on numerical methods for methods used to determine how a finite difference solution converges to the exact solution as time step and mesh size are decreased, Note that the above solutions have converged to within 0.001 K of the solution of the difference equations, not the differential heat conduction equation, Rev-9Exercise 3-141 Given: ‘A 3 cm thick AISI 302 black oxidized stainless steel plate suddenly exposed to solar radiation. Required: (i) Implicit finite difference approximation for the surface node. (ii) Temperatures after three Gauss-Seidel iterations at frst time step, and check using + adirect method. Assumptions: 1. Bdge effects ate negligible and surface conditions are uniform to give one- dimens 1 conduction. 2. Reradiation to the surroundings can be approximately treated by using a radiation heat transfer coefficient. (@® We parallel the derivation of Eq. (3.93). The energy balance on the surface node is Increase in _ net conduction | solar radia- | convection and radiation intemal energy ~ into volume ~ tion absorbed ~ . from surroundings Convection and radiation from the surroundings can be accounted for by a combined heat transfer coefficient h = hth, where h, = 4ceT,,°, because both the air and surroundings are at the same temperature. Then iet_pi ATy) = kL Ate y(1)At pe ax Toe a pe Kc DC, +h(LI(T,” +Ty” At where the conductive and convective fluxes have been evaluated at time step i+1 (instead of i as for the explicit method). Rearranging and solving for Ty!*! with Bi = hAx/k, pitt < 2FOT) BITS saigAW/k) “To, ‘ 1+2F0*2Fo i) Taking T,, = 300 K, h, ~ 6€; = (6)(0.75) = 4 Wim? K (assumed constant), and h = 6+4 = 10 Wim? K. From Table A.1, k = 15 W/m? K for 302 stainless steel. Hence Bi = hAwhk = (10)(0.01)/(15) = 0.006667 ‘Substituting in Eq. (1) with Fo = 0.4, T, = 20°C and constant, i (2)(0.4)[T;"' +(0.006667)(20) +(900)(0.89)(0.01)/15]+T 1+(2)(0.4) +(2)(0.4)(0.006667) int To3-141 (continued) Ty = 0.4431(T;"'+0.6673] +0.55391T. Using Table 3.8, item 1 for the interior nodes and item 4 for the adiabatic node, ia, 0.55556[0.4(1""4T;")+T,] Ty 7; = 0.55556(0.4(T;"+T;")+T,] Ty! = 0.44444T}""+0.55556° Som] [ae | r= 900 Wie 17 1 i I th tom | zip hh, | 20.06°C : = 20.02°C % 10.01°C ‘Adiabatic For direct solution we rearrange the equations in the usual form for simultaneous linear equations, Ty! - 0.44313T,! = 1374 = 0.22222T,' +7, -0.22222T,! = Wil = 0.222227; +T;! = 0.22229T,' = 11.111 - 0.44444T,' 47! = 1d Solving using a standard routine on a hand-held calculator, T,' = 20.264°C, T,' = 20,062°C, T;! = 20.015°C, T,! = 20.006°C. Comments: 1. We see that the Gauss-Seidel iteration has essentially converged to 0.01°C in only three iterations. With a computer a tighter convergence criterion can be used. 2. The physical time step comesponding to a mesh Fourier number of Fo = 0.4 is At = Fox = (0.4)(0.01)'/3.88x10* = 10.31 s (with « from Table A.1a), 3. Notice that we are iterating on T,,,"!; T,,' remains constant (in this case 20°C). Also notice how latest available values are used in Gauss-Seidel iteration. 4. We did not bother to use an index to denote the iteration level since we are already using two indices--m and i.Exercise 3-142 Given: A 3 cm thick AIST 302 black oxidized stainless steel plate suddenly exposed to solar radiation. Required: A computer program to solve for the temperature response using the explicit method, and effect of length of time step and mesh size on the solution. Assumptions: 1. One-dimensional conduction, 2. Constant thermal properties. Equation (3.90) for the interior nodes is THY = Fo(T yg y+Tyoy) +(1-2F oT, ay For the irradiated surface node we perform an energy balance on a control volume Aw/2 thick, \ Ax nerit-1) Te aay + a Preece tates denser Aca wh 7 + GuigAt(1.1) +h ((P,-1,.41(1.1) te where h = h.eb,! = h,seo(T2+T,7)C7,+Ty), and ajar = (0.89900) = 801 Wim?, Thus Ty! = 2FolT y+, AWK) +Bi *T,] +[1-2Fo-2FoBi ‘Ty @ Table 3.8, item 4, is used for the adiabatic surface node, Tj”? = 2FoT,'+(1-Fo)Ty @) ‘The stability criterion is Fo $ 1/2(1+Bi') where Bi’ varies with time due to the variation of h,. ‘Thus, in general, At = FoAx"/o: may have to be reduced as Ty increases with time. However, for this problem, a trial calculation shows that Ty(1000 s) ~ 27°C so that h, varies very little and Bi even less: thus a constant value somewhat less than the allowable value is quite adequate. From Table A.la at 300 K, 0: = 3,88x10° m/s, k = 15.0 Wim K; € = 0.75 is given. The temperature distribution at ¢ = 1000 s using Ax = 0.03 cm and At = 0.0025 s is xem 00 05 10 15 20 25 30 T.°C 26.32 26.34 26.40 26.50 26.64 26.82 27.05 ‘The effect of Ax and At on Ty at t = 1000 s is shown in the table.3-142 (continued) Axom dys Ty °C 1 15 27.024 1 2 27.038 1 1 27.040 05 2 27.047 0.03 © 0.0025 27.049 ‘The graph below shows the temperature response when the stability criterion is violated: Ax = Lom, At = 13.16 5, Atga, = 12.8 s. Responses are shown for nodes N and N-2. 0 100-200 004008006700 BK 8091000 ‘Time [6] Comments: 1, The solution is acceptable for Ax = .5 em (7 nodes). 2. Reducing the time step below the maximum allowable gives a small increase in accuracy. 3. Notice that the treatment of radiation is exact. In the explicit method, h,’ is caleulated exactly.Exercise 3.143 Given: An electrically heated type 316 stainless steel bar, 4 mm thick, cooled by impinging water on each side, Required: (i) Implicit finite difference approximations for the surface and interior nodes. ii) Temperatures after five Gauss-Seidel iterations at first time step, and check using a direct method. Assumptions: 1, Surface conditions do not vary along or across the bar. 2. Constant thermal properties and PR heating. (@ Equation (3.89) becomes AU = Qat + Q,At and adds a term Q,” (1.I)AxAt to the energy balances used to derive Eq. (3.95), which for node 1 becomes Fo(Ty1#T")+FoQ, Ax k+T Tro T w In deriving the approximation for a surface node with convection, the term Q.” (1.1)(Ax/2)At is added to the energy balance; modifying item 2 of Table 3.8 for node 0, in _ 2POCT}"+BiTS)+FOQ)” Ax 2k +Ty gee cen @ T-2Fo-2FoBi Since node 2 is a plane of symmetry (adiabatic) we can set Bi = 0 in Eq. (2) to obtain 2FOT}"'+FoQ,” Ax /k+T 1+2Fo Tr 8 (ii) From Table A.1a for 316 stainless steel at 300 K, k = 13 Wim K, «= 3.37x10% m’vs. The mesh Biot number is Bi = hAw/k = (6500(0.001)/(13) = 0.5 Also, Q”Ax hk = (8.67x107\(0.001)7/(13) = 6.669 ‘The nodal equations with Fo = 0.3 and TT,,° = 20°C become. 1. eosyr;" +(0.5)(20)}-+(0.3)(6.669)+T T2VO3)*2HOINOS) ia C1A.9)0.6(T; +10)+2+Ty)3-143 (continued) THT = (A.6 10.374" +T$)+2+T/4 Ti = (A.610.6)T}"424T,) Using Gauss-Seidel iteration for the first time step, (I/1.9){(0.6)(20410)+2+20) = 21.05°C (1/.6){(0.3)21.05+20)+2+20) = 21.45°C ™ = (1/1 .6[(0.6)(21.45}42+20] = 21.79°C (A/1.9){(0.6)(21.45+10}+22.00] = 21.51°C (/1.6)[(0.3)(21.51421.79}+22.00] = 21.87°C = (/1.6)1(0.6)(21.87)422.00] = 21.95°C Tol = (1/1.9){(0.6)(21.87+10)+22.00} = 21.64°C. Ty! = (1/1.6)[(0.3)(21.64+21.95)+22.00] = 21.92°C = (1/1.6)[(0.6)(21.92)+22.00) = 21.97°C ‘Ty! = (1/1 9){(0.6)21.92+10)+22.00] = 21.66°C Ty! = 11.6){(0.3)(21.66+21.97}+22.00] = 21.93°C Ty! = (1/1 6){(0.6\(21.93)+22.00] = 21.97°C To! = (1/1.9)[(0.6)(21.93+10)+22.00] = 21.66°C T,! = (1/1.6)[(0.3)(21.66+21.97)+22.00} = 21.93°C = (1/1.6){(0.6)(21.93)+22.00] = 21.97°C For a direct solution we arrange the equations in the usual form for simultaneous linear algebraic equations, Ty! = - 031579T,! = 14737 - 0.18750T,! +7 -0.18750T,!' = 13.750 - 0.37507,! +7! 13.750 Solving using a standard routine on a hand-held calculator, = 21.662°C, T,! = 21.932°C, T,! = 21.974°C. Comments: 1. ‘The Gauss-Seidel iteration has converged to 0.01°C in four iterations. 2. The physical time step corresponding to a mesh Fourier number of 0.3 is At = FoAx’/a, = (0.3{0.001)1(3.37x10°) = 0.0890 s. 3. Notice that T,, docs not change while iterating at the first time step—-it remains equal to 20°C.Exercise 3-144 Given: An electrically heated type 316 stainless steel bar, 4 mm thick, cooled by impinging water on each side. Required: Computer program to solve for the temperature response using the implicit finite difference method. id) Effect of mesh size and time step on profile at 100 s. Assumptions: 1. Surface conditions do not vary on the surface of the bar. 2. Constant thermal properties and FR heating. —*—y>—p*—- — bus (Equation (3.89) becomes AU = QAt+Q,At, and ! adds a term Q," (1.1)AxAt to the energy balance used en i to derive Eq. (3.95), which for an interior node m | becomes o Fo(Tyy 1 *Tyrn) *FoQ, le 1+2Fo m In deriving the approximation for the surface node with convection, the termQ, (1.1)(Ax/2)At is added to the energy balance; modifying item 2 of Table 3.8 for node 0, ja. 2FO(T,"! BIT!) +FoQ.” Ax + Te (2) 1+2Fo+2FoBi e Since node M is a plane of symmetry (adiabatic) we can set Bi = 0 in Eq. (2) to obtain 2FoT yy +F0Q, Ax 7k +Ty 1+2Fo ia M From Table A.1a for 316 stainless steel at 300 K, k = 13 W/m K, 0 = 3.37x10* m? s. Hence Fo = aAvAx? = (3.37x10 “*)(AUAx?) Bj = h,Ax/k = (6500)Ax/13 = 500Ax Q)" ax 7k = (8.67x10)(Ax)/(13) = 6.669x10 Ax?3-144 (continued) MATLAB was used to solve the difference equations. The table below shows the effect of mesh size and time step on the centerplane temperature at t = 100 s. ‘The convergence criterion was 0.001 K. At s Ax, m 0.0005 0.0002 0.0001 4 2 cl 0s T.,8C 59.94 60.00 60.01 60.01 59.93 59.99 60.00 60.01 5988 59.95 59.96 59.96 Comments: 1. With a time step of 0.5 s all the mesh sizes give T, within 0.02 K. 2. A time step of 4 5 is too long to get an accurate solution. 3, Consult a text on numerical methods for methods used to determine how a finite ifference solution converges to the exact solution as time step and mesh size are decreased. Note that the above solutions have converged to within 0.001 K of the solution of the difference equations, not the differential heat conduction equation.Exercise 4-1 Given: Laminar flow in a2 cm LD. tube with known velocity and temperature profiles. Required: Bulk temperature, Ty, Assumptions: 1. Constant properties (p and cp). Since we are given the velocity and temperature profiles across the flows, Eq. (4.7) can be used to calculate the bulk temperature, i puT2nede R ‘T= where si can be found from Eq, (4.6) as ti = [ pu2nrdr a ° puT2mdr — fuTrdr Ten, Ty = ——= puted fu { J }.1[1~(£/0.01)?][400-3x109(1.875x10-0.251°+62514)](1) 375t-26,875x10°P-9,37SK1OH1.875x10777 ur=0.1r-1000r* & pene 17.1875R2-6.71875x10°R*—1.5625x10°R9423.4375x10°R® = 17.1875x104-6.71875x104—1.5625x10440.234375K10 = 9.14x10 R J urdr= 0.0SR?-250R4 = 5x10-6-2.50x10 = 2.5x10* ° ‘Then, Ty, = 914x104 /2.5x10 = 365.6 K Comments: 1. From the temperature distribution we can calculate the centerline temperature ‘T, = 343.8 K, and the wall temperature T, = 400.0 K. Then (T,—Ty)/CT.—T.) = 0.612 (CLI/18, which corresponds to fully developed conditions; see the analysis in Section 5.3).Exercise 4.2 Given: Velocity and temperature profiles for a turbulent flow of air between parallel plates. Required: Bulk velocity and bulk temperature. Assumptions: 1. Constant properties (p and cp). For flow between parallel plates spaced 2b apart the bulk. velocity, and bulk temperature are given by » Uy = Jusy a ol it Le ® » 2fpucpTdy fuTay n-2 a2 =e = 2 J pucpdy a ‘The required integrations must be performed numerically; using the trapezoidal rule we obtain uy = 19.4 m/s; Ty =305.9K Comments: 1, The ratio uy/ic, where u, is the centerplane velocity, is 19.4/22.3 = 0.870. 2. ‘The ratio (T,— Ta)((T.—T.), where T, is the wall tomperature, is 0.904. 3. ‘The bulk values are within about 10% of the centerplane values, which is typical of turbulent flow of gases. The ratio (Ts — Te)(T; — Te) can be ‘much lower for liquid metals, but is almost unity for oils. Liquid metals ‘are characterized by low values of the Prandtl number (Pr = cp}/k), gases have Pr= 1, whereas oils have high Pr values.Exercise 4-3 Given: Flow ina duct. Required: Friction factor-skin friction coefficient relationship. Assumptions: 1. Fully developed flow so that there is no fluid acceleration. For generality consider a duct of arbitrary cross section Mr with cross-sectional area A, and perimeter P. A force balance on the fluid contained in a length of duct L Be requires that the pressure force equal the viscous drag on. “ 15PL (PHAP IAS. the wall 77 (P-AP)A.= TPL PA. From Table 4.3, f=4Cy Comments: 1, ‘This result is valid for laminar or turbulent flow. 2. See Exercise 5-19 for an alternative definition of friction factor that is independent of Reynolds number for fully developed laminar flow.Exercise 4-4 Given: A heated horizontal cylinder rotating about its axis, Required: Dimensionless groups pertinent to heat transfer. Assumptions: 1, Constant properties. 2. Natural convection is negligible. 3. The cylinder is long compared to its radius. 4, Viscous dissipation effects are negligible, We first limit the scope of the problem by making the simplifying assumptions listed above. Then we expect the average heat flux to have the following functional dependence: H=KOR WAT. Kd: ‘PR pS WAT ogi In terms of the primary dimensions kg, m, s, K, W, SEG atd ay ‘There are eight variables and five primary dimensions; thus we expect (8-5) = 3 independent fey dimensionless groups. A Nusselt number Nug =,R/KAT, and Pranddl number Pr= cpii/k can be obtained as was done for forced flow over a cylinder in Section 4.2.2. It remains to obtain a dimensionless group corresponding to the Reynolds number of that analysis. A suitable group can be written down immediately if we recognize that the characteristic velocity for this induced flow is the tangential velocity of the cylinder surface, RQ. Thus a Reynolds number can be defined as Reg =(RQVRP/U=R2Qp/. ‘Then Nug =F(Reg.P1) Comments: 1. Itis easy to see that the three groups are independent, since each contains vari- ables found in no other group. 2. When examining a new problem it is unnecessary to derive all pertinent groups from first principles: results obtained previously should be used when appropriate, 3. Clearly, the most difficult problem is deciding which secondary variables to ignore, In this case, ifthe rotation speed is too low, natural convection becomes important and perbaps dominant. Some degree of boldness is required. But all assumptions should be carefully listed, and their validity re-examined as the project proceeds.Exercise 4-5 Given: Central receiver for a solar power plant in the form of a vertical cylinder. Required: Dimensionless groups to correlate the effect of a cross wind on convective heat loss. Assumptions: 1. Constant properties except for the density variation causing a buoyancy force. 2. Negligible effects of viscous dissipation. ‘This is a situation involving both natural and forced con- vection, We expect the average heat flux to have the ee following functional dependence, i = £(0T.B.8,V.psHukscpsH.D) — H where notice that we have included two characteristic lengths, the height H and the diameter D. We expect H i to characterize the natural convection, whereas D should characterize the forced convection. Then peg Ve phtkc, ED WTI which in terms of primary dimensions kg, m, s, K and W, is wf} heel ale] Ey fs] oe bie] EP ‘There are eleven variables and five primary dimensions; thus we expect (11-5) = 6 independent dimensionless groups. From the dimensional analysis of natural convection in Section 4.2.2 we know that the following groups pertinent to heat transfer by natural convection can be obtained, ae 2a Noy = 22 the Nusselt number, a008 : gar; “BE. the Prands number w KAT Dp » For forced convection we can obtain a Reynolds number, Rep = A suitable remaining group is easily seen to be a geometric parameter H/D. Thus Nug= 222 pat, PoReo, B Comments: 1. ‘The choice of H (rather than D) for the Nusselt number Nuy is arbitrary, and was made because we thought about the natural convection problem first. We could ‘equally well use Nup, but then the functional dependence on the geometric parameter H/D will change accordingly. 2, Check that the groups are indeed independent.Exercise 4-6 Given: Packing of a perforated plate heat exchanger for a cryogenic refrigeration system. Required: Dimensional groups for correlating (i) pressure drop and (ji) heat transfer data. (i) Pressure drop. The first step is to identify the pertinent variables. As for any viscous fluid flow, we expect AP to depend on the fiuid velocity through the holes V, the hole diameter d, the hole pitch p (or equivalently the open area ratio ¢,), the fluid density p, the fluid viscosity jt, and the additional geometric parameters, the plate thickness t, and plate spacing s. In addition, the specified low pressure (~100 Pa) indicates that rarefied gas effects may be significant: hence the transport mean free path, 4, (see exercise 2-4) is a possible variable to be included. Also the low pressure suggests flow velocities may be high enough for compressibility effects to be important: hence the speed of sound, a, should be included. Thus AP = £(Vid.ep,PsbhSsbh,a) n= AP Veatestptptstt -fa] ey ere] fey eee sm] [s m| [ms s ‘There are 3 primary dimensions and 10 variables; thus there are (10-3) = 7 independent dimen- sionless groups. These can be written down by inspection. First we select the Euler number (see Eq, (4.17) to characterize the pressure drop across the holes, Eu = AP/pV", and of course, the Reynolds number, Re = Vpd/i will characterize viscous efits, The plate open area ratio ¢, is dimensionless. The other two geometrical variables can be formed into independent groups as s/d and 1/4, as can the mean free path as Jd = Ka, the Knudsen number. The last independent group must be formed with the speed of sound a; V/a = M is the familiar Mach number, Thus Eu = f(Re,€5,8/4,tld,Kn,M) (ii) Heat transfer. The additional variables are the fluid thermal properties, thermal conductivity K and specific heat cp, and a characteristic temperature difference AT =T,—T,,. Thus Ge = £(V,4,ep,P.Hs ep, S,th,2,AT) RG VP ep pul ct kh sue al AT™ Referring to the treatment of high speed flow in Section 4.2.2, we will use four primary dimen- sions, kg, s,m and K.4-6 (continued) Ea] fs} ery es] Lee] Pre] ete eb] With 13 variables, there are (13-4)=9 independent dimensionless groups. Again these may be ‘identified by inspection. The Stanton number is selected to characterize the heat transfer rate between the plates and the fluid, St = 9,/pcp VAT. The Reynolds, Knudsen and Mach numbers arise again, as do the geometrical parameters ép, s/d and td. In addition, the Prandil number, Pr = pik is a dimensionless group of fluid properties. Lastly, for heat transfer in a high speed flow, 1 the Eckert number is a pertinent dimensionless group, Ec = - V"/c,AT. Thus, St= fiRe,e,,8/d,/d,Kn,.M,Pr,Ec) Comments: 1, Check that these groups are independent. 2, Experimental results show that we have overestimated the number of independent variables. Since hydrogen and other candidate gases exhibit near ideal behavior, the fluid properties are related and the number of independent dimensionless groups is reduced. For example, Re, Kn and M can be related.Exercise 4-7 Given: A constant property, low speed boundary layer on a flat plate. Required: Dimensionless groups characterizing the growth of hydrodynamic and thermal boun- dary layers. Assumptions: 1. Heating commences at the leading edge. (@ Hydrodynamic boundary layer 8. For reasons which i 8 will become apparent later, we will take a shortcut by eee recognizing that the growth of the boundary layer is a diffusion process (diffusion of vorticity) characterized by only one fluid property, the kinematic viscosity v (also called the momentum diffusivity). ‘Thus we propose the following functional dependence for 3: B= f(ue,xv) ae fF om fms ‘There are four variables and two primary dimensions; thus we expect two independent groups. By inspection 8/x is a dimensionless group. Then combining x with the remaining variables gives a second independent group, a Reynolds number Re, = uex/v. Hence 8/x=fRe,) x = Subxv' (i) The thermal boundary layer thickness 6. If heating commences at the leading edge the appropriate length variable is again x. We expect 8, to depend on the same variables as 5, and, in addition on the thermal diffusivity, o.= k/pe, which should determine the relative thicknesses of the two layers (for o> v, 8, > 8, and vice versa). Thus, 8. =f(0,.2,V,0) > n= Shubxvial = iny{a| im" [=| 5 s ‘There are now (5-2) = 3 independent dimensionless groups. By inspection 8,/x and v/a. (the Prandtl number) are dimensionless groups, and an independent third group is a Reynolds number Re, = ugx/V. Hence 8,/x = ffRe,,P:) Comments: 1. Ifthe heating commenced at a location x = x,, then x; is an additional variable, and 3/x= fRe,,Prixi/x). 2. Repeat part @) postulating a dependence on } and p, rather than on v. 3. Repeat part Gi) postulating a dependence on p, 1, ¢p and k, rather than on v and o.Exercise 4-8 Given: (@) Water flowing in tube. Gi) Airin free convection on a vertical wall. Required: Equivalent stagnant film thickness. Assumptions: 1. Convective heat transfer coefiicients are given by Eqs. 1.21-1.23 of Chapter 1 (@ From Table A.8 the properties of water at 300 K are k= 0.611 Wim K, p = 996 kg/m?, cy = A178 Jikg K, v= 0.87x10° m’/s, First check the Reynolds number, Rep = VD/v = (5.0)(0.02)/(0.87x10-*) = 1.149x10° ‘Since Rep > 10,000, Eq. (1.22) for turbulent flow applies new 0.003 Lake) _ 10,023)15)%*(0.611*996x4178)%4 peyo4 0.02)"2(0.87x10°6)* Hence, 8, = k/h, = 0.611/1.6x10* = 3,82x10™ m (0.0381 mm). Gi) Estimate air properties at 300 K, which is the average of the plate and ambient temperatures: from Table A.7, k = 0.0267 Wim K, v = 15.66x10~6 m?/s, First check the Grashof number, Gry = (BAT) Bx? /v? = (20/300)(9.81)(0.2)° /(15.66x10°°)? = ‘Since Gr, < 10°, Eq. (1.23a) for laminar flow applies, he = LOT(AT/x)* = 1.07(20/0.2)* = 3.38 W/m? K. 8, =k/h, = 0.0267/3.38 = 7.89x10 m (7.89 mm) = 1.60x10" W/m? K 13x10" ‘Comments: 1. Comparing the results for each situation, 8, for the water flow is very small, only 38 pm; for the air flow 6, is much larger, almost 1 cm. 2, Physical chemists and chemical engineers often use this equivalent stagnant film concept.Exercise 4-9 Given: Experimental data for heat loss from a 8 em major axis elliptical cylinder in cross flow. Required: Convective heat loss from a geometrically similar cylinder of 12 em major axis, at 30°C to 24.5 m/s air low at 23°C. Assumptions: 1, Secondary variables, such as wind-tunnel blockage and free-stream turbulence, had a negligible effect on the data. If the dimensional analysis for forced flow over a circular cylinder of Section 4.2.2 is repeated for an elliptical cylinder of major diameter D,, minor diameter, Dy, with the major axis at angle @ to the flow, the result may be written in the form Nip, =f(Rep, ,Pr,D,/Dy,8) In the tests, Pr, D, /Dy and @ are all constant. Hence the tests can yield only Nup, =f(Rep,); Pr=0.7, D,/D,=1.33, 0=0 Our strategy will be to prepare a graph of Nup, versus Rep, from the test data, and use this graph '2.cm cylinder at the value of Rep, corresponding to the given flow to obtain Nip, forthe D, conditions. Properties will be evaluated at the mean film temperamure, using Table A.7. For each test case we will calculate a Reynolds number and Nusselt number (based on the 8 cm major axis). The results are tabulated below. Test Rep, — deony -Nup, T6420 452 36.1 2 14,040 219 82.1 3 18230 488 = «112 4 30540 238 = 150 5 46400 379 © 236 6 41410 378 198 7 43370 376 = 230 8 43503 517-230 9 56,760 448357 10 56870 522-297 11 73,850 527 400 12 89,220 561_-494 For the 12 cm major axis cylinder with properties evaluated at the mean film temperature of4-9 (continued) 00: 400 300 Mp, 200 ~ 100 20000 40000 60000 80009 ” 100000 Rep, 26.5°C =299.7K, Rep, = (4.5)(0.12)/15.62x10 = 34,570 From the graph, Nu=170 Bh, = (K/L)Nu = (0.0267/0.12)(170) = 37.8 W/m? P= 2n{(a?-+b?)/2]"* = 2n{(0.12°+40.08")/2)" =0,641 m Qeoav /L = he ATP = (37.8)(7)(0.641) = 170 W/m. Comments: 1. Note the scatter in the data, 2. Most of our knowledge conceming convective heat transfer coefficients is based on experiment. It is important for the student to lear to interpret and use experi- ‘mental data correctly.Exercise 4-10 Given: Heat transfer coefficient test data for R-113 flowing over cylindrical studs. Required: Correlation of data, ‘We expect the Nusselt number Nup = hD/k to depend on T the usual groups for forced flow over a cylinder, namely Rep _y and Pr as found in Section 4.2.2. In addition, we expect @ dependence on the geometric group p/D where p is the stud i pitch, because the flow pattern will certainly depend on the stud spacing. Also, since the studs are relatively short, end effects may be important; hence there is a possible depen- dence on the geometric group H/D, where H is the stud height. Hence Nup = f(Rep,Pr.p/D,H/D) However, Pr, Dip and D/H were not varied in the tests; hence we can only correlate the data as Nup = fiRep) ; Pr=8.52, pD=2, Hid=4 After calculating Nup = h,k/D, Re, = VD/v, the following table can be constructed. tet | v | on | Nu] Rep ms | Win? K 1.0 | 1810 | 242 | 2.3x10* 16 | 2430 | 325 | 38x10 24 | 3150 | 422 | S.6x10% 3.1 | 3730 | 499 | 7.3x10¢ 3.6 | 4130 | 553 | 8.sxi0* wenn Plotting on log-log paper and fitting a straight line through the data suggests the following power-law correlation Nup = 0.416ReB™ ; 2.3x10"< Rep <8.5x10"; Pr=8.52, pD=2, H/D= Comments: 1. ‘The data base is clearly quite limited in scope. In particular, data for other values of p/D would be useful in optimizing the stud configuration. 2. Data for the stud temperatures should be supplied to allow properties to be evaluated at the mean film temperature,Exercise 4-11 Given: Flow at V = 3 m/s ina 10 m length of 2 cm LD. tube; T, = 320 K, Ti, = 300 K. Required: Average heat transfer coefficient and pressure drop for water. Assumptions: 1. The correlations in Section 4.3.1 are of acceptable accuracy. ‘We expect h, to be large (moderate k, turbulent flow): assume Tew ~ 320 K and hence the aver- age bulk temperature is Ty, = (4)(300+320) = 310 K. Properties of water at 310 K from Table A.8 are: k= 0.628 Wim K, v=0.70x10% m/s, p = 993 kg/m?, cp = 4174 J/kg K. Pr=4.6. Rep = VDW = (3)(0.02)/(0.70x10) = 85710, >2300, hence turbulent Eq, (4.42): £= (0.790 InRep-1.64)? = 0.0186 . (BYRep-1000)Pr Eq, 445): Noo= To nega pzaay 35 ‘To correct for variable propery effects evaluate U: = WG20K) _ 84x10 _ 9 549 Hy HGIOK) 6.95x10+ From Table 4.6 for heating a liquid in turbulent flow: m=0.25,n=-0.11, f= fy(Us/tig)” = 0.0186(0.840)° = 0.0178 Nu=Nug (s/o)? = 435(0.840)1! = 443 nh, = Nuk/D = (443)(0.628/(0.02) = 13,920 Wim? K APIL = (QD\CAPV) = (0.0178/0.02)(4(993)(3)? = 3.98x10° Pa/m For turbulent flow of a liquid with Pr = 4.6 the entrance length is less than about 10 diameters so entrance effects can be neglected and h, and AP given above can be taken as average values. To check our estimate of Tay we use Eq. (4.11); fepL __bexDL_ ah __4 3.920000) __ Thc, — PORD*/A)Vcp_ DV pcp (0.02)(3(993)(4174) Thou = TT Ty in)®xD(B, xDitincy) = 320-(320-300)e- 2% = 317.9 K Ty = (12)(TouTia) = 309 K ~ 310 K as used. Comments: 1. CONV, Item 1 gives h, = 13,935 Wm? K, AP/L = 3978 Pa/m.Exercise 4-12 Given: Flow at V = 3 mAs in a 10 m length of 2.cm LD. tube; T, = 320 K, Tig = 300 K. Required: Average heat transfer coefficient and pressure drop for SAE 50 cil Assumptions: 1, The correlations in Section 4.3.1 are of acceptable accuracy. ‘We expect he to be small (low k, laminar flow): assume Tay ~Tiy and Ty = 300 K, From Table A.8, oil properties at 300 K are: k = 0.1445 Wim K, v = 570x10° m?/s, Pr= 6600, p = 883 kgim?, c, = 1900 Ike K. Rep = VDA = 105.3 < 2300, laminar 4 —2.06S(DALIRepPr__ 140.04[(D/L)RepPr}*> Eq. 4.50): Nup = Eq. (4.39): f= 64/Rep = 0.608 Hs _ 0.137 _ 0.979; from Table 4.6 for heating a liquid in laminar flow, m = 0.58,n Hy 0.503 £=0,608(0.272)" = 0.286, Nup = 18.8(0.272y°" = 21.7 O11 h, = Nupk/D = 156.8 Wim? K AP/L = (iD) [3 | = (0.286/0.02)(4)(883)(3)? = 5.68x10* Pafm. Notice that the hydrodynamic entrance length is given by Eq. (4.48) as L/D ~ (0.05)(105) = 5 so that entrance effects for f can be neglected. Finally check the outlet temperature using Eq. Gin): ‘Be2RRL he(eDL) Abel 4(156.8)(10) =0.062 they PERD*/AYVcy VD pc, (3X0.02)(883)(1900) Toor = Te—(T eT le = 390-(320-B00)e° PO? = 301.2 K ‘The oil temperature change is small, as assumed. Comments: 1. CONV, Item 2, gives he = 156.4 W/m? K, AP/L = 5.678x10" Pa/m.Exercise 4-13 Given: Flow at V =3 mfsin a 10 m length of 2 cm LD. tube; T, = 320K, T; Required: Average heat transfer coefficient and pressure gradient for mercury. Assumptions: 1. The correlations in Section 4.3.1 are of acceptable accuracy. 300K. We expect h, tobe high: assume Ty ~ 320K and T, = 310K. Mercury properties at 310 K from Table A.9 are: k= 8.54 Wim K, v= 0.1075x10- m/s, Pr= 0.0241, p = 13505 kgin?, cp = 140 Jikg K. Rep = VDN 58,000, turbulent Eq. (4.46): Nup=4.8+0.0156ReB SPP = 42.2 Eq. (442): f= (0.790 InRep-L.64y? = 0.0129 Bs 15x10 Eg LAI = 0.975 Hp 1L.451x10% Since j1,/Hy ~ 1.0 variable property effects are small and can be neglected (Eq. (4.46) could give errors greater than 10%). hg = Nupk/D = 18,020 W/m? K AP=(fD) ev" = (0.0129/0.02) [3]essi0 =3.92x10" Pa/in Entrance effects on heat transfer are negligible for turbulent flow of liquid metals. To check the outlet temperature use Eq, (4.11): H2eRL BDL) abe they pRD?/4)Vo, VDP cp Comments: 1. CONV cannot be used to check these results since it does not contain heat transfer coefficient correlations or properties for liquid metals. 2. Notice that the Re and Pr range restrictions on Eq. (4.46) are satisfied. 320 K, as assumed. $6, lage => ThowExercise 4-14 Given: Flow at V = 3 m/s in a 10 m length of 2 em LD. tube; T, = 320K, Ty = 300K. ‘Required: Average heat transfer coefficient and pressure gradient for air at 1 atm pressure, Assumptions: 1. The correlations in Section 4.3.1 are of acceptable accuracy. Although h, will be small, so is th and we expect a large Nu anid Tox ~320K. From Table A.7 air properties at 310 K, 1 atm pressure are k = 0.0274 W/m K, p = 1.141 kg/m?, v = 16.54x10 js, Pr= 0.69, cp = 1005 kg K. Rep = VDWv = 3628, barely turbulent Eq. (4.42): £= (0.790InRep-1.64)? = 0.0428 i (£18 (Rep-1000)Pr Bq. 445): Nov = => caayeepeaay 722 T J, 7 1032 ~ 1.0: variable property eflects can be ignored 'y he = Nu k/D = 16.7 Wim? K (0.0428/0.02) [soso = 11.0 Pafm For L/D = 200 diameters, Table 4.4 shows that entrance effects will be at most 4% and thus should be ignored. Finally, use Eq. (4.11) to check Tox! h2mRL _he(xDL) 4h 7 = 210=> Thou Te = 320 ed. icy pCRD*/A)Ve, VDP, ‘ peenarnae Comments: 1. CONY, Item 1, gives h, = 16.38 Wim? K, AP/L = 10,90 Pa/m. 2. Inthis transitional regime the correlations for f and Nu are not reliable,Exercise 4-15 Given: Flow at V=3 m/s in a 10m Jength of 2 cm LD. tube: T, = 320K, Ty = 300K, P= 1.013x10° Pa Required: Average heat transfer coefficient and pressure gradient for helium. Assumptions: The correlations in Section 4.3.1 are of adequate accuracy. Although h will be small, o is rh, and we expect a large No, and Tyy, ~ 320K. From Table A.7, properties of helium at Ty = 310 K are k = 0.152 W/m K, v= 131x10 m/s, Pr p=0.157 gim?, cp = 5200 J/kg K. Rep = VDV = (3)(0.02/131x10 = 458, laminar 0.065(D/L)RepPr Peete rere rere ar om ge 140.04{(D/L)RepPr?* Eq. (4.39): f= 64/Rep = 0.140 Eq, (4.50): Nap =3.66+ From Table 4.6, m= I, n= 0; hence correcting f, f= (0.140)(1.032)= 0.144 he = NuQW/D) = (3.70)(0.152/0.02 = 28.1 Wim? K PIL = (1D) 1 oye Vv 2 Pl =(0.144/0.02) [3] 157)G)=5.09 Pam Finally, use Eq. (4.11) to check Tou: Romer _Re@DL)__4hel = = = 20=> Thou =Ts = 320K, as assumed. tic, PERD?/A)Ve, VD cp ee Comments: 1. CONV, Item 2, gives h, = 28.14 Win? K, AP/L. 5.087 Paftn.Exercise 4-16 Given: ‘Therminol 60 at 3 m/s in a 0.05 m LD. pipe. Required: Variation of h, and AP/L with average temperature. Assumptions: 1. No entrance effects. ‘CONV, Configurations 1 and 2 can be used. ‘Therminol 60 is Quid No. 8. T, is set equal to Ty, to obtain results characterizing heat transfer at small AT. Property values are available for 230 10,000 ‘The flow is turbulent and Eqs. (4.46) and (4.47) are applicable. @ Uniform wall temperature: Nup = 4.8+0.0156Re Pro = 4.8+(0.0156)(77900)°55(0.025) h, = Nupk/D = (12.06)(50.7)/(0.01) = 61,200 W/m? K. Gi) Uniform wall heat flux: Nup = 6.3+0.0167Ref$ Po = 14.07 he = Nupk/D = 71,350 Win? K = 12.06 Comments: 1. These heat transfer coefficients are very large. It is for this reason that lithium is used as a coolant when high heat fluxes are to be absorbed, notwithstanding the severe materials handling problems associated with liquid metals. 2. The uniform wall heat flux boundary condition gives the higher heat transfer coefi- cient (12% higher. 3. CONV cannot be used to check this result since it does not contain correlations for liquid metals. 4, Notice that the Re range restriction on Eqs. (4.46) and (4.47) is satisfied, but that the Pr range restriction is not satisfied. Thus our results may not be accurate enough, and more appropriate correlations should be obtained.Exercise 4-18 Given: Air flowing in a 1 em LD. pipe. Required: Power input to increase bulk temperature from 300 K to 330 K. ‘The steady flow energy equation, Eq. (1.4) applied to the 20 Ty, K mn length of heated pipe requires that 330 Q= me, ATS 300 Evaluate properties at the average bulk temperature of 315 K: Foran ideal gasp = PM = 1810929) _1 107 g/m? }~20em—>| RT (8314)(315) — From Table A.7, cp = 1005 J/kg K. -107)€2)(/4)(0.017 = 1.74x10~ kg/s Q= (1.74x10}(1005)(30) = 5.24 W Comments: 1. Notice that Q does not depend on the length of the heated section. 2. It was not necessary to calculate the heat transfer coefficient to obtain this result, However, if we were required to find the corresponding wall temperature, a calcula- tion of he would be necessary.Exercise 4-19 Given: SAE 50 oil flowing at 0.5 kg/s through a 25 mm diameter tube, 60 m long. Required: Tew: if Tiy = 300 K and T, = 370 K. Assumptions: 1. Fully developed hydrodynamics at commencement of heating. Guess Ty, = 320 K, and hence evaluate properties at 310 K: k = 0.1435 W/m K, p = 877 kg/m?, Cp = 1950 J/kg K, 1 = 2500x10~ kg/m s, Pr = 3400. Rep = VpD/it= maD/A, 11 = (0.5)(0.025)/(n/4)(0.025)2(0.25) = 102 Since the flow is laminar, Eq. (4.50) applies, 0.065 (D/L)RepPr 140.04[(D/L)RepPr]** (AL RepPr = (0.025/60)(102\3400) = 144.5, Np = 3.66 + 0.005) 044.5)__ _g 13 14(0.04)(144.5°% Correcting for variable properties; using Table 4.6, Nu = Nug(pt/lp) 1 = (8.13)(189/2500)™! = 10.80 Ihe = (K/D)Nup = (0.1435/0.025)(10.80) = 62.0 Wim? K ‘The outlet temperature is obtained using Eq, (4.11): henDLitincy = (62.0)(n)(0.025)(60)/0.5)(1950) = 0.300 Tron TT Thinde oO = 370-370-300)e = 318.1 K which is close enough to our guess of 320 K for property evaluation. Nup =3.66+ Comments: 1. Since the flow is laminar the heat transfer coefficient is relatively low for a liquid, and even though the tube is 60 m long, the temperature change of the oil is relatively small. 2. Check using CONV.Exercise 4-20 Given: Liquid sodium flowing at 3 ms in a 5 cm LD. pipe. Required: Nusselt number and heat transfer coefficient for (i) uniform wall temperature and Gi) uniform wall flux, ‘Assumptions: 1. Fully developed hydrodynamics. From Table A.9 evaluated at 1000 K, k = 59.3 Wim K, p = 772 kg/m®, w= 1.8x10~ kg/m s, and Prs 0.0038 Rep = VpDipt = (3(772)(0.05Y(1.8x10~) = 6.43x10° ‘The flow is turbulent; thus Eqs. (4.46) and (4.47) apply. @ Uniform wall temperature: Nup = 4.840.0156ReB PO = 4.8+(0.0156)(6.43x10°)"*5(0.0038)°%? = 12.4 hh, = Nupk/D = (12.4)(59.3)(0.05) = 14,700 Wim? K (i) Uniform wall heat fux: Nup = 6.3+0.0167ReB SP = 14.4 hg = Nupk/D = (14.4)(59.3)(0.05) = 17,100 Wim? K Comments: 1. Entrance effects are negligible for turbulent flow of liquid metals, hence the length of the pipe is not relevant. 2. These heat transfer coefficients are very large. For this reason sodium is often used a a coolant when large heat fluxes must be absorbed, for example, in breeder-type nuclear reactors, and in internal combustion engine valves. Unfortunately the use of sodium presents many problems due to its explosion potential, toxicity, and incompatibility with other materials. 3, Notice that the Pr value of 0.0038 is not within the range specified for Eqs. (4.46) and (4.47); however, the extrapolation is modest.Exercise 4-21 Given: Aqueous solutions of ethylene glycol flowing in a 10 m tong, 3 em I.D. tube. ‘Required: Pressute drop and heat transfer coefficient for bulk velocities in the range 0.2-4.0 m/s. Assumptions: 1. Negligible entrance effects for turbulent flow. 2. Properties can be evaluated at 270 K. CONV is well suited to a parametric study of this kind. We first check the Reynolds number range, The lowest value will be obtained at V = 0.2 m/s and 60% glycol (highest viscosity); using Table A.13a, Remin = PVD/L = (1089)(0.2)(0.03(14.1x10™) = 463, laminar ‘The highest value corresponds to V=4 m/s and 20% glycol, Remax = 1034)(4)(0.03)(3.6x10°°) = 34,470, turbulent ‘Thus, depending on flow velocity and % glycol we can have either laminar or turbulent flow, and must use Configuration 2 or Configuration 1 of CONV accordingly. Solution using CONV. aaesi ‘The required input is: Configuration number = 2 (Re < 2300), 1 (Re > 3000) Fluid: 3,4,5,6 and 7 1,=210 7 T)=270 wo] os P=any value ae ‘= any vi oe D=0.03 sist Sane via ‘The results are graphed opposite. em Ed i Se a “ E oo = 7 Vas) Comments: 1. With increasing % glycol, h, decreases and AP increases (except in a small range of velocity). ‘Thus when using glycol as an anti-freeze one should not use a stronget ‘concentration than necessary to prevent freezing.Exercise 4-22 Given: Pressure drop for flow of a crude oil through a 3 m long, 1.2 cm LD. tube. Required: (i) Dynamic and kinematic viscosities, (i) convective heat transfer coefficient! Assumptions: 1. Thermal properties approximately the same as those of SAE 50 oil. 2. The inside of the pipe is smooth. 3. Uniform wall temperature, ‘We first calculate the friction factor from Eq, (4.15), f= AP(DIL/U2)pV* From Table A.8 the density of water at 370 K is 960 kg/m?, thus the oil density is approximately (0.862)(960) = 828 kg/m?. The oil velocity is then V = ma/pA, = (0.2)/(828)(m/4)(0.012)? = 2.14 ms £= (31,6x10*)(0.012/3)/(0.5)(828)(2.14)* = 0.0666 ‘The Moody chart, Fig. 4.49, shows that the flow must be laminar for a smooth pipe. Hence from Eq, 4.39) Rep = 64/F = 960 B= VDpiRep = (2.14)(0.012)(828)960. v= p/p =2.68x10% mss. 21x10* kg/m s Gi) From Table A.8 for SAE 50 oil at 370 K, k = 0.137 W/m K, c, = 2200 J/kg K. Hence Pr= cpt/k = (2200)(2.21x10°}/0,137 = 355. For laminar flow Eq. (4.50) gives the average Nusselt number, Nip = 3.664 —2205@/LRepPr__ 140.04((D/LJRepPr}** =3.66 150.01; (355) + = 18.6 140.04((0.012/3)(960)(355)]°° he = Nupk/D = (18,6)(0.137Y0.012 = 212 Wim? K Comments: 1. Notice the large entrance effect for heat transfer: Nup is about five times the fully developed value.Exercise 4-23 Giver. Flow of oil between 0.25 m long parallel plates spaced 3 mm apart. Required: Average heat transfer coefficient and pressure drop. Assumptions: 1. Fully developed hydrodynamics at inlet. Evaluate properties at T, = 360 K; from Table A.8, k= 7 0.138 W/m K, p = 848 ke/m?, v = 30.1x10 m/s, { 255x10~ kg/m s, Pr= 400, 8005 3 m/s 1.006 m Dy = (20.003) = Rep, = VD,/V = (3)(0.006)/30.1x10 = 598 < 2800, hence laminar. From Table 4.5, fy = 96/Rep, = 96/598 = 0.161 Eq, (4.51) gives the average Nusselt number, _ 0.03(D,/L)Rep, Pr (0.03)(6/250)(598)(400) see ea ag 35. Nun. 754+ Tiel Dy/DRen PDP” * Tro ore(c6/250\98 40072" °° Correct for variable properties. At T, = 400 K, 1 = 88.4x10~ kg/m s; from Table 4.6, m = 0.58, n=-0.11; £= fy (Hs/He)” = 0.161(88.4/255)°* = 0.0871 Na = Ng (4, /p)™ = 35.6(88.4/255) 4 = 40.0 ‘h, =(K/Dy )Nup, = (0.138/0.006)(40.0) = 920 W/m? K ar=snoy|t vj = (0.0871)(250/6)(0.5)(848)(3)* = 13.85 kPa Comments: 1. Use CONV, Item 3, to check these results: with Ty, = 360, T, value, plate spacing = 0.003, L = 0.250, V = 3, CONV gives 40.0 and h, = 921 W/m? K = 0.0868, Nu =Exercise 4-24 Given: SAE 50 oil flowing at 4 m/s in a 3 m long, 1 cm x 1 cm duct. T, = 330 K, Tq = 290K. Required: Average heat transfer coefficient, outlet temperature, and pressure drop. Assumptions: 1. Pally developed hydrodynamics at commencement of heating. Guess Tox ~ Tin and evaluate properties at 290 K: k = 0.146 W/m K, p = 889 kg/m?, c, = 1850 Jikg K, v = 1230x10~ m/s, Pr= 13,900. Dy =4A,/P = 4(0.01)7/4(0.01) = 0.01 m Rep, = VDyW = (4)(0.01/1230x10~ = 32.5 < 2800 , laminar From Eq. (4.49), Lgy/Dy = 0.017(32.5)(13,900) = 7680; Len = (7680)(0.01) = 76.8 m i.e., thermal entrance effects are important. Modify Eq. (4.50) for a square duct by replacing 3.66 by 2.976 (from Table 4.5) and D by Dy. 8.065Dy/LMREn.P)_ 2.976 + —0.06510.01/31(32.5)(13,900). 140,04((D,/L)RepPr]? 140.04{(0.01/3)(32.5)(13, 900)? Nup, =2.976+ Be = (/Dy)Nup, = (0.146/0.01)(18.6) = 272 Win? K ‘To obtain the outlet temperature use Eq. (4.11) with 2xR replaced by the perimeter P, RePL 272)(0.04(3) an = 0.0496 tcp —_4)(889)(0.01)°(1850) Text © Te(TyTr ne = 330-(330-290)e- 00S = 291.9 ‘Hence our use of 290 K as bulk temperature for evaluation of properties is adequate. Correcting for variable properties using Tables 4.6 and A.8, 7 ot =e [i] = [es] = (272)(0.0730)%"! = 363 Win? K HPLitincy = 0.0662, Thou = 292.6 K From Table 4.5 f= 57/Rep, = 57/325 = 1.75. Correcting for variable property effects, £= fy(Us/bte)® = 1.75(0.0730)° = 0.383 Pressure drop AP = fL/Dy (1/2 pV*) = (0.383)(3/0.01X0.5)(889)(4)? = 0.817 MPa Comments: 1. The correction for the thermal entrance effect is an approximation, but is probably good enough for an engineering calculation,Exercise 4-25 Given: Annular duct 9 m long, 3 cm LD.,4 cm O.D. Air flow 0.05 kgs; P= 10 atm, Tia = 245K. Steam at 1.1x10° Pa condensing inside. Required: He, Tous AP. Assumptions: 1. Airside thermal resistance dominates: hence the inner tube wall is at uniform temperature of Tyy(1.1x10° Pa) = 375.3 K. Guess Toe ~ 375 K, and evaluate air properties at the ‘mean bulk temperature (245+375)/2= 310K. k= 0.0274 Wim K, p = 11.41 kg/m, w= 18.87x10° kg/m s, Pr= 0.69, ¢p = 1005 kg K. Dy=D.-D;=0.01 m, Ae = (R/AMD3-Df) = (a/4)(0.04?-0.037) = 5.50x10 m? 2D, (0.05)(0.01) Aa” Gsourhasenars Eq. (4.42): £= (0.79 In Rep,—1.64)? = (0.79 In 48,200-1.64)? = 0.0211 (G8)(Rep,-1000)Pr___(0,0211/8)(48,200-1000\0.69) _ 4.45): Nup, = = = 1002 Fa G49): Noy 7s @2PA) ~ 1+12.7(0.0011/8)%(0.692-1) ‘The Nusselt number is next corrected using Eq. (4.52a), p,)* 3)" Nup, = Nup, (0.86) (2 = (100.2)(0.86) [3] =902 Correcting for variable properties, Table 4.6 gives m = -0.2,n=-0.55 for heating of a turbulent 8 flow, . [T; }* 315.3 Nup, = Nupp F ona 310 Ty fh4.25 (continued) hh, = (K/D, )Nup, = (0.0274/0.01)(81.2) = 222.5 Win? K Since L/D, = 9/0.01 = 900, entrance effects wil be negligible. To calculate Ty, we adapt Bq. G1) as, De2RRy _ 222.5V2RV0.015)9) _ 5 76 Titey (0.05)(1005) Toot = Te(T. “Tine = 375.3-(375.3-245)e* = 372.3 K, which is close to the ori- ginal estimate of 375 K for evaluation of properties. AP=f(L/D,) [: ov) wa] 0m which is small compared to the inlet pressure of ~10° kPa, Hence the change in density due to pressure change along the duct can be ignored. 6.62 kPa, PeeeEEEeEE DUO ceeeeeeeee car (2)(11.41)5.50x104)* Comments: 1. Check these results using CONV as follows: (i) CONV gives AP/L = 737.4 Pa/m; hence AP = (737.4)(9) = 6,64 kPa Gi) CONV gives h, = 247.6 Wim? K; for an annular duct heated on the inside Eq, (4.522) gives the required correction: the = (247.6)(0.86)(3/4) 97° = 223.0 Wim? K. 2, Notice the use of the radius of the inner tube wall (across which the heat transfer occurs) in the modified Eq, (4.11). 3, Table A.7 was used for k, 11, Pr and cp since these properties can be taken to be independent of pressure at moderate pressures.Exercise 4-26 Given: Air at 1 atm flowing at 0.05 kg/s through an annular duct. Steam at 1.1x10° Pa conden- sing inside. Required: Average heat transfer coefficient, outlet temperature and pressure drop. Assumptions: 1, Inner wall isothermal at steam saturation temperature, ° 2. Fully developed hydrodynamic and thermal conditions. Dj =0.03 m;D, = 0.04 m ‘T, =375.3K (Table A.12a) Ta = 240K Guess Tay = 360 K and hence evaluate air properties at (1/2)(360+240) = 300 K; from Table A. p= L177 kein, c, = 1005 Jhkg K, p= 18.43x10° kg/m s. Dy =4A,/P =D,-D, = 0.01 m; Ac = (R/4)(D3-DP) = (n/4)(0.047-0.037) = 5.50x10~4 m? Rep, = ThDy/Acht = (0.05)(0.01V(5.50x10™)(18.43x10) = 49,300 ‘The flow is turbulent. Using CONV Configuration 1, we first obtain: AP/L = 7053 Pain. ‘Nu =904; hy = 241.3 Win? K ‘Then we correct h for an annular duct with heat transfer through the inner wall using Eq. (4.528), he = 0.86(D/Dp)°19(241,3) = 217.3 Win? K ‘The outlet temperature is obtained by adapting Eq. (4.11) to an annulus, RexD\L = hePLsitcy = (217.3)(0.03)(9)/(0.05)(1005) = 3.67 Tout = Ts(T Tin EXP AD Ley) = 375.3-375.3-240)¢ 2" = 371.8 K Since our guessed outlet temperature was 360 K, a second iteration will be made using Toy = 372, Repeating the above procedure gives AP/L = 7245 Pafm and he = 220.7 Win? K. Then AP = (72A5)(9) = 6.52x10* Pa. Comments: 1. Notice the use of of the inner wall diameter in Eq. (4.11) because heat transfer is through the inner wall only.Exercise 4-27 Given: Helium flowing at 0.01 kg/s in a bundle of tubes of a shell and tube heat exchanger. Required: Graphs of h,, h,P and AP/Ax versus tube diameter for a fixed flow ‘cross-sectional area of ()) 0.007 m* for laminar flow, and (ii) 3x10“ m? for turbulent flow. Assumptions: 1. All properties may be evaluated at 200K, 10° Pa. 2. Fully developed fiow conditions. 3. Constant wall temperature. From Table A.7, helium properties at 200 K are k= 0.116 Wim K, ¢» = 5200 Ikg K, p= 15.6x10~ kg/m s, Pr= 0.70, p = (10°)(4\/(8314)(200) = 0.241. (0) Laminar flow. CONV will not be used for this case because D and A. cannot be specified independently in item 2. V=1iypA, = 0.01/(0.241)(0.007) = 5.93 m/s Rep = pVD/= (0.241)(5.93(DYI15.6x10 = 9.16x10'D ‘Number of tubes N= A> 2300 the flow is certainly turbulent. Now using Fig. 4.20 the following table is wall temperature this occurs when T, = T,, constructed. 90° elbow 90° round bend x Wd dh dh hy cm Win? K Wim? K 1 22 339 16 246 6 2 18 27 14 216 3 16 24600125198 aaieeceaa: 14 216 12 185 18 6 13 200 12 185 mw 8 12 1500 SIT 5. )psesees Cpeeesaas ©) 85 caeeeeT ds cateec 7 4 15 iS Lt 169 Comments: 1. A third figure is given for h,, but not more than two figures are significant owing. to the limitations of the graphs. 2. Since Ty, vaties at most 10 K in this situation, we are justified in neglecting the effect of property variations along the duct. Exercise 4-35 considers a case where property variations along a duct should be accounted for.Exercise 4-34 Given: Five ducts of various cross sections. Required: Merit indicator Svf for air in the Rep, range 107-3x10*, Assumptions: 1. Fully developed hydrodynamic and thermal conditions. 2. For turbulent flow, circular tube correlations apply with D = Dy, 3. Heating at uniform wall heat flux. For laminar flow (Rep, < 2300-2800), Nup, and fRep, can be obtained from Table 4.5. Then, Suf = NWRePrf = Nu/(0.69)(fRe), for air with Pr = 0.69. Cross Section NuyfRep, svt @ Circular 3.657 64 0.0828 Gi) Square 2976 37 0.0757 Gi) Triangular 24 33 0.0656 (iy) 1:3 rectangle 4.0 oo 0.0840 (¥) Lis rectangle 7.541 96 o.114 For turbulent flow (Rep, > 2300-2800) we cannot use the simple power laws Eqs. (4.43) and (4.44) since Eq. (4.43) is only valid for Re > 4x10‘, Hence the need to use Eqs. (4.42) and (4.45), though, Eq. (4.42) is strictly valid only for Re > 10°, (#18)(Rep, -1000)Pr 1+12.7(6/8)"7(Pr 29-1), f = (0.790 InRep - 1.64); Nup, Rep, f Nup, Su 6x1 0.0365 194.134 10* 0.0315 296 0.136 2xlot 0.0262 SLO 4x10 0.0221, 87.1 (0.143. Comments: 1. SUf is higher for turbulent flow than for laminar flow—the higher pressure drop is offset by an even larger increase in heat transfer. 2. Use of hydraulic diameter in Eqs. (4.42) and (4.43) for the 1:6 rectangular is not too accurate, Check an advanced text to obtain a more accurate correlation. 3. Use of the simple power laws Eqs. (4.43) and (4.44) gives SUf = 0.156, independent Reynolds number. Even at Re = 4x10‘ it differs by 11%.Exercise 4-35 Given: Air flowing at 0.14 kg/s in a 5 cm LD, tube, with walls heated uniformly at 9x10* W/m?, Required: Local heat transfer coefficients for a 90° elbow entrance and Ty, = 300 K, Assumptions: 1. Turbulent flow. 2. A stmooth tube wall ‘The variation of bulk temperature along the tube is obtained from a simple energy balance--as was done to derive Eg. (4.8), which gives ay2nRax dT, 2nR a =a From Table A.7 we take ¢, = 1010 Jikg K as a constant. Then 76 —} > Wha AT, x10*)(2)(8 10.025) 199 Kim g = x G1HA0I0) Consider the location x = 5 cm; then x/D = 1.0, and from Fig. 4.20, h/h. 300+(0.05)(100) = 305 K. The wall temperature is T, = Ty+a/, = 305 +(9x104/(2.2)h,_ Since by... depends on T, we must iterate using CONV (item 1) to obtain T, = 715 K, h,., = 99.8 ‘Win? K, h, = 216 Wim? K. In this manner the following table is constructed. x wD Ty te Te teh cm K Ko Wh? K Win? K Client detste sectieg Cliteiet Stier <7 )ot 10 2 310 18 85 916 165 15 3 315 16 6962870139 2 4 320 14 10 814 4 306 «6-330 «431195 80.0104 40 8 340 122 1275 789 «© 963 50 10 350 118 1310 794 = 93.7 75 15 375 11S 1315 834 © 959 x Comments: 1. ‘The graph shows a schematic of the behavior of b, and b,, for heating of a gas (this case) and for cooling of a gas. 2. Check the pressure drop: is it adequate to assume the pressure remains at 1 atm in order to calculate h,? 3. Investigate the anomalous behavior of fa.Exercise 4-36 Given: Double-tube heat exchanger to cool water against 0.4 kg/s of 20% ethylene glycol solution; a2 em LD., 1 mm wall thickness inner copper tube. Required: Outer diameter of annular duct to give equal hot- and cold-side heat transfer coefficients. Assumptions: 1. Bully developed turbulent flow. 2. Outer shell perfectly insulated--Eq. (4.52a) applies. 3. ‘Thermal resistance of copper tube negligible. First we obtain the inside heat transfer coefficient. Since b,, and h,. are to be equal and the copper tube thermal resistance can be assumed negligible, the wall temperature is T, (12YTyp#Tyc) = (1/2)(290+280) = 285 K. Assuming fully developed conditions and turbulent flow (check later), we can use CONY, item 1. With T,, = 280 K, T, = 285 K, D = 0.02 m, we obtain h,; = 3197 Wim? K. Next we use CONY, item 1, to obtain h,. for a range of hydraulic diameters (D,, = D,-D;, and A, =(x/4)(D,”-D,*) for an annulus). Then Eq. (4.52a) is applied, namely, the correction factor 0.86(D/D,)™'* is used to obtain the h,,, for an annulus from h.,, for a circular duct. Interpolation gives the diameter required to match heat transfer coefficients. The results are summarized below, 6 Dy Ae h, (CONV) _ h, (annulus) m m mm Wien? K Win? K 003 0.008) 3.27K10* 5313 0035 = 0.013 5.8210 2999 0.034 0.012 5.28x10* 3303 3045 0.0335 0.0115 5.01x10* 3478 3199 Interpolating gives D, = 0.0335 m (3.35 em) Comments: 1. Check the CONV output to see that both flows are turbulent. 2. Compare the hot- and cold-side pressure drops.Exercise 4-37 Given: R-22 flowing in a 10 m long, 2.cm LD. tube. Required: Pressure drop and heat transfer coefficient for 0.01 Rey, a precise specification of Re, is unnecessary. 2. Radiation heat transfer must be included to complete the surface energy balance.Exercise 4-41 Given: Automobile oil pan on vehicle traveling at 80 km/h, Required: ‘Heat loss from five exposed sides of pan. Assumptions: 1. The sides and bottom can be treated as flat plates. 2. Rey = 5x10". 3. The average heat transfer coefficient on the front and back is the same as that on the sides. Fora complicated flow such as this, our model must be rela- tively crude; however an accurate answeris not required. A | 142293 low value of Rey is taken because of the large scale tur @15m[ 1.0395 a bulence expected undemeath the automobile, CONV will Ff. 9 7m ope bbe used to objain the average heat transfer coefficient. The required input is Configuration number = 5 (Flat plate: turbulent boundary layer) Fluid = 21 (Ain) P= 1.0133x10° 7 V = 80/3.6 = 22.22 Rey = 5x10" ‘The output includes: Re, = 7.93x10° fh, = 66.0 Wim? K ‘The total surface area is A= (0.7)(0.15+0,3040.15)+(2)(0.15)(0.3) = 0.51 m? Q=h,ACT,=T.) = (66.0)(0.51)(393-293) = 3370 W Comments: 1. The front and back contribute only 20% to the total area. Hence our assumed value of h, does not have a large impact on the total heat loss. 2. The boundary layers on the sides and bottom do not commence at a sharp leading edge, is as required by our flat plate correlations. But the error incurred is probably small.Exercise 4-42 Given: Diver 1.8 m tall swimming at 8 km/hr in 286 K water. Required: Rate of heat loss from diver. 1, Rey = 10. 2. Diver's skin temperature = 297 K. Assumptions: To estimate h, suppose a flat plate 1.8 m long. If the wet suit is a good insulator most of the tem- perature drop will be across the suit so assume T, close to water temperature, and hence T, ~ 286 K. From Table A.8, properties of water at 286 K are: k= 0.582 Wim K, v= 1.22x10 ms, Pr=8.8 Rey = (8/3.6)(1.8)/1.22x10* = 3,28x10%, turbulent at trailing edge. Using Eq. (4.65), with Re, = 1x10°, Nu os 0.664Re Pr” + 0.036Re2 Po? [ a (ie ] os = 0.664(105)*8.8)"" + 0.036. 28x105)°(8.8)°* [: 7 [zs] ] = 434 + 14,050 = 14,480 ‘hy = Nuk/L = (14,480)(0.582)/(1.8) = 4680 Wim? K water layer suit convection 297k Hee 0.003 2.05.4 8TH Referring to the thermal circuit, Q 297-286 a A 178 Wim? A 0.001 , 0.003, _1_ 4.746040.2)10% 0.606 * 0.05 * 4680 ‘where k for the trapped layer of water has been evaluated at 297 K. Also, A~(1.8)(8)(0.35) = 1.98 m* Q=(178)(1.98) = 352 W (= 303 kcal/hr) |. Using CONV, liem 5, with T, = T, = 286 K, V = 2.222 mA, Rey = 10°, gives bh, =4675 Wim? K. 2. Since ii, is very large its precise value is unimportant. Indeed, for wet suit ‘design we should simply assume that the outer surface is at the water temperature! CommentsExercise 4-43 Given: A 10m by 10m heat rejection area at the bow of a submarine available for heat rejection. Required: Average temperature of heat rejection surface when rejecting 15 MW. Assumptions: 1, The surface can be modeled as a flat plate with a transition Reynolds number taken as 10°, Guess a mean film temperature of 310 K; from Table A.8, v= 0.70x10° m/s, Rez = VLW = (4)(10Y(0.70x10) = 5.71x10" ‘The boundary layer is turbulent over all but the initial 0.2% of the surface. Eq. (4.65) applies, and CONV will be used to obtain the result. The surface temperature needs to be guessed, and iterated upon if necessary. The required input is Configuration number = 5 (Flat plate: turbulent boundary layer) Fiuid = 1 (Water) ‘T, = 320 K (first guess) ‘The output includes: h, = 6611 Wim? K ‘The surface temperature is then T, = QA) + Te = (15x10°Y(6611)(10") + 288 = 311 K Using this value as a second estimate of T,, CONV gives h, = 6315 Win? K, and T, = (15x10°)/(6315)(10") + 288 = 312 K No further iteration is necessary. Comments: 1, Heating of a water boundary layer tends to delay transition from laminar to turbulent flow. Thus our value of Rey = 10° may be too small.Exercise 4-44 Giver: Flow along a flat plate, 0.1 < V < 100 m/s. Required: Average heat transfer coefficient and drag force on a 1 m length of plate for air at @ 1 atm, and Gi) 0.01 atm. Assumptions: 1. Transition Reynolds number = 10° 2. Properties can be evaluated at 295 K. ‘Parametric studies such as this are easily done using CONV. The required input is: Configuration number = 4 (Re < 10°), or 5 (10° < Re < 3x10") Fluid = 21 (Ain) 7, =295K T.=295K P= 1.0133x10°, 1.0133x10° L=10 V=0.1-100 Rey = 10° ‘The following table can be constructed from the output v @ Air at 1 aim Gi) Air at 0.01 am Rey, F ih Rey, F hy mys Nin Win? K Nim Wim? K 01] 6580 9.8x10% 1.26 66 © 98x10 = 0.126 03 | 19740 Saxlo* 2.18 197 5.Ax105 0.218 1.0} 65800 31x17 3.98 658 3.1x10* 0.398 3.0 | 197x10° 25x10? 10.8 1974 16x10* 0,689 100 | 6ssx10° 0.255 33.4 | 6580 98x10? 1.26 30.0 | 1.97x10° 1.99 849 | 19,700 S.AxlO? - 2.18 100.0 | 6.58x10° 18.6 228 «| 65800 © 0.31 3.98 Comments: 1. Transition begins to occur on the plate at 1 atm for V between 2 and 3 m/s; at 0.01 atm the boundary layer is always laminar. 2. The heat transfer coefficients at 0.01 atm are very small; even at 100 m/s they are of the same order as natural convection coefficients at 1 atm.Exercise 4-45 Given: Flow of four liquids along a 1 m long flat plate at velocities of 0.01, 0.1, 1, and 10 mis. Required: Average heat transfer coefficient and drag force per unit width. Assumptions: 1. Properties can be evaluated at 300 K. 2. A transition Reynolds number of 10°. 3. Natural convection negligible. Due to the large range of velocities and fluid viscosities, transition may or may not occur on the plate, If x, <1 m, the average heat transfer coefficient can be obtained from Eq. (4.57) as 0.664(K/L)Re,”Pr", Pr> 05 a) For mercury Pr << 1 and Eq. (4.58) must be used, = 1.128Q/L)Re,”Pr!? @) ‘The drag force on the plate per unit width is F = 7L(1) = TI) = 7, and 7, f, is obtained from Eq. (4.55) as (12)pV 2, = (1/2)pV2(1.328)Rey", Re, > 10? @ If x, > 1 m, Bq. (4.65) gives the average heat transfer coefficient as (K/L)10.664Re,.Pr'? +0.036Re,Pr°“[1 +(ReJRe,)°*)) @ for Pr > 0.7. For Pr < 0.7, which applies to mercury, the text does not contain a correlation for heat transfer (see Comment No. 1). The average drag coefficient is obtained from Eq, (4.62), fie 0523 [Re 0.523, © 1.328Re,"j — |»_9°23 _ Re, | in?0.06Re, (Re | 1n?0.06Re, CONV will be used to do the calculations for water, SAE 50 oil and refrigerant R134-a. Since CONV does not contain heat transfer comrelations and properties for mercury, the mercury calculations must be done by hand. At 300 K, Table A.9 gives mercury properties as k = 84 Wim K, p = 13530 kg/m’, v = 0.110x10* m?/s, Pr = 0.025. ‘Then for V = 0.01 m/s Re, = VLWv = (0.01)(1)/0.110x10* = 9,09x104 < 10° i.e,, laminar. ‘Then from Eq. (2)4-45 (continued) fh, = (1.128)(8.4/1.0)(9.09x10 40.025)" = 452 Wim? K and from Eq. (3). F 1/2)(13530)(0.01)7(1.328)(9.09x10 4)" = 29.8 N For V = 0.1 m and larger, x > x, and we cannot calculate heat transfer. The results for the four liquids are tabulated below. h, Wim? K FN Liquid V,ms | oor 01 1 10 | oo on 1 0 Water 786 287 3090 20.790 | crvat0* 21600? 198 143 SAE 50 oll 763 (238754 28 140x107 0.443 140 442 ‘Mercury 452 - - 7 298x10° 0.277 20.1 1468 R-134a 1992 182 128.1 8218 | 333x104 259x107 190 138.1 Comments: 1. Always check to see if the convection correlation is valid in the range of parameters of concern. 2. A good reference for liquid metal heat transfer R. Stein, "Liquid Metal Heat Transfer", Advances in Heat Transfer, vol. 3, 1965.Exercise 4-46 Given: Helium at 1 atm, 300 K flowing at 30 m/s along a 1 m long flat plate, also at 300 K. Required: Plot of local shear stress and heat transfer coefficient. Assumptions: 1. The transition Reynolds number is 1.5x10°. From Table A.7, properties of helium at 1 atm and 300 K are k = 0.149 W/m K, p = 0.1624 kg/m’, v = 124.0x10° m/s, Pr = 0.70. The local Reynolds number is Re, = Vx/v = (30)x/124.0x10 © = 2.419x10°x, a) For Re, < Re, = 1.5x10°, Eq. (4.54) gives the local shear stress as Taq = (1/2)PV 2(0.664)Re “Y= (1/2)(0.1624)(30)(0.664)Re,"” = 48.53Re,"” @Q) For Re, > 1.53108, Eq, (4.60) gives Tq = (1/2)(0.1624)(30)°(0.455)(In 0.06Re,)* = 33.25 In(0.06Re,)? @ For Re, < 1.5x10%, Eq. (4.56) gives h, = (K/x)(0.332)Re, "Pr? ha, = (0.149)(0.332)(30/124.0x10 )7(0.7)!2x “Y= 21.60x a) For Re, > 1.5x10°, Bq. (4.64) gives h,, = (k/x)(0.029)Rey “Pr? t h, = (0.149)(0.029)(30/124.0x10 “)(0.7)"2x © = 75.15x 2 (6) eer ae ria hy 197—y las 09] |] a srearstwes wnt +00 m Nim? Wh? K gg] | ~#-- Haat rnafercoecont Wink ae 0.0 0 a0 ae & 07 O1 02419 0312 83 Z oe. m5 02 04838 0221-483. gg ane 03 0.7257 0.180. 394 & 0 = 04 09677 0156 342 & 4° 3 05 1209 0.140 «305 B89 Pole 06 «1451 0127-279 o2 0 0.7 1.693 0.394 80.7 a 7 08 1935 0.380 78.6 = is 09 2177 0370768 > Onin SUMNER’ ISRISTET PURIST TIEUETEE 10-2419 0362-752 a Comments: 1. Notice that the virtual origin of the turbulent boundary layer is the leading edge (x = 0) as an engineering approximation, Equations (2) and (4) indicate t,, and h,, are infinite at the leading edge. This is not true since boundary layer theory is not valid for Re, < 300. CONV was not used here because it does not calculate local coefficients.Exercise 4-47 Given: Air at 360 K, 110 kPa flowing at 15 m/s over 2 1 m long flat plate maintained at 300 K. Required: Average shear stress and heat transfer coefficient as a function of Re, for'5x10* < Re, < 5x10°, Assumptions: 1. Virtual origin of the turbulent boundary layer is the leading edge. 2. Variable property effects can be accounted for by evaluating all properties at the mean film temperature. CONV can be used for the calculations. Using item 5 the required input is Fluid number = 21 (air) Wall temperature = 300 Ambient temperature = 360 : Pressure = 1.1x10° Length = Free stream velocity = 15 Transition Reynolds number = 5x10‘ - Sx10* °° ‘The table below summarizes the results. ggg aay Resiuaaee b, oso. Nim? Wim? K = Eos Sxi0* 0.5614 49.19 - 1x10° 0.5377 47.00 aa 2x10° 0.4904 42.68 3x10° 0.4439 38.49 4x10 03982 34.43 ne Sx10° 0.3532 30.46 Cr ee er Roy Commenis: 1. Note that CONV gives Re, = 8.862x10°. Hence, at the lowest value of Rey, Ky = (0.5/8.86)(1) = 0.056 m, whereas at the highest value of Re,, X, = 0.56 m. 2. Note that all inputs into CONV are in SI units 3. The appropriate value of Re, is usually not known precisely. It is useful to perform a parametric calculation such as this in order to gain an appreciation of the impact of the chosen value of Rey. Oqwna) “4Exercise 4.48 Given: Air at ~10°C blowing at 10 m/s over a 30 cm O.D. pipeline. Required: Beat loss per 100m. Assuraptions: 1, Outer surface of the pipeline is maintained at 40°C. Te z at, Wim K, v = 14.610 mss, Pr = 0.69. 2sava13) = 288 K. Properties of air at 288 K: from Table A.7, k= 0.0260 Rep = VD/v = (10)(0.3)/(14.6x10) = 2.055x10° Eq. (4.71b) applies = _ _0.62Re5PrF Np 08+ TT oan (0.69)18 ys 03 + 2.8212.055x105 0.69) [+ =] ] [140.470.6977] ae ‘he = Nupk/D = (404)(0.0260)/(0.3) = 35.0 Wim? K Q= b.xDL(T,-Te) = (35.0(4(0.3)(100)(40--10)) = 165 kW Comments: 1. Use of CONV (tem 6) with T, = 263, T, = 313, P= 1.013x10°, D = 0.30, V = 10, gives Nup = 404 and h, = 35.0 Wim? K, 2. The oil temperature decreases along the pipeline: Eq, (4.11) et. seq. can be used to determine the outlet temperature if given the inlet temperature and oil flow rate. However, first check to see if the oil-side and wall thermal resistances are negti- gible.Exercise 4-49 Given: Oxygen bubble of 1 mm diameter rising from 3 m below the surface of a water pool at 300K. Required: Rise velocity. Assumptions: 1. Bubble has reached its terminal velocity. A force balance on the rising bubble gives Buoyancy force (Fx) = Drag force (Fp) (1) where Fa = (p-p_XnD*/6)g, and Fp = Cp(nD?/4) > pv? ‘To solve for the velocity an expression for Cp is needed. Assume 2 Pg? Assume 500 < Rep < 2x10°; hence from Fig. 4.28, Cp = 0.44, Assurne the hailstone is at 0°C = 273 K and evaluate the properties of air at a mean film temperature of (27342872 = 280K: v= 13.91x10~ m/s, Pr= 0.69, py = 1.265 kgfm?, k = 0.0255 W/m K. % 6 ye | 2@PeeP (BYO10-1.265)(9.8110.008) |" _ 11.39 mjg 6Cop, (6)(0.44)(1.265) ‘The Reynolds numberis Rep = YD. = 11.3210.006) . 4553 v 13.91x10% Eq. (4.76) applies, Nop =2+(0.4Refs + 0.06ReB?)Pr°* = 2+{(0.4)(4883)"+(0.06)(4883)"*(0.69)"* = 41.0 Liny= [eae 41.0 = 174 Wim? K D 0.006 Comments: 1. Notice it is the frontal area (nD?/4) that appears in the expression for drag force. 2. Rep = 4883 is within the range specified for Eq, (4.76). 3, Use CONV to check Nup and i,Exercise 4-51 Giver: An electrically heated 1.905 cm diameter cylinder cooled by a 20 m/s cross flow of air at 97 kPa and 290K. Required: Stagnation line and average cylinder surface temperabures. Assumptions: 1, The surface of the oylinder is heated uniformly. 2, Circumferential conduction in the wall is negligible. @ Stagnation line temperature, = QA=h(T,-T.) QUA = 60/(n)(0.01905)(0.12) = 8355 Win? ‘The heat transfer coeflicient can be obtained from Eq, (4.70). To evaluate properties we guess T, (1/2)(330+290) = 310K, From Table A.7, = 18.87x10~ kg/m s, k= 669. Density is obtained from the ideal gas law as p = PM/RT = (@7x10°}(29)/(8314)(310) = 1.091 kg/m? Rep = pVD/AL = (1.091)(20)(0.01905Y(18.87x10) = 22,030 Nup = 1.15Re{Pr! = (1.15)(22,030)*(0.69)" = 151 1h, = Nugk/D = (151(0.0274)/(0.01905) = 217 ‘Ty = (Que) + Te = (8355/217}4290 = 328.5 K hich is close to our guess of 330 K, so that the property values are satisfactory. ii) Average surface temperature. Referring to Eq. (4.82) we can define an average heat transfer coefficient for a uniformly heated wall by the relation > i eee ge Ratt: T= J Te)dx ‘This value of h, can be taken equal to the value for an isothermal wall, that is, as given by Eq. (4.71). We expect T, to be higher than the stagnation line value, and so guess T, = 350 K and evaluate properties at T, = 320 K: = 19.29x10 kg/m s, Pr= 0.69, k = 0.0281 W/m K, p= 1.057 kgim?. Rep = pVD/t = (1.057)(20)(0.01905)/(19.29x10) = 20,880 Equation (4.71b) gives the average Nusselt number as, 0.62Re Pr? | - | Rep }) Nop = 03+ oan 282,000 0.3+(0.62)(20,880)"(0.69)'"[14+(20,880/282,000)"Y/[1+(0.4/0.69)°9 = 88.6 hh, = Nupk/D = (88.6)(0.0281/(0.01905) = 131 Win? K, T, = fig) + Te = (8355/131)+290 = 353.8 K, and again, no iteration on T, is required. Comments: 1. Kays and Crawford [15] give an exponent of 0.4 on Prandtl number for the stagnation Nup. This value will give more accurate results for Pr ~ 1: the 1/3 exponent gives less accurate results for Pr~ 1 but more accurate results for large Pr. Notice that in both cases, the correlations are not valid for Pr< 0.5.Exercise 4-52 Given: A cylindrical hot-film sensor to measure ocean speeds. Required: (j) Flow speed for given data, (ji) expected error, and (iii) thickness of bio-fouting deposit to give 5% error. Assumptions: 1, Cylinder is perpendicular to flow. 2, Seawater properties can be approximated by those of pure water from Table A.8. (@ Anenergy balance on the sensor gives Q=hAST: R= QAAT hh, = (1.101)/()(0.009\(0.03)(21.35-20.54) = 1602 Wim? K Evaluate water properties at a mean film temperature of 21°C = 295 K; from Table A.8, ‘Wim K, p = 998 ke/m?, v= 0.97x10"S m7, Pr Nup = h.D/k = (1602)(0,009y0.602 = 23.95 ‘We suspect a relatively low Reynolds number, using Eq. (4.71a), meas 0.62RefPr!> {ap = 0,3 + ——— I Tory] 0.62Re $6.7)!" 23,95 = 03+ 2 1+0.46.77 Solving, Rep = 440 (<10* so Eq. (4.71a) applies). V=(W/D)Rep = (0.97x10%/0.009)(440) = 0.0474 mvs (4.74 cm/s) Gi) V~ Rep « Nu « (T,-T.)? (0.0140.01)100 081 ‘Thus % error in V is (2)(2.47)=4.9% % error in (T,—T,) is Gi) We now have two resistances in series with U = 1602 W/m? K. usenet un Kk’ OR Uk Since V « he, a 5% error in V results from a 2.5% error in hi: thus we require 2. 0.025 [3] + or 8% (0.025\(1)/1602 = 15.6 ym Comments: 1. Hot-film velocity sensors are more robust than hot-wire sensors, and are used to measure steady flows (the thermal capacity of the quartz cylinder is too large to allow measurement of fluctuating velocity components in turbulent flows).Exercise 4-53 Given: Water at 300 K flowing at 2 m/s over a 3 cm diameter sphere maintained at 340 K. Required: (i) Drag force, Gi) heo/t,, and Gi) Q. @ Drag force F. Evaluate properties at T, = (1/2)(T,+T,) = (1/2)(340+300) = 320 K. From Table A.8, p = 989 kg/m?, v = 0.59x10 m/s Rep = VD = (2)(0.03)(0.59x10) = 101,700 From Figure 4.28, Cp = 0.44, F= (1/2)(pV?)A¢Cp = (1/2989)(2)"(00)(0.015)"(0.44) = 0.615 N Gi) The additional properties required are Pr = 3.8, k = 0.641 W/m K. Equation (4.75) gives the stagnation point Nusselt number as, Nup = 1,32Re$Pr! = (1.32)(101700)*(3.8)' = 657 ‘Equation (4.76) gives the average Nusselt number as Nup =2+(0.4Re$+0.06Ref)Pio+ = 24(0.4x101700%+0.06x101700"7)(3.8)" = 443 Nup/Nup = 657/443 = 1.48= hf, iil) b = Nupk/D = (443)(0.641)/(0.03) = 9465 W/m? K Q=h,AAT = (9465)(n)(0.03)°(40)= 1070 W Comments: 1. Check these results using CONV. 2. Recalculate Nup using the recommended viscosity ratio correction (a/h)"+ n= -0.25, rather than the mean film temperature. The effect will be seen to be small, 3, Notice the use of frontal area Ay = xR? to obiain F from Cp. 4, Kays and Crawford (15] give an exponent of 0.4 on Prandtl number for the stagnation point Nup. This value will give more accurate results for Pr ~ 1; the 1/3 exponent gives less accurate results for Pr ~ 1, but more accurate results for large Prandtl namber. Notice that in both cases, the correlations are not valid for Pr, = 19.63x10° kg/m s, Pr = 0.69. Rep = pVD/t = (1.080)(2.5)(0.01)(19,63x10) = 1375 From Eq. (4.712), 0,62Re$Pr!? 040.4" h, = Nupk/D = (18.1)(0.0286)((0.01) = 51.8 Wim? K (@ To find the temperature at the midpoint of the rod we use Eq, (2.36); with kay = 237 Wim K Nop =0.3 + = 0,3+(0.62)(1375)*(0.69)!9/1140.4/0.69)"9 = 18.1 y (Table A.ia), B= (heP/KA.)” = [51.8R 0.0L Y237IER/40.01P 1* = 9.35 Ta-Te 363-293 = 7, = 08 4.293 = 356 K Te= Goanpe * 7 costyo.a5x005) *223* 356 Gi) Using Eq. (2.38), Q =2(h,P/B)(Tp—Te)tanh BL (the rod is two fins) = (2)(51.8/9.35)(%)(0.01)(70)tanh(9.35x0.05) = 10.62 W Comments: 1. Check h, using CONV.Exercise 4.55 | Given: Aix at 10°C blowing across a4 cm O.D. uninsulated hot water pipe at 45°C. Required: Cost of heat loss/meter length/day. Assumptions: 1. The inside and pipe wall resistances are negligible. 2. A power cost of 8 cents/ KW h Evaluate properties at T, = (1/2)(T,+T,) = 1/2(318+283) = 300K, From Table A.7, k ‘Wem K, p = 1.177 kgfm?, w= 18.43x10~ kg/m s, Pr= 0.69. Rep = PVD/HL= (1.177)(2)(0.04Y(18.43x10-°) = 5109 Equation (4.71a) applies, 0.62RefPr® O40. ares) ie (0.6215109*(0.69) 7034 Tecoan.seey he = Nupk/D = (34.6)(0.0267(0.04) = 23.1 Win? K ‘The heat lost per meter length is QL =H, PCT ~Te) = (23.1((0.04)318-283) = 101.6 Wim, ‘The cost per day is Cost = (101.6 Wim)(24 1)(0.008 cents/W h)= 19.5 cents/m Nup =0.3+ Comments: 1. The importance of insulating hot water pipes in cold climates is evident. 2. Check h, using CONV.Exercise 4-56 Given: A 1 cm diameter alloy sphere suspended inside a furnace. Required: Steady state temperature of sphere. Assumptions: 1. A small gray body in nearly black surroundings for the radiation calculation. =350K P= 11 bar= Lixl0® Pa Vesms D=1em 2=040 Fora steady state the net heat flow out of the sphere is 2270: aa + Geoov = 0. colt 14) + TT) =0 2 Guess T, = 550K, and evaluate helium properties at T, = (1/2)(T,+T-)= 450K: from Table A.7, k= 0.192 Wim K, p= 26,3x10~ kg/m s, Pr=0.715. The density is obtained as p =PM/RT = (1.1x10°)(4)/(8314)(450) = 0.118 kg/m’. Rep = pVDjpt = (0.118)(5)(0.01(26.3x10) = 224 Equation (4.76) gives the average Nusselt number, Nup= 2+(0.4Ref + 0.06Ref?)Pr°4 = 24{(0.4)(224)* + (0.06)224)? 100.715)" = 9.17 ih, = Nupk/D = (9.17(0.192)(0.01) = 176 Wim? K ‘Substituting in Eq. (1), 6.67)(0.4)[(T./100)*~7.0°] + (176)(T,-350) = 0 Solving by iteration gives T, = 378.3 K Since our initial guess was much higher (550 K), we repeat the evaluation of h using T, = 378 K,T, = (1/2)(378+350) = 364 K, to obtain i, = 178.6 Win? K, T, = 377.9 K. Comments: 1. Check he using CONV. 2. Notice that we do not use v from Table A.7 when P # 1 atm, since v is pressure dependent, Instead we obtain p from Table A.7, and calculate p from the ideal gas law.Exercise 4-57 Given: A.2.cm diameter ice sensor on the Ieading edge of an airplane flying at 50 més. Required: Minimum heater power required to raise sensor to 0°C when the air is at -26°C. Assumptions: 1, The leading edge of the wing can be modeled as a 30 cm diameter cylinder. 2, Radiation heat loss is negligible, Neglecting the radiation heat loss, the minimum power is that required to balance the convective heat loss from the sensor when itis at 0°C. Q=hACT-T) where h, can be obtained from Eq. (4.70), Nup = 1.15RePr'3 Air properties are evaluated at T, = (1/2)(T,+T.) = (1/2)(273+247) = 260 K. From Table A.7, k= 0.0242 Win K, j1= 16.63x10 kg/m s, Pr= 0.69. The density is obtained from the ideal gas law = PM/RT = (61x10°)(29)/(8314)(260) = 0.818 ke/m?. PVD/t = (0.818)(50)(0.3(16.63x10) = 7.38x10° Nup = (1.15)(7.38x10)!7(0.69)"9 = 873, he = Nupk/D = (873)(0.0242)(0.3) = 70.4 Wim? K Q=h.ACT HT.) = (70.4)(R/4)(0.027[0--26)] = 0.58 W Comments: 1. Since h, is at most ~6 Wm? K, draa is negligible compared t0 qoany- 2. ‘The minimum power requirement is very small. 3. Check h, using CONV. 4, Atypical jet airliner speeds (~250 m/s) the effects of aerodynamic heating must be included (see Section 5.2). 5. Notice that we do not use v from Table A.7 when P # 1, since v is pressure dependent. Instead we obtain jt from Table A.7, and calculate p from the ideal. gas law.Exercise 4.58 Given: A.0.12 mm diameter platinum hot wire anemometer. Required: Current to maintain wire at 200°C for air velocities from 0.3-7 m/s. 1. The wire axis is normal to the flow velocity. 2. Radiation heat loss is negligible. Assumptions: V=03-7m/s T.= 20°C T, = 200°C 12mm ‘The electrical power input must be balanced by the convective heat loss, QL=PR=HPCTe) Win here the resistance R is per unit length, and R = (10x10™*/(x/4)(0.00012)? = 8.84 Q/m. CONV will be used to calculate h,. ‘The required input is: Configuration number = 6 (transverse cylinder) Fluid = 21 (air) T, =473,T, = 293 co P= 1.0133x10° ) 112x107 7 After obtaining h,, we can calculate Q/L = ii,(m)(0.12x10*)(473-293), and I= (Q/R)* = {(Q/Ly8.84}*, ‘The results are given in the table below. Comments: 1. The heat transfer coefficients are large due to the small diameter. Since, at most, /h, ~ 6 Wim? K, radiation losses are negligible. 2. Notice that at V=0.3 mis, Rep = 1.5Exercise 4-59 Given: A thermistor used to measure air temperature in a forced air heating system duct. Required: True air temperature when the thermistor records 46.8°C and the duct walls are at 412°C. Assumptions: 1. The thermistor can be modeled as a 3 mm diameter sphere with € = 0.8. 2. Radiation exchange can be calculated assuming a small gray body is large, nearly black surrounds. 3. Air properties may be evaluated at 1 atm pressure AC steady state, the surface energy balance Eq. (1.33) reduces to ne ear + Qras = 0 hA(T,-T,) +0¢,A(T/-T5 = 0 w ‘The heat transfer coefficient can be obtained using Eq. (4.76), Nip = 2+(04Rey” +0.06Reg")Pr°* Since we expect the error in T, to be small, we can evaluate air properties at T, = 46.8°C = 320.0 K. From Table A.7, air properties at 320 K, I atm include k = 0.0281 Wim K, p = 1.106 kg/m’, 1 = 19.29x10° kg/m s, and Pr = 0.69. Rep = VpDit = (1.3)(1.106)(0.003)/19.29x10% = 223.6 Nu, = 2+[(0.4)(223.6)"? +(0.06)(223.6)77}(0.69)"* = 9.06 KNWD 84.9 W/m? K (0.0281)(9.06)/(0.003) Substituting in Eq. (1), (84.9)(T.~T,) +(5.67x10 *)(0.8)(320.0°-314.4%) = 0 Solving, T-’ 382 K T, = 320.0+0.382 = 320.38 K = 47.2°C Comments: 1. An error of 0.4 K is probably acceptable in this sitnation. Note that a typical hand-held digital thermometer may have a certified accuracy of only 0.5 K. 2. Use CONV to check b,.Exercise 4-60 Given: Dolphins swimming at 3 m/s. Required: Minimum possible length for ocean temperatures in the range 0°C < T, < 25°C. Assumptions: 1. Length to diameter ratio L/D = 12 2. Blubber thickness to diameter ratio, UD = 0.05. Convective heat transfer can be calculated using a flat plate boundary layer model. We would expect that the minimum possible length is determined by the dolphin’s ability to maintain a steady internal temperature for a specified water temperature. An energy balance on a dolphin gives Q)'v = UAC,-T) ay ‘Approximating a dolphin as a cylinder, V = (RD7/4)L = (W/4)(L/12)"L = $.45x10 *L3 A = nDL+(2)(nD 7/4) = (m)(L/I2)L + (2)(/4)(L/12)? = 0.273L? WA = 5.45x10 *L4/0.273L? = 2.00x10*L, @ Lets}; ¢=0.05p = (005/12). = 4.17010 7L Uk h 417x107L | 1 a 2.08x107L +) @ Now we would expect that dolphins have evolved to have rather efficient insulation, that is, the thermal resistance of the blubber layer is much larger than the convective resistance. To check this assertion, use CONV to calculate some typical values of fi,. A few trials show that for 5°C 17.44x10 © At the desired conditions (7)(0.06) LRT. = 241x108, which is in the range tested 17.44x10 Re, = ‘A simple power law for the Nusselt number correlation is Nu = CRe "Pr ™ hD 7 eS H2 = {YP Ve, ob, = Caer WV "D™ k v If we ignore the effect of property variations, 30)", solving for n = 0.607. iB esses (ed: eer (TS Comments: 1. Notice that fs is much lower than the values measured in the tests. However 0-1 = ; solving, h,, = 36.0 Wim? K we have not extrapolated outside the range of the experiment because the Reynolds number is within the Reynolds number range tested. 2. Check that the worst case Biot number h L/k, for the test is 0.003 << 0.1. The lumped thermal capacity model is excellent for this situation.Exercise 4-67 Given: Water at 290 K flowing over a 10 cm square uniformly heated flat plate. Required: Water velocity and temperature distribution if the average plate temperature is 311.5 K when the heat loss is 1025 W. Assumptions: 1. Laminar boundary layer. 2. Back side of Teflon plate is perfectly insulated 3. Side effects are negligible. 4, Conduction along the plate is negligible. Equation (4.83) gives the relation between heat flux and average plate temperature for a laminar boundary on a flat plate. g,L/K(T-T,) = 0.680Re,!"Pr Properties are evaluated at the mean film temperature, T, = 1/2(2904311.5) = 300.8 K. From Table A.8, k = 0.6125, Pr = 5.78, v = 8.54x107 m’ss. 1. = (1025/0.12)(0.1)/(0.6125)(311.5-290) = 778 + = Nu, /0.680Pr 3)? = [(778)/(0.680)(5.78)"7F- = 4.07x105 V = Rew/L = (4.07x105)(8.54x107)/0.1 = 3.47 m/s ‘The local Nusselt number for a uniformly heated flat plate is given by Eq. (4.79), Nu, = 0.453Re, "Pr! ay. 2 8% 9.453 Y% "Pr; solving AT = iY KAT 0.453KV Pr (1.025x105)(8.54x10 “7)!x 4 EE. = 29010217 (0.453)(0.6125)(3.47)'(5.778)'* 290+. xm 0.001 0.01 0.05 o4 TK 2932-3002, 31283223 Comments: 1. Note that the coefficient in Eq. (4.83), namely 0.680, is only 2.3% different to the isothermal wall result (0.664 in Eq. (4.57). 2. Owing to the temperature variation there will be conduction along the plate with kd? T/dx? not zero. Hence the surface heat flux will not be exactly constant. However, since the stainless steel and ‘Teflon have a low thermal conductivity the heat conduction effect will be small (see also Exercise 2.97).Exercise 4-68 Given: Spheres with insulations having k valves of 0.03 and 0.1 Wim K. Required: Heat loss in air as a function of insulation outer radius. Assumptions: 1. One-dimensional steady conduction, 2. (TT) = 10 K to give a typical natural convection heat transfer coefficient. 3. Radiation heat transfer coefficient = 5 W/m? K. (@ From the thermal circuit and using Eq. (2.25b) for the resistance of a spherical shell, a and, Q/4n(T,-T,) = hg? fort, ,, and = kr, for 1, = Gi) CONV, item 11, can be used to calculate bh, for T, = 305 K, T, = 295 K, P=1 bar. Then |, #5 Wim? K. These results are tabulated with the results for parts (ii) and (iv). ii) For k = 0.03 Wim K and using the results of part (ji) the following table is constructed 2 fy mm | hy Wim! K | Qyan(T,-T,), Wim? Kx10° O1mm | 5=05 mm - - 0.3000 1.500 10 13.24 0.3023 1.560 5 1696 0.3039 1.604 25 23.55 0.3060 1.663 15 3175 03076 wu 1 41.60 0.3086 1743 08 48,84 0.3090 1.755 06 60.75 03091 1.759 0s 70.17 0.3090 1.754 03 1074 03071 a ou 290.1 0.2901 - We see that Q has a maximum of r, = r, = 0.6 mm, independent of 5, (iv) For k = 0.01 W/m K a similar table is constructed,4-68 (continued) Comments: 1 hy Wi? K | eyancr.=7,), Win? Kx10* 1.0mm | 1, =10mm ~ - 40.00 100.00 100 8.40 10.09 109.6 50 9.26 10.16 118.6 25, 10.50 10.25 132.9 15 1181 10.30 1409 10 13.24 10.25 132.4 5 16.96 9.65 S 25 23.55 782 _ 10 3175 3.18 ~ j, « 15 mm, independent of r, For r, = 0.1 mm and k = 0.03 W/m K, Q is essentially independent of r,, ie, the thickness of insulation (r,-r). If, is assumed constant, Eq. (1) gives a critical Biot number h,,/k = results show values of 1.22 and 1.77 for k = 0.03 and 0.1 Wim K, respectively. 9. Our More importantly, the Bi = 2 criterion cannot be used to give even an approximate result without iteration, because h, varies strongly with r,, Ifthe calculations are repeated for k = 0.025 and r; as small as 1 ym, there proves to be no critical radius~Q increases monotonically with r,. ‘The above results are rather insensitive to the exact convective situation since the Nusselt number is close to its lower limit value of 2.0. Use CONV to check this assertion for r = 1 mm by using AT = 2 K for natural convection, and forced convection with V = 1 cms. ‘This exercise demonstrates the importance of always checking that the assumptions used to obtain an analytical result are indeed valid for the situation of concem.Exercise 4-69 Given: ‘Thermocouple response in a high temperature argon plasma. Required: Local plasma temperature. Assumptions 1. Lumped thermal capacity model valid. 2. Negligible radiation heat loss to surroundings. ‘To use Eq. (1.38) for a lumped thermal capacity model of the bead response we assume negligible radiation heat loss from the bead and a constant heat transfer coefficient, TT, Ty-T, 1, = pe(V/AYhh, Equation (4.75) gives the average Nusselt number as Nu, 1p = 2+(0.4Ren”+0.06Rey Pr Rep = VDIV = (200)(0.003V7x10* = 85.8 Nuy = 2+[(0.4)(85.8)"? + (0.06)(85.8)°7(0.67)* = 6.15 1500-T, 500: Comments: 1. 2, 3 4, 1- = Nupk/D = (6.15)(0.17)(0.003) = 349 W/m? K pe(R/3yhh, = (7.3x105(0.0015/3)/349 = 10.5 s ose WS; 7, = 26,000 K To check radiation losses to the surroundings take Ty, = 300 K, € = 0.15 and evaluate at the maximum bead temperature of 1500 K, h, = 4oeT,,* = (4)(5.67x10°)(0.15){(1500+300)2) = 24.8 Win? Kysmall cf. h, = 349 Wim? K. ‘The elapsed time of 0.42 s is small and difficult to measure accurately. But, if a final temperature of 2500 K is chosen (still below the melting point), the radiation heat loss cannot be ignored. ‘A rather large bead is required to give a measurable response time--hence spatial resolution of the temperature field is poor. ‘We conclude that this temperature measurement scheme does not appear to be ‘promising. But the important lesson is that calculations are required to assess any proposal. Until calculations are made, the proposal is only speculation.Exercise 4-70 Given: Natural convection from a3 cm diameter horizontal pipe at 320 K. Required: ‘, when immersed in @) air at 2 atm pressure and (ii) water, at 300 K. Assumptions: 1. A long pipe so that end effects are negligible. (@) Air. T, = (1/2(300+320)=310K Propertiesofairat310Karek =0.0274WimK,@nu = (16.54x10-§/2) = 8.2710 m/s; Pr = 0.69. Also B= 1/T, = 1/310 K“. The Rayleigh number is 7 gBATD'Pr _ (9.81)(U/310¥20¥0.03)°(0.69) _ 1 ay4n105 < 10? Rap = Grppr= ABR aren ‘The flow is laminar, hence use Eq. (4.87), cai 7 . Nup =0.36+ 0.518Ra5 ¢ 0.518. 724x105)" _ U+0.559/P)"59 (14(0.559/0.69)"6}" bh, = Napk/D = (8.32)(0.0274)(0.03) = 7.60 Wim? K. ) Water. T; = (1/2)(300+320) = 310 K. Properties of water at 310 K are k = 0.628 Wim K, V=0.70x10- m’js, Pr=4.6, B= 361.9x10 K+ (Tables A.8 and A.10b). (9.81)(361.9x10-20)(0.03)°(4.6) _ 7 ie Rap = OSNGELIx107Y200.09°4-6) _ 1 g0x107 , laminar. i (0.70510)? ae 0.518(1.80x10")* Nap = 036+ 25180. 80x10 _ 393 036+ Ta 55014.6)96 07 h, = Nupk/D = (30.3)(0.628)(0.03) = 634 Wim? K Comments: 1. Use CONV, Item 10, to check these results, 2. For air, iy is low and thus radiation heat transfer may also be significant.Exercise 4-71 Given: Tank of water heated by a steam coil. Required: Time to heat water from 290 to 340 K, and the amount of steam condensed. Assumptions: 1. The thermal resistances of the tube wall and on the condensing side are negligible; thus T, =T,,, (1 atm) = 373 K. 2. Lumped thermal capacity model for the tank. 3. ‘The coil pitch is small and the lower tums have a negligible effect on the upper ‘tums: then the convection can be modeled as free convection on a horizontal cylinder. ‘The average bulk water temperature is T. = (1/2)(290+340) = 315 K. Then take T, = (12)(T.+T,) = (1/2)(3154373) = 344 K. From Tables A.8 and A.10, water properties at 344K are: k= 0.665 Wim K, v= 0.42x10% m?/s, Pr= 2.57 and B = 5.90x10 K"!. Equation (4.87) applies. 7 Rap = GrpPr= BATE De = G.90x104)(373-315)(9.8)O.0IP 25D _ 4 g0.04105 < 10° @.42x10°P Nip = 0.36 + sii ease ae 2212 Fi, = (e/D)Nu = (0.665/0.01)(21.2) = 1409 Win? K Ifthe tank is insulated and natural convection stirs the water well, we can assume a lumped ther- ‘mal capacity model o obtain the temperature response. From Eq, (1.38) A= Acai = RDL = ()(0.01)(8\(10)(0.5) = 0.493 m?, For water at 315 K, p = 991 kg/m?; ¢= 4174 J/kg K; also V = 107 m°. te = (91)(4174)(1074/(1409)(0.493) = 340-373 290-373 ‘Table A.12a gives the enthalpy of vaporization of steam at 1 atm as 2.257x10° J/kg. The amount of steam condensed is given by a simple energy balance: why, = pVcAT. Substituting, w(2.257K108) = (991)(101)(4174)(340-290); hence w= 9.16 kg, 95 5 5, t= 549 s=9.1 minutes Comments: 1. Asan alternative, model the coil as a vertical cylindrical shell. 2. Use CONV to check it.Exercise 4.72 Given: 2 cm diameter power cable with e = 0.7, resistance = 8.33x10~ Q/m. Required: T, for T; = 290 K when the current is 600 A. Assumptions: 1. The cable temperature can be taken as uniform due to the high thermal conductivity of aluminum, An energy balance per meter length of cable gives PR =(h,+h,)xD(T,-T.), or eth, = PR. (600)°(8.33x105) _ _ 477 RD(Ty-Te) (#)(0.02)(T,-290) T,-290 Guess h, «7 W/m? K for natural convection around a small object, and h, ‘W/m? K. Then an estimate of T, is T,-290 =477/(744.2)=42.6K; T,=332.6K To calculate h, precisely we now use a reference temperature T, = (332.6+290)/2 = 311 ~ 310K. From Table A.7, k = 0.0274 Wim K, v = 16.54x10°S m/s, Pr= 0.69; also B= 1/T, = 1/310. For Jaminar natural convection on a horizontal cylinder Eq, (4.87) applies, \TgD*Pr _ (1/310)(42.6)9.81)0.02° 0.9) _ 97 99 < 109 =42 Rap = GtpPr= "2 (16.54x10- Neos 37 meee Teossorpp > >” ie = (K/D)Nip = (0.0274/0.02)(5.37) = 7.36 W/m? K 1, = 40€T i = (4)(5.67x10*)(0.7)(310)° = 4.73 W/m? K Heth, = 12.1 W/m? K New estimate of T,: T,-290 =477/12.1= 39.4 K, T, =329.4K. Since this is close to our first estimate of 332.6 K, a further iteration is not required. Comments: 1. Natural convection in air is characterized by relatively low convective heat transfer coefficients. ‘Thus radiation can seldom be ignored. 2. Check b, using CONV, Item 10.Exercise 4-73 Given: A 1 mm diameter sphere in air at 300 K, 1 aum pressure, Required: (i) V to give 50% greater h, than conduction limit value. Gi) Similarly AT ="T,~T, for free convection, Also, corresponding results for water at 300 K. @ Air: for a conduction limit of Nup = 2.0, h, = 2k/D = (2)(0.0267)/(0.001) = 53.4 Wim? K. For 250% greater value, h, = 80.1 Wim? K. Now use CONY, item 7, and increase V until ic = 80.1 ‘W/m? K: the result is V = 0.091 ms. Water: For Niip = 2.0, , = (2(0.611)(0.001) = 1222 Wim? K. Fora 50% greater value, he = 1833 Win? K. Again using CONV, V = 0.00099 m/s. Gi) Air: Use CONY, item 11, to find T, to give h, = 80.1 Wim? K for T, = 300K and P= 1 atm. We obtain T, = 400K; AT=T,-T, = 100K. Water: Use CONY, item 11, to find T, to give h, = 1833 Win? K for T, = 300 K; the result is T, =300.56K; AT =T,-T, = 056K. Comments: 1, Notice that Nup for free convection in air at T, = 300K, 1 atm, with T, >Te, has a maximum value =2.69. The table below illustrates this point. AT, K 1000 Np Explain this behavior.Exercise 4-74 Given: A 60 W light bulb modeled as a 5 cm diameter sphere. Required: Fraction of lamp power lost by natural convection. Assumptions: 1. The flow around the bulb can be modeled by the flow around a S cm diameter sphere. 2. The bulb is a small gray body in large surroundings. ) Evaluate properties at the mean film temperature of (1/2)(408+293) ~ 350 K; B = 1/350, k= 0.0300 W/m K, v = 20.30x10" m/s, Pr= 0.69. The Rayleigh number is Rap = GrpPr= BATgD* Pr” = (1/350)(115\9.81X0.05)°(0.69/(20.3x10°)? = Since 10° < Rap < 10", equation (4.90) applies, . O.S89Rap : [+@.469/R)"75)° (1+(0.469/0.69)" Ihe = (K/D)Nup = (0.03/0.05)(15.0) = 9.00 Wim? K Qooae = H,ACT, ~ T,)= (9.00.05) (135 -20)=8.1W Qeoav/Qeice = 8.1/60 = 14% @ Qua =0eT}-TDA = (5.67x10 *)(0.8)(408%-2934)()(0.05)" = 7.2 W Qea/Qatee = 7.2160 = 12% 75X10 Nup = Comments: 1. The total power lost is 26% of the electrical power input. 2. Of course, Qaa Calculated above is the infrared radiative loss due to ‘emission by the glass bulb. The balance of daice (100-26 = 74%) is ‘transmitted through the glass as somewhat shorter wavelength visible radiation (Gee Section 6.2). 3. Use CONV to check hg.Exercise 4-75 Given: W.H. McAdams’ correlation for natural convection from horizontal cylinders Required: Comparison with Eq, (4.87) of the text. ‘The McAdams correlation is Nup =0.53Raf; Pr> 05, 10° 7-20 a 3 =65 [v0 | [cvv2.09%-2,090]c.ss6xi0%) Be = (K/L)Nuty = (0.0274/0.1)(10.9) = 2.99 W/m? K Q=h, AAT = (2.99)(0.1)(0.2)20) = 1.20 W Comments: 1. Use CONV, Item 1210 check B.. 2. Notice that the heat transfer coefficient is very small: if the surface has a bigh emittance, radiation heat transfer will be larger than the convective heat transfer.Exercise 4-80 Given: Steam supply pipe carrying saturated steam at 1.2 MPa in still ambient air. Required: Heat loss per meter length of pipe and the surface temperature of the insulation. Assumptions: 1, Negligible radiative heat loss-insulation perhaps contained in an aluminum foil sheath. T,=293K T, = 461 K (Table A.12a) Kx = 0.068 Wim K (Table A.3, 350K) LSem n=7Som inkey/t) 1 ake Darl, ‘The overall resistance to heat transfer is a _ eid , 1 55), 1 7142.12the UA ~ 2nk,L * Qrebh, ~ @n}(0.068)0) ” (2ny(0.075)(1)h, 1. = Qe ot (O.1Sx)h.(T.-Te) B.TT#212Mag ‘To solve for, we have to find h, which is dependent on T,. Guess T, = 310K. Solution using CONV ‘The required input is: ‘Configuration number = 10 (horizontal cylinder) Fluid = 21 (air) 1, =310 Te= 293 = 1.0133x10° D=015 ‘The output is: ih, =3.39 Win? K. Hence, aT) (461-293) at ~ GasmGTIz. I * 77 WAEMIETIG.IRI *777709K Further iteration is needed and we guess T,= 316 K. Using CONV, the output is i = 3.64 W/m? K; and T, =315.5 K. No further iteration is required. Hence Q = 38.6 Wim.Exercise 4-81 Given: A 20 cm square plate inclined at 45° in air at 1 atm and 300K. Required: Convective heat loss from each side when the plate is at 320 K. Evaluate air properties at the mean film temperature of 1/2(320+300) = 310 K: k = 0.0274 W/m K, v= 16.54x10° ms, Pr = 0.69, B= 1/310. The Rayleigh number is calculated using g cos @ where 0 is the inclination to the vertical, 3 520.20)? rypr= BATHOOS OL? py. US10)G20-300)0.81e0845° 0209 (969) - 9 3108 v (6.54x10 *Y Since 6 is not close to 90°, Eq. (4.85) applies to both sides, with Ray replaced by Ra,cos® Ra Nu=0.68+0.670Ra, cosy) 9767-165 - 0.492 = Hy |0492, veel 1.68 + 0.6701(9.03x10°K(cos45°)(1.402)!" = 37.3 9n6)-169 ] = 1402 (KILYNuy, = (0.0274/0,20)37.3) = 5.11 Wim? Q=bACT,-Te) = (5.1110.2)°(320-300) = 4.09 W Comments: 1. Unless the plate emittance is very small, radiation losses will also be significant. 2. Notice that it is incorrect to use Eas. (4.92) and (4.95) with g replaced by goos®.Exercise 4-82 Given: A sphere-cone at 340 K immersed in water at 300 K. Required: Rate of heat loss from sides. Assumptions: 1. An isothermal surface. 2. ‘The surrounding water is stagnant. pn ——4 7, =340K T. =300K Sem Evaluate properties at the mean film temperature of (1/2)(340-300) = 320 K. From Tables A.8, A.10b, k = 0.641 Wim K, p = 989 kg/m’, v = 0,59x10 m?/s, Pr= 3.80, B = 436.7x10° K™! Equation (4.91) applies where L isthe length of the boundary layer, L = (a)(0.01/4)+(0.0057-0.02)* = 0.00785+0.0206 = 0.0285 m 3 3 Ra, =Gr,Pr= EAE p= beam ea 3:80) _ 4 335107 | SARal = (0.52(4.33x10")" = 42.2 Nu k/L = (42.2)(0.641)/(0.0285) = 949 Wh? K Q= ACT.) = (949){(2n)(0.01)7+()(0.025)(0.0206)](40-300) = 85.3 W Nu Comments: 1. The required calculations can be simplified by us using configurations 8 or 9. 2. Lienhard’s formula, Eq, (4.91) is approximate, but nevertheless very useful. ‘The shapes encountered by the practicing engineer often do not correspond to the simple plates, cylinders and spheres studied by academics! CONV to calculate Ray,Exercise 4-83 Given: Tank of water heated by a steam coil, Required: Time to heat water from 300 K to 360 K, and amount of steam condensed: Assumptions: 1. ‘The thermal resistances of the tube wall and on the condensing side are negligible; thus T, * Tyg(1-1 atm) = T,,(1.114x10° Pa) = 375.2 K, 2. Lumped thermal capacity model for the water. 3. The convection can be modeled as natural convection on a horizontal cylinder. ‘The average bulk water temperature is T, = (1/2)(300+360) = 330 K. Then take T, = (12)(T,#T,) =(1/2X3304375.2) = 352 K. From Tables A.8 and A.10, water properties at 352 K are k = 0.670 W/m K, v = 0.38x10° m/s, Pr = 2.3, and B = 639x10° Kt 4 + 2 3 Ray = Grppr = BATED?Pr _ (639x10-9975.7-330\9.81}(0.01)°2.3) al (0.38x10 ©)? evaluated at an average T and AT for the heating process. Eq. (4.87) applies: Nu = 0.36+0.518(4.56x10 4/11 + (0.559/2.3)""" = 20.6 1h, = (DNup = (0.670/0.01)(20.6) = 1380 Wim? K If the tank is insulated and natural convection stirs the water well we can assume a lumped thermal capacity model to obtain the temperature response. From Eq, (1.38), T+, = 456x105 a, St = BY A A= Avy = MDL = (n)(0.01)(5)(0.4)(12) = 0.474 m?, For water at 330 K, p = 985 kg/m’, c = ANTS kg K. Hence, t, = (985)(4178)(0.08)/(1380)(0.474) = 503 s t= 804s = 13.4 min Table A.12a gives by, for steam at 1.1 atm = 1.114x10" Pa as 2.25x10° J/kg. The amount of steam condensed is given by a simple energy balance: why, = pVcAT. Substituting, w(2.25x108) = (985)(0.08)(4178)(360-300); w = 8.78 kg Comments: 1. Use CONV to check ih, 2. Hlow would you attempt to justify the use of a lumped thermal capacity model for the water?Exercise 4-84 Given: A 2.2 cm diameter power cable, resistance = 6.9x10° Q/m, with a 3 mm layer of electrical insulation, Required: Wire temperature for T, = 295 K when the current is 800 A. Assumptions: 1. ‘The cable temperature is uniform due to its high thermal conductivity. 2. Insulation k-value is 0.2 Wim K, and € = 0.85. 3. Surroundings are nearly black at 295 K. ‘The heat transfer per unit length is QL = PR’ a or = (800)°(6.9x10 *) In Ft.) Oak = 44.2 Win From the thermal circuit 1,-295 Q. eee a L Fl Indavity T x TWh) «BHO “WHoIHa a) +5 = 12 Wim? Since h, and h, depend on T,, an iterative solution is required. First guess h.+h, K, then Now use CONV to give a new estimate of h,; using item 10, h, = 6.64 Wim? K. also, b, « 4oeT,,' = (4)(5.67x10°)(0.85)((338+295)/2)* = 6.11 Wim? K. Hence h,+h, = 6.64+6.11 = 12.75 ‘W/m K, and a new estimate of T, is 334 K. A second iteration is hardly justified. Thus using Eq. (1), T,-295 42 = ist, O91 3012.75 42.9 K Comments: 1. Perform a second iteration to see how much the result changes.Exercise 4-85 Given: A.2 em diameter polished copper sphere, originally at 20°C, is suddenly immersed in air at 81°C, 1 atm, Required: ‘Time required to heat up to 80°C. Assumptions: 1. Radiation heat transfer negligible. 2. ‘Lamped thermal capacity model valid. ‘The low emittance of polished copper allows us to neglect radiation heat transfer. We will fist obtain an approximate answer by assuming h, constant at the average temperature difference. Using CONV, item 11, with T, = 1/2(80+20) = 50°C = 323.15 K, and T, = 354.15 K, givesh, = 10.07 Wim? K, From Tables Ala and A.le, p = 8933 ke/m* and c = 390 I/kg K for pure ‘copper. Assuming that the lumped thermal capacity model is valid, Eq. (1.38) applies, TT. Lo; Vln.A; VIA = RB TT, Lae 1. = (8933}(0.01/3)(390)/10.07 = 1153 s 740 s = 80 min fy (80-81)(20-81) =e“; solving, t = Now hy varies with T, and AT, both due to the AT in the Rayleigh number, and due to variable property effects. To obtain an improved solution we will use LUMP, and calculate the heat up in four stages. First we use CONV to generate a set of values of i, as follows: I 0 0 Vv 50 10 18 80 31 uw 6 1 10.07 845 7.685 6.005 ‘Then EAt = 5939 s (99.0 min) Comments: 1. Our approximate value is 80% of the improved value. The improved value is within a few percent of the exact answer, and is quite accurate enough for engineering purposes.Exercise 4-86 Given: A 12 om square vertical plate uniformly dissipating 4 W. Required: Maximum temperature in air at 300 K, 1 atm. Assumptions: 1. One side is perfectly insulated. 2. Radiative heat transfer is negligible. 3. A laminar boundary layer. First we calculate the temperature midway up the wall following the procedure of Sparrow and Gregg. Using CONV, item 8, we guess wall temperatures until a wall heat flux of q, = 4/(0.12)* = 277 Wim? is obtained. Inerpolating, T, = 346.4 K, h, = 2 TeR ATK wine x de Wie 335 35 5.598 196.0 340 40 5.169 230.8 345 45 5.923 266.5 350 50 6.062 303.1 963 Wim? K. (-T) = 2"8yy-T. = (2)"9(46.4) = 53.3 K Ty = 353.3 K (802°C) Comments: 1. ‘The Rayleigh number is 5.33x10° << 10°: the flow is certainly laminar as required by the theory. Vertical circuit boards are often exposed to quiescent air for cooling. Since the heat dissipated by components on the board is often distributed quite uniformly, ‘one sometimes sees the Sparrow and Gregg procedure used for such situations. However, such boards have relatively high emittances (desirable) and the radiation ony * Geos ANF at (TT. decay Will not be uniform even if q, is uniform. Indeed, if heat conduction along the board is negligible, T, -> T, as x -> 0 (bh, »»). Hence . Asa Joss is about equal to the convective heat loss. Since rus > 0, and increases up the board to be about equal t0 quegy at X= result, dueqy decreases by about 50% up the board. Such situations require advanced analysis,Exercise 4-87 Given: A 1 m high vertical flat plate at 320 K immersed in a fluid at 300 K. Required: Estimate of characteristic natural convection velocity for six fluids. Assumptions: 1. Viscous forces can be ignored in estimating the characteristic natural convection velocity. Equation (4.106) gives an estimate of the characteristic natural convection velocity by ignoring viscous forces. Fquating the gain in kinetic energy of a fluid element to the work done by the buoyancy forces gives v- (2 a)" = (PATEL)! Using the ideal gas law or Table A.10 to obtain B (evaluated at the mean film temperature of 310 ) gives the following results Fluid BK" V, emis @ Air at 1 atm 1/310 80 ii) Air at 10 Pa 1/310 80 Gil) Water 361.9x10% 20 (iv) SAE 50 oil 0.70x10% 37 (¥) Thermninol 60 0.70x10% 37 (vi) Mercury 0.18x107, 19 Comments: 1. Notice that the air pressure has no effect on V. However the thickness of the natural convection boundary layer increases with decreasing pressure (8 « p™, see Eq. (5.79). 2. Gases have relatively high natural convection velocities, liquid metals have relatively low values. 3. These values of V are a guide to assessing whether natural convection or forced convection dominates in a particular situation.Exercise 4 - 88 Given: A vertical wall maintained at 320 K in a fluid at 300 K. Required: Location of transition from a laminar to turbulent boundary layer. Assumptions: 1, Transition occurs at a Rayleigh number of 10°. From Table 4.3 the Rayleigh number is ae BATgL? BATgL* ‘vo. ve If transition takes place at Ra = 10°, the location of transition is X,, = (10°v'/BATgPr) Using Tables A.7, A.8, A.10 and the ideal gas law, the following results are obtained. Fluid vy, ms, 8,K? Pr Xm @Airatiam —16.54xi0® 1/310 069 (0.856 Gi) Air at 10° Pa 16.76x10% 1/310 0.69 18.6 Gi) Water 0.70x10% 361910 46 ond Giv) SAE SO oil 25x10 0.70x10° 3400 0.558 (¥) Therminol S.Ax10% —0.70x10% 65 0.143 (vi) Mercury 0.107x10% —0.18x10% 0.024 0.238 Comments: 1, Notice that x, is very large for gases at low pressures. Because x, ev’, and = wp & Pl, x, o P™, 2. Notice that x, is relatively small for water-we often encounter 2 turbulent natural convection boundary layer when using water as a heat transfer fluid.Exercise 4-89 Given: Immersion heater 14 cm high, 3 cm diameter, used to heat a bath of oil from 20°C to 60". Required: Maximum power if the heater temperature should not exceed 150°C. ‘Assumptions: 1, Uniform heat flux along heater. 2. Curvature effects negligible-can use vertical plate natural convection contelation. ‘The procedure recommended by Sparrow and Gregg can be used for this uniforraly heated surface (see Exercise 4-86). We first note that the temperature half way up the heater is obtained as. (Typ-T,) = 20, -T,) = 2-"(150-60) = 78.3 K Typ = 78.3+60 = 138.3°C = 411.5 K (Worst case - when pool has heated up to 60°C.) ae We now find ih, for a vertical plate at a uniform temperature of 411.5 K. Using CONV, item 8, > ul a h, = 113.9 Wim? K. Q = BAC )-T,) = (113.9)(0.14)()(0.03)(138.3-60) = 118 WExercise 4.90 Given: Steel cylinders 1 m long, 10 cm diameter to be cooled from 200°C to 60°C in a 20°C oil bath. Required: (i) Configuration for most rapid cooling. i) Estimate of cooling time. Assumptions: 1, Negligible forced convection effects due to immersion process. 2. Lumped thermal capacity model valid. 3. The vertical cylinder can be modeled as a flat plate (D >> boundary layer thickness). (@ We need to compare average heat transfer coefficients for the vertical and horizontal configurations. Since h, varies due to the variation in AT and in fluid properties, we will evaluate hh, at the average cylinder temperature of (1/2)(200+60) = 130°C. Using CONV, items 9 and 10 gives h, = 125.3 Wim? K for a vertical cylinder, and h, = 87.8 Wim? K for a horizontal cylinder. Thus the cylinders will cool more quickly in the vertical position (see Comment No. 1). (ii) From Table A.1 take p = 7840 kg/m®, k = 42 Wim K, c = 500 J/kg K for AISI 4130 steel. ‘The Biot number based on our estimate of h, = 125.3 Wim K, and V/A = DL(4L+2D) is a WA) (125.3)(0.1)(1.0)(4x1 +2x0.1) E B Bi 0.071 < 0.1 and thus use of a lumped thermal capacity model is valid. For an approximate answer we can assume h, constant; then Eq. (1.38) applies with = p(V/A)cih, = (7840)(2.38x10 )(500)/125.3 = 745 s 60-20 0.222 =e; t= 1121s (18.7 min) Comments: 1. ‘The fact that h, is higher for the vertical configuration perhaps goes against our intuition—usually h, decreases with increasing boundary layer length (see Eq. (4.91)). Look carefully at the CONV output to assist your understanding of this, result.Exercise 4.91 Given: A 6 mm O.D. insulated tube losing heat by natural convection, Required: ‘The critical radius of insulation and corresponding heat loss. Assumptions: 1, One-dimensional steady conduction. 2. Negligible radiation heat loss. 3. Equation (4.87) can be used to give h, y Inte, /r,) 1 2mLk 2Trrlhe ‘From the thermal circuit, T-T, | TT, ing Jr) T mEk —-2nLr,h, where hi, = h,(I,,) is obtained from Eq. (4.87), 0.518Ra,* Teta sr Nig tala esses ccs Goer aes [1+(0.559/P yy? Substituting in Eq. (1), @ _ 350-T, _ T,-300 TL InG,/0.003) T CmyO0s) = -2nrh 350-T, __ryln(r,/0.003)h, a 0.08 or (QUL) = 6.283r,b,(T,-300) a Q) Ga) (3b) CONY, item 10, will be used to generate a table of h(rq:T,). A number of r, values are chosen and T, varied. Then Eq. (3a) is solved by iteration for T,, and Eq. (3b) used to give Q/L. From4-91 (continued) a plot of Q/L versus r, Q.,,,/L is obtained. The result is F, = 0.00525 m, T, = 338.0 K, h, = 8.73 Wim? K, Qua/L = 10.94 Wim. Comments: 1. The values of hy, calculated by CONV are listed below. Dom hy, Win? at T, K 336 338 340 342 344 0.009 9099.19 9.29 9.38 9.47 0.0095 8929.02, 912.921 9.30 0010 877 887 897 906 9.14 00105 863° 873882891 (9.00 0oll 850 860869878886 2. The problem is easily solved using CONV: writing a special computer program to effect the solution is not necessary (unless many similar problems require solution). 3. The result for Q,ay/L obtained in Exercise 2-30 is 10% higher than the more accurate value obtained here.Exercise 4-92 Given: A 1 KW cylindrical electric heater, 2 cm O.D. and 30 cm long. Required: (i) Surface temperature when horizontal in an unstirred water tank at 330'K. ii) Surface temperature in air at 310 K for a surface emittance of 0.8. Assumptions: 1, Circumferential conduction in the metal sheath of the heater results in a nearly isothermal surface. 2. In part Gi) the tank walls remain at 330 K for the duration of the transient. 3. Negligible end effects. (i) An energy balance on the heater gives Q = Quigys hence Q=hAC,-T,); T, = T,+QhA T, 330 + 1000/h,(m)(0.02)(0.30) = 330+5.305x104 Q) ‘Since Ih, depends on T, an iterative solution is required. A value of T, is guessed, f, calculated using CONV, item 10, and the result substituted in Eq. (1). After a few guesses we find T, = 375.6 K and h, = 1163 Wim? K satisfy Eq. (1). (ii) The energy balance is now Q = Quony* Quai Q = bAT-T)-oeacrs 1) (1000)/(7)(0.02)(0.3) = h,(T,-310) +(5.67x10 *)(0.8)(T. 4-330) $.305x10* = b,(T,-310) +4.53x10 *(T4-1.186x10 ) Q ‘Again an iterative solution is required. Using CONV, T, = 1002 K and h, = 11.4 Wim? K satisfy Bq. 2). Comments: 1. The wiring inside the heater may bum out before the surface reaches 1002 K. 2. Notice that, in air, the heat loss is dominated by radiation: an accurate value off, is not required, Note that T, > Tgp in part (i). After you have studied Chapter 7, you can check to see if bubbles of vapor form on the heater surface.Exercise 4-93 Given: Coil of 8 mm O.D., 0.75 mm wall thickness copper, mean coil diameter 0.4 m, in a water tank maintained at 280 K. Required: Length of tube required to cool 0.15 kg/s of a test liquid from 360 K to 285 K. Assumptions: 1. Test Tiquid properties can be approximated by those for water. 2. Convection from the coil into the water can be modeled as natural convection on a single horizontal cylinder. Copper tube thermal resistance is negligible. 4, The overall heat transfer coefficient is constant along the tube. Since the water in the tank is maintained at a constant temperature, the system can be modeled as a single stream heat exchanger, if we assume that the overall heat transfer coefficient is constant along the tube, From Eq. (8.17) Carine, 3607285. 73. 9.9375 TT, 360-280 80 L = (2:773)(0.15)(1005) _ 1.663x104 a ((0.008)0 u To calculate O we will neglect the copper tube thermal resistance. Since U above is based on outside area, ‘Both h,,, and h.; vary along the tube (especially h,, because h.,, depends on (T,-T,)"* in addition to property variations for natural convection). We will evaluate the assumed constant U at the4-93 (continued) average bulk temperature of the test liquid, T,, (1/2)(360+285) = 322.5 K and guess a wall temperature of 320 K. Then CONY, items 2 and 10 give h,; = 2.53x10' W/m? K, h,,, = 944 Wim? K. A check of the thermal circuit shows that the corresponding wall temperature is 320.6 K, so our guessed value of 320 K is adequate. ited ease 1 ta; (0 = 903 Wim? K 5 2.53x104* 944 L = 1.663x104/903 = 18.42 m ‘The number of tums is L/D = 18.42/(R)(0.4) = 14.7 ‘Comments: 1. ‘The inside convection resistance is much smaller than the outside convection resistance-forced versus natural convection, 2. Check that the thermal resistance of the tube is indeed negligible. Instead of Eq. (8.17) for a single stream heat exchanger, we could use Eq, (4.11) with h, replaced by UExercise 4-94 Given: A 20 om square hot plate maintained at 150°C, in ambient air at 24°C, 1 atm. Required: Convective and radiative heat losses from its upper surface if € = 0.8 Assumptions: 1. The ambient ait is stationary. 2. Radiation heat transfer can be modeled as a small gray surface in large, nearly black surroundings, also at 24°C. For this natural convection situation we first calculate the Rayleigh number, Ra, = BATgL?Pr/\?. Properties are evaluated at the mean film temperature of (1/2)(150+24) = 87°C = 360 K. Using Table A.7, Pr = 0.69, v = 21.30x10% mis, k = 0.0306 Wim K. With B = L/T = 1/360, Ra, = (1/360)(150-24)(9.81)(0.2)°(21.30x10 “)? = 6.05x107 ‘Equation (4.96) applies for Ra, > 2x10’, 18 Nu, = 0.14Ra)” = (0.14)(6.05x107)!? = 54.9 1h, = KNu/L = (0.0306)(54.9¥(0.2) = 8.41 Wim? Qeone = RACT,-T) = (8.41)(0.27(150-24) = 42.4 W For the radiation heat loss we assume a small gray body in Jarge, nearly black, surroundings also at 24°C; then Eq. (1.18) applies, Quy = OAT TS = (0.8)(5.67x10 *)(0.2)°(423*-2974) = 440 W Comments: 1. ‘The convective and radiative heat losses are almost equal. 2. Check ih, using CONV.