Sizing Piping for Process Plants Good design practice for sizing pipe for process plants requires a working knowledge of fluid dynamics. Here are techniques for solving most sizina problems encountered in the piping that interconnects the process equipment. Lar SUFSON. non Corie Corp upstream pressure and a low downstream pre eter sag methods ae of ret enone impor entimiaton ie ene he dsiher tance te the chemical voce industries becaae ie SH Drone the simaest pipe size tat will ve tia of pica nag Tops 10 fp 0s ofthe rederaod peu dup antes ee Reiner eras fovrecver the enerey through a trbine for use Althogh modern technology has replaced many sae B® a the pan, lng rules-t thom within the past 20 yeara the de: PUMDING of compression, i needed to boast situer & prseat mt ave naa cd ee cetnecey pressure, complicates the sing optimisation Fa Sizing methods. Nota, eonon andthe deterioration X61 mate eae the invertment in pip ad ‘of fltion factor both ncnatate the une of artitrary teases with Increasing ppe alae, wheres the fining techniqss, Hopetally, the economic incerive <2SPEY cot decreaees with ocrasng la. Aa 8 wil stimulate future technol Improvements, tHe Wile ene minimam-caet or eptimim ioe Tn addition aia eect on initial piping Investnet,Atalaes of this tape ave be made beth graph Pipeline aisng infuencs a plan's operating and (PI 2) and algebraic. The minimum cot | icrtapciot, With tworgnace (eediqhd) few goor tah option velsty Wat i largely a function Geri erection Con thoes te et we Sent Asi enity in turbulent fo. Tete T pests line and ran in a rvistn expe pus acatiy doe of optima vlog, for fuids In tarulet | Other pretice may afc ste mare ably. Forex: #00 £0 sours | amples ifthe pipeline careying «controled fw i xT 18 apparent from Table I that econamle w Sets ised he eee ee a SE cape on fastre ther Hn id. Senay SsGeiga inter rana oneg those are pipe eats, energy vat, and the F rt fittings and vavea por anit length of pipe Me eaten ‘Uist, optimiation anaes of pip lta encampans ore of each anit proces then has A typiea sising problem je a proce optimization comsldered 20 far. Inclusion of wore of the fl In which the designer i earehing fora line ane that lements into an optimization would nmprore Wil provid the masmim economic gain fr hs eom- txhnlquea All ofthese clements bnve costs pany. If pipn Is to e Snetalled between 4 fsed high end on pipe size 1. Cost of inalation and heat lose could be to the piping investment and energy cont Th ‘vould bo an optimum inaulation thikoe fe 8 Die size, and an overall optimom pipe site fptimum size ould be smaller then tht ob Sngultion cota ware igcred 2 Contro-vale prossore drop, if neaded sualyal, would tend to increase! the rensired piping Process equipment (og, hoat exchange?) ttlen with plying for a Geed presnire drop Important If piping and equipment coe were ‘ne order of magnitude The designer frequently cannot sue seonoals cause of arlouecunstalnt Wt mist beeluded anger) ‘These constraint, which provide the frame fr most ne sag, can be cined in the fo. ar + EROSION LIMITATIONS TPRoCESS CONTROL, PRESSURE DROP Scompressinte FLOW ‘two-pitase FLOW IFLOW DISTRIBUTION Limitations wh hgh veloiin aometines cause pacivity ng matetalt the engineer is more frequetly swith velocity induced ecoion High veloc shorten the potenti! life of « metal pipe, king the Sostlaton soynomially weatleae Hor example, em eopper nickel alloys Ia 20 rcvice have acceptable eraion rates when the iB to 10 fee; however, when eboity ex thoat 16 fe/4ec, the eacion vate hen tacos tie teigner should be aware of the following con: hat eau Ineeuse ereslon and Impoee velocity on erosion problms. Production engineers are ect eciee af erion know-how. When the 4 has no Nstorial daa or constants avalabe ‘rosion, ho ehosid srhitreiiy lit the ‘eoety. Par liguds, a maximum velty of fs recommended Control contol of fd ie coentel to many piping “Athan sning of contol componente shes to those experienced In thie fl the engines the piping shoud familrize himelf with apace of procses conta glare goad fw eootrel i a pipeline the control st dissipate considerable pressure drop. AS a ieige? he control valve shonld tke 2 lost Ue total freon prenure drop inthe com system. (Ste the accompanying formation ‘on page 196 for-a totter cotinats of the ‘eop Perentage) The reason ie apparent. If vey tle pressure drop in taken across 4 valve, tha valve atom will havo to move a large ge of ie sealable stoke to change the fos rolling at lower fowrates, the cantel valve then be nearly loved, and small change te ive opening. would effect large change fo ol valves can be roughly ined with the f Einsrng—Jone 17,1968 ‘CONTROL VALVE ispetes much pressure dop—Fia 2 (Optimum velocity for turbulent ow in stel pipe—Table| ‘aid Denity optimum Yloiy’ Optimum aay thou re resee "ese heads, Por turbulent incompressie oof nonib ing aids, the contolvatve.peccure drop (heaa toa Example: ControlVaive Size ‘The following smplifed example ‘iustratse the above poeple, Weter et 80 pl dows through Ube heat exchanger chown in Fig. 2. he host exchanger Aluspates a presure drop of 12 psi when the Aow- fate 100 gpm ‘A Zin, pipeline produces sn estinsted 1¢pH,tle- 2 pel. A Zin ine le too sual! beau the avalable resare drop forth entrl valve on (100) (8012 14)/80 ~ 13% of the tla syatem deop, and therefore the control valve would aot have a useable range. ‘The tse of Bin. piping le dscunted becatee the plant does not inventory thie ena Tho Sin. line makes2 i ‘ails Jet an iE + = [al ee S i Oe — rae, : Cintas eomonis i ee saw Aeneid thy. Jean ee vam anya Repeat, - are aa Rete sips te van Mi aay aoe ee perce yokes eam FANRUING ition actor fo pine flow sons tof spe roughne—Fa 3 (90-12-2) = 16 pa. In useful enginering unit, Ba @) i Subetituting the appropriate value, the velocity be ¥ (9280 16,0) 6238) ~ ws ne ‘Converting the velocity oa correpending pipe lam. ter, we perform the flowing calculate: = oan. din? Danese = fl) 0) aa = 8 ‘The contel valve mould probably be 18 c 2 in, de- pending onthe osteome of& more compete anys by Instrument specaiste, Spesfeally, a itter estimate le neded ofthe umber of velocity headaK. a8 Pressure Drop: Incompressible Flids Before more-complex phenomena are cor I ble review the conventional cor Tle for prenare drop estimation nth Ne compressible single-phase Aan. "The sted Tosepie mechanial eneray balan. (cam Bernoultequatio?) ie useful with such AP 4 AO! 4 AE + gm ae where AP/p is the presaar change, (Vi Kinet energy change, Zp /2e 8 te dpraieh sind fs the Read oe, allIn feb "Ba. (2) aesumas that: (1) an incor fe foming, (2) no work is done on BP (3) head ows Uathermally, and (4) Sm 20 17, 1948—Ch ca} es iy be the re iP ote th red rl fi aw ‘he of Ma ineis ak The rate of irreversible conversion of Ppscanicl to thermal energy, aio ealed fection ls, fe sor lat wrk, i dencied B: This bead los wil feral positive. An exception to this le can ocrar Be Pe the sence t's tow change that Se inherently et 3b, [MiiMint bat arosay alters the wlosity prose, sheredy £601 IM io:ing Ascumption 4 stove, Such lathe cate with « foes pipe-ize expansion were sie change is al? INete that Bq. (2) clearly datinguishes betwen ire drop abd fiction nee. In contrast to rie feet encrer, pressure shergy cat ether Increase of Mcrcose trom sleation. and weloelty changes Hend los vil exe! preesure op only with constan-diam- per, coseant-slevaton piping. Te nw used to predict ~aP ot tn laminar pipe ior was evolved experimently by Poosalle® Today sth chemin! engtneering terse, his lw in atted intems of « Fanning trietion factor f and Repos faunber Ns, both of which a7e dimensocins (The Bayside number ls the ratio of ner to vious ices) 1 182% amine ow) w 2) t#( 109, + a twat cao Fo “mee sible oe ewe nnn onic epinering—Jne 17,1968 Flow, Oa Ba, (4) and (5) daiae Np, and f respectively; both ‘tuations are correct fr turbulent aswell ay lama ow. Combination of Ba. (8), (4) and (5) eves the Poel equation for lsminar pie few: ‘Laminar fow is encountered tee frequent than urbu Jent ae, where the reintonatipetwasn fiion facto snd Reyaulde numbers ore comple, "The cnet of turbulence in pipe flow ie now well Aerstood. A recent thocetic! atu" indian thatthe tritieal of tanational epmole nomber ls dependent fm the ampitade of te disturbance (es faded by Foughnoss or Stings). Practically, ths rita! Rey ols numbor Is often akan &o be 3100, Experizental ‘Work has ahown that cw wl ec ayiar Below Ni 2000, even Inthe presen of large leturbsncer, Above oughly Nay = 4000, ie ow il ike te tubule, although Iain ow haa been noted at » Rayo ‘omber of 40,000 nn setem of exceptionally ow dis tarbancs, n'a later section polymer ation to ly lds abn shown to netard tee transition fo turbulent fom. Tor tarbulent tow, Moadye feiton-astor graph i widely usd in enginering today. Ta graph has ben eplotied” 20a Fanning fiction fst ands shows in Fig. 3. Tho perametar .D on Hig. 3 le telative rough © PRessUne DROP for water in turbulent flow trough Schedule 40 sel pipe—Fia.Nomenclature ‘gle zaman np Senin na), dennis a A Qo, Sitimum Sow of water at ¢8 F. theovgh suet pesure aop of te pais ene Teteoal pie dems, Tiere! pipe dame’ rite foe awd em 18 Mh ‘rae names, cna ‘Fanning fh tat, dimcnlolse Bat wl, Ih oe eat) Seoveon ak S217 a pop ay Ble iee Bidea et eta te, 2 Maniac ty (raiment Lanethotyine fe ea Ee RT cma ow, dimensioniese a [nn ‘Brave aa. Gr hi a>) Acton "tamale "Bowral, settle ‘il tan owe gem ‘tan auc fora ‘Unter! pas omeant = 2 eth (hy Ce ai of ip eng, yaaa amber, dnsesn racion ope torsion ecupled by gas Eitstion Gf pipe cae sition: ceo op kemnmeessctzmas 9 eee os Pressure Drop Generally a ese Sof the ttl tictoal pre see Ric il tte a ber etme af ts pcre 8 Contra ates have a eed, aie oti range {BRE cried yo rnguatity keene eS Helkook? thi rangeabity ver oe fet Meal ‘ves 18 for larger valven with en tresses ene apgeniiy it related to preeare drop and ow rate By the lowing onto here the ssberpte min, and. mac, ref to be n'a tyelom Jn which tho’ upsteamn and down. ream pracrer are eed (ewe Fg Dh deca ‘op for in valve af inti fo she es teat af maximum fw Pace pine te prcoine Sane {om dae tthe Tower fowrata te Rage {em The spate praaure drop Pr Gpsing. Rees snd ea changer ba exeading a ane args & {hat is ten throogh the sortsl valve Hescpetyae ‘imate ie foo phase doving i eee Marin paramcter ded by a, (18) Frosueduycoaaenin £ Forel il cod tea Subscripts Refers wo condition psteeam rservle Beles tn deeoooeas ana lcs pecure op fo 200 ta ia Mente fe thermal compre ow Seller Noi elice é ; as ale in tle Ia in Contr Vatves be assumed fo vany wih hw square of the Sheela nay i Pe the tston of he tata pe sop that i taken vag the salon a few, Ba. (A) ‘and (8) combine ives P= Golfeg =1 ‘An exam wil toot the ote of Ba (G oe 16 em uae 2 90 gs ad {Eventi he ta een dep fb 2/007 = 1 _ og ‘Thorsone, 185 of the total teen for (cing control ave) ast be keh etcntal ae” Tigh Reynols numbers, the curves in Fi naspmptotieteltion fstr. Thi fleton factor sleted by the von Karman eguaton:® roughnew + is the prteasion helt of telly ‘shad that ig glued to the wall of pipes i the rents of J. Nikaradses The valves in FIC. 3 the aut hydrate performance a¢ androish pipe et high Reyoots numbers. the sranitlon region, represented by che carved ‘ot each line, the fltion factor devsts from Miradve curves. Commercial pipe causes a mre by tanaition to folly developed turbulence than sand rourhenod pipe ighest cannot be maeaured atvetiy st present tips measured hydeutisuly ae sand Poutnes, pesmi insurmountable cistrction to direct, ements the efec an inne varable rough omelzy has on pipedoundary layers, An exa ft tle diticlty ie presented str in Table Th. The fn ocior for flexible metal hose Increases with futher Bete Na =10" and Nau = 20 Mphavior cannat be predicted by the curves fs mates ot fsetion facto oF band healoas from S tho be within 10%" for new pine. Asie fan unpredictable fost on fiction. Usually, ero Corvation and sols deposition will increas by Toughness snd the fietion factor. One refel~ upstate that afer 10 ye, water flowing through il incrence the friction factor to 22 times Its ale in bin throug 10in pape: and to 1.6 times alse In 22m. through 6Oin. pipe, Por the Syplel, lestcossoaie environments the dealener fee 2 20 to 0% anfey factor for elngleghase e-rop estimates. = bg Wi +238 » on os sisting — Joe 17,1968 ‘Sizing Shorteuts -requntly, the dosigner must eae a particule pigee tne before if hasbeen rosted. As 2 reul,he ony has trough sen of the total ine length and a specu {ate cout of the fittings Spied methods ae aaile {Die for thissvatin aud lao forthe one tht equies Srongh estimate to liminatetraband-orror eaelae fiona of velacty or diameter “simplified rising graph canbe constructed for lige ide. Shown sa Fig this graph permits an estimate of preagre drop, welacity, Noweste or diameter, If two Ur ene four entity are known. Fig. ¢ aa plated from the flowing torbaten-flaw equation: f= oostnie® © ‘eens from Fig. are nota acurat a that fom rip’ Hoover, the eimplied graph wil be adeauste for many icing peoblema Loving commercial see fr iron pive and turbulent ow. A teloitie Tes thn 1 fancy eles from Pig. Wil bo within 10% of Moody's frien correlation. Pig was constracted for water at 202.6. (684 P) Peanare drop fa corrected for fuld density and vs: foslty with the folowing relationship, derived om Eq, (8) ad potted on Fig. 5 (2)-(2)__(ewtor)"(as)" ‘De chart prepared by Crane provides «convenient sising technique for vapor ings Reprodnced 28 Fig 6, [Wie baeed on the Bighly torbulnt region decribed by Bq (7); rand roughes «i 0.00015 fA simple equs- tion erties hoses o peacae drop in pa 100 = 9) ‘To estimate peesute drop, the constants Cy and Care he ON for pressure drop of Fig due to viscosity ond éensty—Fia. 5 7Resistance coeticints fr valves ad fitings—Table I +4 sana] [+E sere Note K rene th nating al lea oF be ne 2 ee a ok Sd i ns Ne a ye "Hes L fe 8 4 & ie ss ih 4 a ae e ‘ee=r aa ~ sian Resistance coeficients for valves and fitings—Table “yet neers ameat t g eet) ty va 0820 0280 Fring Wien fr ole maa hos 4900 0.187 w/e 1500 0.200 2 2000 ou 4 4000 0.081 titi savant, 9r5* OCS Sinan tome spe 026 tig eg bn ip say ey seat Spe ve lenirag tone 17,948 ro ‘nicl: t i i i A the valves and stings with straight pine for tat onal purposes. ‘The velocity head concept, re, permits a move direst estimate of presse {ase Ba. (1). sah the number of veleclty Reade of lo ld be eonatant for fittings that are geometrically ine, the A of commercial fines (as shown in TD seems to derense with nerening size Patt Ye varition Is due to lack of geumetricsclecp; Te canted by aualvee that ignore «possible Rey: umber dependence, ssre-drop estimates with ether oquvaientJength ity bead are leas acerate then that for teaght ubslations similar to thoee in Table Il cat be tod to predict ead los to raaghly 25% fe: Pump-Sucton Piping este piping in Fig. 7 eta be alaed ao that the the pump euetian set least 104 pigs at fot 100 gpm. Pid vlausty and density ate 03 ep, ob Ibyeu fe seapectiey ing a, (2), we fat ebabloh the head oe avail in term ofthe veeity Y tie k= wo) ue) Ba 85 i025) ct locity of Row, V, depends on pipe ise, we alcaate am approximate bead los tha le om tos prowure: AP" (58) (0) (18) = 2.98 pt 1 Fig 4, 8 2m. pipe sppesre marginal without lowarce for fitings Estimated AP” for in. isle a ne with total straight pipe length of 's conputed from data In Fig. a5. The prev drop from Wig. 4 is corvected for donalty ad ty, end then for seta! ie length (088) ~ 0.728 yt 10 fr he AP" (078) (4/00 = 033 locity head i let in one contraction, Smo lly nega valve The Hern eet wil be sje the Telowing Attog let: 1 Obstruction 2 OE regular danged (ow of 100 gpm ie equ to 4.000 Ihr.) For 8 Sin tne, Ba, C10) lve ‘he pipe las estimated from Big. 8 e025 pe ‘Thocecaculatine sDuetrata th tlative inportance of valve and Sting oases They generate over 40% ef the total head las this exer WHA Piping fn a ‘more congested are ling lores wl cote larger $Eracion ofthe taal os, Compressible Flow [aing eat indatia lines, «gato vapor canbe treated at Icompretble if guitable average debates are deed, Occasional the lie pressure drop will ex eed roughly 40% o€ the abeolue uptreae pressure fd denityverlatione must be consieted. Ove etch Instance acears whm prasture ig reieve through a rupture dls Compesrbiefow techniques tout be "aed when the banetng pease exceeds 10 ele ‘To eure compresble-fow problems, Ea’ (2) and (8). the Bernoull snd. bealse equations, are ey pressed in diferntal form. Tho renting eqation fltlne a presere-denaty tat eae ‘This tarm must be orate forthe particular type of expansion that oeeurs inthe pipe Fer dneomoreeile flow, the Integral comes aP/P;, whereas, for lat "erinal expansion of en ideal gas it Booimes (RT/M) In(PaP) When tae integration is compat for ise- ‘hrm ‘xpancion ofan ideal ea, the Tchanicl-oe ray balance becomes Pen ree an where for sralght pipe N = K = 4{LyD. The equ ton enumes: the vecityprofle is Mat; the fletlon {cori constant ove tho pipe length: eleration eects ste nglgibey and the ieal gael le vl ‘upstream and dbwneteeam prenures are Known the designer ean eae Ba. (12) dre for masa vel: city. He can then dstermine a in tian, sine @ W/A. The moro dial task of ealculating a prenare ‘with @ known involve ralend-etror technigue 1 the downstveam prestare Py in Eq. 02) it ity trom Fig. ls roughly 4. owe, a0 that or valves and tings is es (24 68 = 64 aes "a8 ™ tal head lone in (88 + 0.28) = O47 pel, < 295, Therefor, a Sn. ine adequate eof Fig i alo ostated with thi exampleValues of i i * i i [CONSTANTS rom chart ana tbe, sada E10), anele api computation of pressure crop im steel pipe ia 6 found trom Fig. 6. Ba. (10) is then nad to ealeuete (CaP) When the rato of preemore drop to sbaclate upstream pressure —-aP/P, = 04, uae upstream of Abwnstzeam density, Whon Di < ~aP/P, < 04, use fan average density, In slater section, we wil srt fampresibe Aow in deta ‘ig. 6 and Bq. (10) can be ued for Ugalde, In fast, such an etimate may be elightlybetar than Fig for Tow viseastes and high veloc | ems = SB @ f i ‘ao we HE “2 a ie rat Valves and Fittings ead neon in valves and Stig have hi bean correlated both ava number of velo tnd as equivalent length of pipe Zn. Tho 19 Inte bythe eavation Ka dy/D ‘The equivalent-ength concept Js «usefulwe rice, @/E a tion of mas wlosy above shed ne Poco: 8 iced while the upetream pressure P is held constant fhe mase velosty @ wil inereme fo 2 maximum value ‘Rurtherredstion in Ps wil not increase he mase-Alow fate, This maximum velocity equals the tone velocity fer adiabatic tow of a hypotbetiea gas whose heat Capacity ratio y = 1, Very roughly, soni velocity Is feneraied in. fay frictionless converging geomet" (Gre, a Beane) Where the absolnte dawnstream pres ive ie Tes than 80% of the absolute upetream pres En sure, For frietiongonerating proeses, the percentage trl be fess than 80. ‘Practically, the isotbermal-ow ansamon i athep oor, presley Sn short ines. pansion ae hag {be mre neatly ediabats, ace Httle beat can Be ttansforred in the short prlod required for exagion the gas Ta long ines (N > 100), adiabatic nnd lage ‘thermal assumptions ive nearly ibe same Novrate fetimates, The isoermalaasureptin So ql vee Secauee permit a roatvaly ice presere rep ea ‘ate (Ha. (12) ] and Deeaue crore are peer lag ‘han 10% in even rentvelyfietionlese pressing (w= 02) "To elve comprestbiedow problema, Lappe ame structed charts that bave found wie enginering age ‘eptance, The chars reproduced os Fig. 8, permit de ect estnation of aay vlolty through 2 pple if ‘sn pstream reserve presure and a downstea pre Sve are know, Pipe sie ean, therefore, a0 be ete tated directly for this ese. I one wishes to climate presure dsep in piping, however, the chate still ie {hie tela and-eror promsdure Moyt of the nomenclature for Lappe chara iven in Fg. which aio lluatrates the bis fr deulatlone, The parameter Vin Fig. 8 ls ‘eoci-head count hat can be ned interange drth the standard namber of Welly heads the lo is used Ly a redaction in fow aes. If flow are ie reduoed nar the pipe ext, so vl fan be genraied ia the reduced aren As rou ‘owrate can be mush lees than cotiated with charts apile's charts are based on the follwing ssimpting’ 1) the gas cbape the iden) eat () the pie a oraonal and (8) the Esti Is constant over the pipe Toth ‘Assumption (3) is general a goed one, snc rayne number wil bavally he sesciated With fone velotise, By (7) (high-bolene 1lon i {or} can be uead to entimte friction foto aloe of = 0.0085 ean Be asemed and then When 2 Reyna number te know "Thy eritcl mts eloety GI altained st Pal fora frettonletlethermalcntrsctin from ha recruit Pint. Tt sn elated to the presse epaty a¢ Point 0 a8 flow ‘whore ¢ = base ofthe natural logarithms Example: Lappe Charts A liqud-phsee reactor Isto be inetaled with tore disk that is to prevent the restor presate {um vaporization rate a sctimatd to be 1500 ‘Alo the fllowing information ls known 1 Bettve resistence of the venting ol celimated to be N= 2 2 Maxtavem density ofthe vapor ep = 94 2, Vapor ie eating tothe atmosphere 4 Ment eapalty eat y = 2 one 17, 19482 147/647 = 0227, From Fig. 8 Pa/Ps = 2u, = 05 EATERS) ~210 one) (og 10) = 16 ae) [ER ee ai oot jp vent line and rupture dik are roqire. 350 Vapor-Liquid Flow naderstanding of two-phase Aow is necessary for piping design. Bacensaly, every chemical proctes forse twosbase flow problams. A smeanue of the importance and Ure compleity of ths area ty Govte’s tibllography"® of the two-phase Merstare that a2 of 1966, inlided over 5,000 in the proceasdsign stage, the enginter tempt ta nderarand the Ao patterns present two-phase pipeline. Particularly In vertical pipe, ier alone often geterminos the prope line ie Trneral Row pattern maze, Balers” le recom oe borzonat flow, and Govier’ i suggeted (el upto. Cons the orinal ates for de ender shouldbe aomowhat wary of the vallity erm prodctd by any general fw-pattern mp. funky that this approach to two-phnce flow wil bs wholly euecestul. Fores that control one 9a sy bo tningortant when a diferent pstterh aap, ow of lange vapor babes ee! om gravitational free, and is independent of forse, Conversely, ow of aml vapor bubbles vom rurface forces but not on gravitational Important aspect of vertical low le that aso with large vapor babble, Termed ale ow, es ‘ astocited with reseore pulation and pipe on. When bubbles an pipes are larger than 150 fod the liguldviscosty Is leos than perhaps 100 "Froude numbers, Thess re related to the Ia: ‘oupraitational force ratio an are dated ‘The velocities Ve and Vi are superteal velctis nage on the total ip eross-cection. Ta the range 0 < (Nrden (Mie < & these Froude numBere exit rang Interesting charactors Downflow in Vertical Piping Consider the pipe shown in Fig. 10 to be intially comply I water i trned and fowrate le gradually Inereased, a falling fm il form on the pipe wal ‘When the fowate Se neeased to roughly (Mr = 2 the pipe will Hood snd the total erearsetlon Will Be swat fled "E thosame procedure ia followed when, asin Fig. 10, 1 liguid i used fo seal the boom of the pipe, lane Tubbios wil tend to be teapped in the pipe below Gp = 031, Abore (Nu = 034, the bubbles wil Teewent downward and onto the ine The author ran the experimente descrlbed above with ¢ 8. cection of 18/164n. LD. pipe. Bach valve (Nha) reported ip Eke average of tree measure: Wot” stud th effect of dwntow rate on pre ‘ace pation, srng the apparat pictured n Fig 11 ‘White attempting to maintain a Lg lel ald ver” tee tee eectons of 1 1, 2, 24,8 and 4. pipe, be Fhised the water rte and strved the pease Pls tion n'a 6. feed pe, Web's data forthe 2in pipe fare parlieularly ineronting baeao, instead of noting the low puletion Mwrate, he measured In oeioe of ‘water the amplitude of the pulestion. Tare data Showa on Fig. IL with the dowrate replated ar a Froude mumber. Bxrpolation af th data gives a zore- surge Froude number of (Nr )y ~ 052 Ths rat ‘tgs remarkably well with the ilo. pipe date, at ‘wel 9g with Webb's other experiments ‘Wabi'e data probly do nat rete! buble hydraulis lone. The 2in. feed pipe discharging ino the in Die could have eaceet am entrance eect. Bnteance Dhenomens are camidered in Inter sion, ‘Down has leo boen tied aa siphon he semenon. In an excelent article on this abject, aly* {epored minimum siphon velotles (Le, the maim, Aowrate at which vapor is removed from the pipe) In 2and 104m pie Ha date sre converted to dem metric Froads namters a follows DDOWNFLOW in verte pine may cause plson—Fig. 10ey eae 7 ne Qa : a oo Fo] These results deamatclly asteate the mp Repo _" Fe str: aoe sgveater than (Np), = 031 canbe attebuted in {the Soweneds at which pons fmm With toe hy felony pipe, for exumple, a alpon forme euphiy a minute when the Froude nomber vas (is). By comparison, Kellys 10-in. pipe was B47 ‘tho apex to the outlet At (Np). = O38, the go ‘would be moving downs at roughy (8) (a Angle o rts petin woe “He PP] 83/088) = as eels. The babble raid on a 7 S4:7/14 = 89 enc, travel down the pipe. In veal V [ tance the gar sigs are remove, the orate Wil De ie fo increase becatse of pressure recovery. As ra on | ‘minimum siphon velelties ae dealt (opp DE 04 GS 08 LO te The velocity t which large bubbles in a tbe Froude mente Ie, © WD ‘rough » quiescent Higa bas eon staded bath retialy and experinentaliy- A theoretic] sy Davies and Teyoy indicted hat he buble PULSATION dota of Webb in Frage number trm=—fi. 11 Vj = O58 (g D)s. Harmathy's® rst bbe comple ant Mato the tg. )* Tigaia tre ed th en he hight iment ined ly by jon 3 bre the ties, ofl oe 0 08 ye 3 fi i ‘won ces ei raion ‘SIZING char for pipe handling Fouts in vers! downto allows one of fit dymamies—fa, 12 rman eo en 17, 1968 Chel ennto be the hast experimental result it ie (ae DY These results agree with thse for ing liga arcund stationary bubble. ary, large (dz 1 tn) long babblas become is vetialy dowafowing liquids when # < ent: venom VHD a) tho wooly is less than predicted by a. (18), vl rise. And at higher veloc, bubbles wi ‘downward end removed from the pips, If ‘coaree of vapor it sealable, 81. = Vu) ean be expected to predice presrare pul viraon. hs been prepared to ald in the solation of The lwrate fa converted to gpa, the fas changed to ing and py ln assumed fo be much fat ls for (Neu — 1, nese Which peeaure ‘mpltader will be high and eiphoar Wil Penis. The second line ts encountered when fe inrenced further; fetal free oacts ena) fore, and no Pretare eradient~il be Hn vertical dowallow. This latter line pends Reyoolds numbor and pipe roughnoss. Tvs line is for biph turbalena conditions Coa, water) eoowors | Downtlow From Process Vessels onal downfow je often confused sc the ble phenomenon inside a vertical pipe Flow ae the ext of «process vessel, HONEWE, AEE se than th simple able dynamics Lig and eatrance goometry besme impartan: Yar st compleate the analyse she geometry down in Fig. 28, guid forme’ reir in the yesal when the Froude mutibet Di leae than seughly 03 or Dee tha sid ows down the ppe ina falling ls. The ft the pipe and weet contain vapor coe thet pececlahy sucked into the dowafowing iti faatventing Gow ie not completly pretstale real may occur at Prone numer some ipestor shan 02. D. 8. Ulack™ for eximpe Ut the transition eeu at (Np). = O58. {he Froude momber exeade 0.3, vapor willbe into the dowalovingHawid aloes suBeient ght s maintained in the proces ved. The nt put ce etc! qu elebt hae been id experinentally by Kalina and shea. by Harieman and others” Heriomas oats ren In Fg. 18 ahould be a conservative design, B vapor wil not be eucked Inta the dovaipe st the gad height pied by. tie ean rite! qld heat t which entree fs ‘a troquetiy unknowa, ‘To test far vapor nt, the critical height Js conpared with that void exist if entrainment were ignored 1f the aight ie greater, por wil be auakad Into nila gud pr aad setive withdrawl of saline Wa | geering—ne 17,1868 sane ter apatons. Te stems Tikely that his equation could th be apeled to tre tmineibe igo paoes (ee ter deslgn). Hovever, the coufelent of 8:2 given in 2. In adliions the Ligbier phase denaty shoud Te lee gon Ba. (14) Wiiripools in Process Vessels Compared with irottional fw, whirpoo! forma lon ls uaualy very unpredictable beeauee the fore History of 2 Downilow Problem tom tthe S0aE ne wan ear evaberi Pre: SSIS APE oe Rees coche Be ecto oe ‘puleation sot vibration was apparenly coming The Has od fave ben de ‘ate 3 eat i oe Ry tra ie eeti ofan mare ot Froude mmoere Fg oe vertical flow. Only when the two-phase pras drop is high wll tho eatimatee be veativey ide af pipe erentation, W's Correlation corrcltion of Lockhart and Martina has wide acceptance by engineers. Though it wat pein the 1940's, researchers in this eld en et i ays wtandard of comparison for thelr ov ona, ‘The popularity ofthe Martin werk ore fein ita simplety then fom it acearsy. timate tworphace presaure rp with Mar te correlation, it le dest necessary to estimate prssire drop each phase woe have i€ it were ‘by telf inthe pipline ‘These slephave Pe droge are denot AP and AP. We cn then the peramater X: ow is further cased into Uheflowing fou TF cach Ja considered ono in the pips, Both the ‘the gar are in turbulent few (earbulet lent ow) Ligud flow is viseos, ad gus is turbulent (vie. 1s viscous (tue Both phaove ane viscous (viecoue-viseous), ‘the ow has been categorized and the param, AF hs been eaculated, ether the graphical or af reults in Fig. 16 ave used to determine a oirng—toe 17,1968 1 FRO FAT ose Tae pasamster 4 that i related to the two-phase pressure Aroptheough By. (18) BP =(48aPe = BP a) —aPoy Je total ead le; elevation fects must be saad to “Pro. Ry and Ry are estimates ofthe frace ‘on of the pe erosestetion cewpied by gaa and fas undar operating conditions (Be + Ry = 1). Ths free fraction ot holdup fa in ghnera, diferent from fhe volumetric Bow feaston. Ay and Ko ae used to estimate tn ate sverage mixture denlls for verte ow prosuredvvp emimates The mine Seat i = Rape Reso a) Martinel’s holdup correlstion is considered int oe t0 the spore complex correlation of Mughmsrk Conseqaenty, ther presuredrop creelatons that (epend dreey on holdup should not be used with Example: Pressure Drop in Two-Phase Flow timate the pressure drop in $00 fof horlontal ‘Lin, Schedule 40 stel pipe careying 60,000 Ib. af Wale and 8,00 lb/hr of a at 70. patra pres ue i100 psig. Vises of the vapor and atid are 1018 op. amd 1 ey ropectively. Ligold ena The lguld rate i alternately 100 gpm. Considering tigud to fow lone, Fig. inlcatea that flow would Ipetarbulety and tit total presto drop —SPs Weald be (0.28) ($00) /(100) = O86 pa "The Reyoolds number for ga alone is from Big. 8 Nae = (632) (500)/ (0018) (4026) = 436,00. Fig. 3 ‘would Indlete Row at heh turbulence whare Fig. 6RE ' 15S FING.» tnd 1a, (10) ett Be uted, Atsome two-phase pret fre drup is 10 pel, average air density 056 Tb/ fu fe and aie premure drop —aPe = 1Cx100 p (3) (0.005) (BI) / (086) = O89 get. (Pig. 6). There fore x stapyaPst = osni(@ane= 1 For dorbulent-tantolent ow, de = Finally naPie = nee APs = (208) = 38 pe. The a8 Tamed and envied preaeore dros aro dose enough fa that the average vapor denalty doesnot have tobe ‘covlantad, The eetual twohare presoure drep ca Bo expected to Tie betwoen rqcehly and 20 pl kler’s Homogeneous Model Dukler’s Homoger si Dukler® bas propaced correlation based on the sumptions that gus ad gud Mow at he same vos {iy too slip) and thatthe properties ofthe uid ean 1 cuttabiy evoraged, The correct averages are pent eet D @ rem Qe/lQs +00) ‘once the average density and vieslty are lated an average to-slip Reyaclde number ca he Madge « itt. = 20 (abe = Dia me ‘The feiona pressure drop is ter: Pea pane Inbese f 8 the conventional Panning feetion Sstimatod from Fir 3. "dition to te fictional pressure loa, thane tert oot anointed eth expanalon of the pa, ‘eometimes term an acslorationLxs oF maa (ec To study thie momentam ‘fete ple tnomentam tale for atendystate Bow fret intended were = fe ote ldo slid anny Sid my = te) mao ald. Bh ar (Gn) io eae engtetng portant use for ansign of ert dtibator a den hing evlagsoenty s8d-eqpment {Sree The equation in yesor equation an therefore be ont fn a angle dreton, Probably Insert seunpons inte ogaton re a ely pole efit ove the pipe roseseion That the crustal ava ie Constant if a torphate fil soment se couse! ta Pint {to Flak 2nsitntl e E be simple to By. (25) “a 2 ‘The fod doos exerts feetonsl dra force 2 bie wall, This face has been excluded from Bi for convenience Tell be aaded tp tho frm Iter. Pho Tato gravitational term in Jae 17, 1940—CheiclBE ES BR BR AR RM TUR A BR BS Re He ae aR ee = PRR! EERE fees YR te RE fp GE EP GP ESP HP GE GR : ee ee Elen ona toe ane con ea ps tag Bo Ow He & oo RR a a ocauae the seclertion of gravity, a doesnot the borlzonfat erection, When velocity Vit to maerfom eate wy, Eq, (26) Becomes: the equation that deine ao-lip density is, in of tao Faas and (28) are combined to eve Ba, (29) iy, Ba. (29) i divided by the tolat two-phase ire poy —aPep and she eal gas lew is asd 10 Laweeen noe ae Pry. and py. a the egthmetc average pes aod gas Seni, reaectvely. ah =) 5: Dukler’s Corelation fers corrltion wil be zed to check the pres: rep in the previous example. Ayal itis neces to fst saunoe » presaure deop—10 ps wil be ‘ae = sno 23 = 503 0 Ge ~ 00/046 = S00 oC we ill ealealnts the volumetele traction of aud pase, the nverae densty, abd the average rtp from fhe several rostions, as fellows: ae = lie. O9, 35 2 Bie P hon, = aos one-one 17,1968 aa. ‘Example: Pressure drop with flashing flow—Table I Boas Fee eae B seaeeds + pall =a) a ove = LSP SIS BJ. vs = (OV ESREAHL? TA -o0m 9. From Fig. $ and Ba, (23), we evaluat the noalip Reyasda number Inte easy" From Fig 9 we gt the fstion factor f, and com- puting the tio wD /D = (ooms) 02/6005 = 00mm yoo Instead of Ba, (24), the more convenient dime sional equation in Fig. Swill be astd to estimate oar Oe carey ~ (sss") ns song won esos ‘Th soeleration term la no cleat as a= [ants (ais) = (oaPeny = ‘This eatimato is close enough to the assumed 10-3 presse de,PROCESS PRN ‘he acceleration term Au ie son to be restive animportan. I, however, the preture had ben ower tr the manly @ higher, Ay would ave hod a ‘Stowtantalefect on preere drop. Such fa the case Sth the Gta used in Barinlfe correlation, As {osu Barinelis correlation can be exposed to give high prestredeop etimate where Momentum is ron the above ditcussion ad calulaton, wih oth Marines and Baker's odes "bet gate” of the fworghase pease drop forth examle: 6 ple < Pan < 1 pal. Design of tho proses should be stl lentiytexbl to alow for this Fang of presse drop {or tworhise fw. Loss in Valves and Fittings Although at present there are no generally accepted techniques for estimating tre-phase rity loas ales and fittings, progress is Being made. A method ‘hat hae. moet with tome succes ie the oli ‘model The sngle-pae vlelty-headcoeficent 1 eee Table ID ie used togethor with no-alip vlelly and novo density [dened hy lther Ba. (20) oF (23), Be. (1) ie then wsed to estimate the two-phase te- [MALOSTRIBUTION of flow n mano pling depenes on flow decton ofthe fds. 18 ‘onal head oes. 1 Is appacent that this mst ‘ot Slade the sceleration tarm noted above, Pa reason, Eg, (00) and (81) should lp wel to the frletional esd los win mass ellis ang and/or presoret are low. The Delp anageg eq (92), when relates the ead loss to the welty an ot) Where Aa agsln the dimensionless stelersin timated from Ba, (30) Vs2 K ae ibe was pa = Gi =v Flashing in Lines int With tw-pate dow, the god fren Souls se oling pss Prive deo came ld lls Beier fa sly gmesting Sin aay gl, freaing the mrxtore vlely and reluing the ‘he caslcal approach to ashing fow wes sien Benjaohn and iiler*= They sasumed Da tbe sod Tigi ere at egoibluna and tat bt awa ieveame velociy Coo al). A diferente! tga the macroeorc mochanical energy bales wa ated fo give By (89)! [nian] is weer es a, 09) cate atte aleleal ni Ol ‘ eae coma aeccl a me Wotineaster octet wae re ee a pat Beanie td Mier see Git tr Dielee tetotayypeN al a Beara iret ene a epee Siac tensor Oly pane fo Mo, = Svat yasue ve fh a SG come Stone abe Fay x es ‘ecort shpat toate tre Ee le ‘eSleined fe yest te fom Ft cm trek erate gil oe sane’ clan eee ar avy vite er itt mint rly 5 Soe doth Walte of Benumin and 8 a ee eis tied nue a Fraarevate Wek te pec ated Conall 200171968 Chemiel‘quently ows fe the Daa Sonal vapor facing’ the ow was even hat the both ow rent) form 4 er vay. One mass voloclty can be ealelatel direct from aoe Cone = NGF OP TTC an example, Benjamin and Miles show that for voter sytem at fe = Q0802 and P = 22 oc LAB Wendt end APrr/Sype = 3/018 (pei) (cade). Therefore: TEC T = calelate valu, glven in thet paper, ls 251 1h Sea). Later experimental wove by ebin and att indents that Gar, = 680 Thy (ote. auf) “hie example: Benji aod Miles mathod wil rontiy ate main ass Wein that re Tow the experimental Gv 1 approached oly ‘he weeh fractian gus 2, approaches Mery the noalip assumption ino: valld for ext How, At present the best analels of ertial mse sites te probaly thst of Fanske, The reader is rot to Tong's book for» suumary of this 20 wn meri sd Dulle’s anstaiscan be ellsby wife fo dash Gow. Inetead of Eq. (2), the romentum term iy, the total two-phate pressure dap is: eaten] 6 tt ay» BF yr aca ets moran aioe siesta wi atts be ty eek te St ea eet iat ae Se Goat ama ls Sims ost 20 te te pang ge oe artes) gate ace aes Se teat i ed Sr intra RS HE wry ocean ns tnd ae intr fame lapde Waar oan Srl Fase Psu eee Se a siti tee ere Benson ra ‘correction factor fo frictional pressures ‘rop—Table omacon Factor 6, When ‘Side ovis Content con Freon Factor co Factortet Constante) 2 0250 257 i oars 310 4 ase 8390 5 300 336 é B08 Ost a eat aes spied to large posesure intervals with only amall Computational ernen, A much larger eee Introduced ty testing that the tivo phases ow at the enme ‘eocty ‘This error can bo expected to be larger than ‘ha for iethormal two-phare ow Fuid Distribution in Manifold With the praent-day emphasis on large low-prese sured sit processes and the ererdncreesing wee Of parallel equipment, tS csental teat the design fhglneer become familar with Aaiditrbutiontech- fplogy, He must be eure that Me piping design will fot cause an andeivable preferential ow 40 8 paI~ Feuer piece of eatipment "We can analyze that manifold configurations with the sume Bq. (25) momentum batance used for Dxo- ‘hace Row enayais. Again the Aalé clement, shown in Figs 10, ie ssected Inride the plpe boundaries. Since nly Berisontalweoity eomponents are considered, a 0. The stance between Points Tand 2 2 ema fothal a= O and fection ean be noglat, Again he cocltyprofl ie aeomod Mt. I, itand, the velocity I been azeumed proportional é2 the one-eventh power of the distance from #he wall (6tandard power Tie) can be shown that V ta Bq, (25) wotld be ‘eplaced Dy 1.02. "Be. (23) again reduces to Eq, (26), oF tothe foe owing tore convenient fray, Get proposed by Acrivor Bok Minar # Nee that this equation, ke Rerpoull, pedis prete ‘hre Hoe from velocity redton The shown in Fig, 16, Suid doco not jeave the hoe soil pipe perpendiclar to the axe of the hoeionta fr main pipe: Consegienty, the fuld clement loses ome of che horizontal omentum fox, resulting in fowce preotare recovery than predilel by 4, (2). aFriction of Prom? The momeutum coeient & accounts for there non lava. “Walser ofthe diversing-manifol coticlent k= ke were estinsid from several sparees®« "2%" The “ale for fy was found fo vary from abost 0 for Dow bredominentiy throgh the branch, te 9 €or high flow {nthe main pipe. Despite large vatiatlons in the data land the obvious mine dependence on pipe geomstry bd flow in the brash, the follwing average fe 5 fultabe for mart eapicering abaloes bm 08 2) or conversing fows, the destion of Strcam 3 is roversod, giving it a tendoney to jet perpendicularly {nto the min piping. Thus the momenta cooticient 1 howl be samewhat larger than fy in dvorying fo. Indoed, the data’ Indesteehat Fr, varies from O75 0 12 Ba, (40) gives good average value: =a wo 1m adition to the moment effects decribed sbove, friction can cause mulistribtion in long manifole. Fig. 17 deplete viding manifold of length L, diam ster D, and Sale veloty Vy. To propre for estating Imaldstribution in the branches or sideports, we fist ‘toamo that the selasity ie equal in each sidopore ‘Thon the fectional resure drop ja eatimated from ‘ho inlet slot V, the length the diameter D, and Inlet freon factor . The friston factor i aasimed to be either contant over the manifeld length on vorssy proportional “0 the ene fourth power af the Reyane number The parameter , tabulated In Table IV tn thenrelatd tothe fitinal pressure drop based tn inlet eandilons by Ba. (HL) Fy = CaP a whore AP; = fretozal presare drop over length Pp. = frictional peesure drop based on Vin Ly and ‘he tietion tact st the manifold Snkt (or exit far combining mania) TOMS phenornnon: ation to pete ie 6 Tor diverts equally paced ports having eau dom tnzough each port, the C factors Table TY Car, aviation tn both the frlction factor and umber of sldeports has relatively unimportant {ee on tho estate frfetonal pressure ose Design Methods: Manifold Piping Detaled lang of Sow-istebatin ins i ine taming, and in net caoeaimwaranted. Use of fat methods wl generally reste im answers of ‘galt. Tp vedo all probs to a commen bai ‘B'euvonient to dei messure of tal is a As fst sppresimation the Rowe eho port is acamed equal. Bq, (S8) and (31) can the vee (Babee of Ian eel to alelate the peecure tipbalance in the piece ‘pir, which int uae asa measure of the pe iaee lee in the eideparts, Tho equare root of th ae Gp aco ach parle stem i then an 372 aren meme of te omate Bq 42) reduces efter = Akboath Te aoe (Sie -1)m ral at For most pices equipment, the following cl eee sil pve good dietrbution Eng mt Ma 55% i aa ‘he preceding prncpln can be ostratet yma fet emple shown Pg. 8. Water owe inne ont wth 500 gym, foving tothe righ AMS £03 ‘on gpm fo tho nt she Ba. (88) throug (40), head toll toPorte 18 and £8 bce then 208 ener ev ert wil generally giv the assem mali on cn Tait mods Calaat the presrre dma a et te onto et and Elton eect piesa. Jone 17,1968-—ChaielMomentum efect: The horizontal component of e- ity atthe exit Vs = 0. The alt velacty Vr 55 t/t. for 500 gpm. in« in, Schaule 10 pipe There from Ea, (38)! Pin (2) 00 Sse Friction eft: Pour oideporis, constant tection or From Fig. 4 and Table 1 (PsP = ($38) 0.2m - 0 5 Ortiet manifold: Again caleulta the pesesre drops ue to the momontom eet and friction ec "Momentum fee. wer == ( Priction ofects (Poy = 000 the pressure drop setoar the ead exchangee (Pe — ) 8 Bet von; adatom 0 sets having e088) re Tate TV Seton tector an Seely weimortant ‘pressure looes Poet Py 4D Bt (o> Plat Poy 0 on nas tn ‘rented. Us of ata conmen b ure of mal vo ((BE =» no he above analyse shows that shout 396 amore water ow Chrough the 1 porte than through he 43 2 ood Aw dieteition, red Flow Pattern by contrast withthe est configuration in Fig. 18, ther two versions will give extimated maldistt, a of IT and 28%, respectively, These examples to the folowing conchae ‘The countercurrent How patterns are prefered ie covarrent pattern even though ‘leis! of ‘in the Inlet nd outlet ports are ates is sree fw Although Table I indicates that @ 6-0, tine may namical size for 3,000 gpm, maldaribton cat reat the recltant high velocity unless the mas. Bring methods prosented above hayg been ap 0 uniform dlametsr masifull pipingdeheaoe spacing. The same principles can be applied to old ping of varying. croesction, designed for constant vlelty Iti doubltul, how ‘thatthe additonal est of reducers wal elec iametereaving, except fn unugaal cleumstancn tier, a enstent manifold velocity wl aga fect at frition on malate ‘ado-gpm. inet low inthe ft example in 18 will separate from the wall cf the berlontal 1% “Soave the neta “asia eal Te presse dont ‘tion fest bipe at the entrance 1f Port re too chow to tht Tnlet pipe, the proscure P would be consersbly highs than predicted. Ifthe 1,000e0m. pipe were Peres Aicalar to the plane of the gmt, the presuee P Wold be’ nearer expectation, but enprectble, 7 Iunimie thia entrance effet the entrance pipe (01 fom.) shold 6a mlntmum of anat rising dam ters from the rst inlet port Alternately, toe pipe ‘ina ean be increased 60 that ob oto weoclty head, ‘€ prescure variation Will not upeet the flow dite Two-Phase Flow in Manifold 1 is sometimes suggestd" thet two-phase ow die tribution problems ean bo ested hy use of eymaet ‘el siping. Untorbinatady, this entrain i 4 bocesury But never g miflcat base for goad fom distribution, even In atnghphase tow. Dwo-piage aye tems should racive the utmartstention By the de- slr, “ven. undor he maatSavorsblesecumstanes site Iely that a two-phase micure, when spit In Ianiflé piping, wil be poor datsbate Te mal {efbtion of vepor and hgaid cant be tolerated, the twophase mista sould be fa separated ithe in sn entrainment seperator oe a Mash tank, The two ‘Dhases can thon be recombined at etch etry pat, A relatively small yolumesre‘raion of vapor wold be tparged into a continaoaelqld pss, whereas a tow lqud rate could bo either sprayed troughs mse or fed through 4 perforate dlttbutor bige ‘Non-Newlonian Flow ‘hos far, it haa boen seaumad that all ude ace Newtonian; that i in Tatinar aw, shear steed fe proportional ta the shear rate. Atfhigh this rump Won ts tee forall gaan nnd most nos comely countered, thre afe many mata that do not 2 low this simple relationan. Ther are the nom New. tontan fad, examples of which ae concentrated elite ea and polgmerie materia ‘To aul enzinering probes with nox-Newtonian fuids, the enpineor mat Set hae information abt the reltinshlp between shear ses a chen tte ‘his relationship must be determined experimentally fn the way that vstosity Ie found fer» Newtonian laud When the dita are avails, they canbe fied to one of ecverl mathematical models corently in tse. Then A Tange number of laminar-low eneineeting pela fan be solved, ‘Turbulent non-Newtonian fow is mire complex ‘Thore ave, however, 2 few tools svalae forthe ei uation, The render ia referred to recent seren of ticles in Chemicet Enatneoring by Mei, Wohl Friction Reducing Agents (One cls of materials that does aot lato the con- ventlonal clasifestion of ether Hewtonian or not NX“ Newioian folds are the fection seducing agents ‘Thr behavior i termed the Toma phenomenon. They frehave as Newtonian fide inthe lnaluar reson, PROCIsS ING ‘Ata given Reynolds number, the Triction factor is tetween laminar aa tbulens ow values. Payoxt i a gelsmer that, When added to water, seduces turbulee frelon, The mime effect pro. {led by concentrations in ppm anges of thi polynee Ie depicted in Plge 180% A Nyy = po feeton i) Shi lee then that for smooth pipe ‘Oe rather satafacary qualitative exjanaton of the Toms pltenimenon fe that the fritlonredution Sesto are pesont ina given Iqid as lng ebainmole- ‘Slee The molecules align shemaeves withthe flow and ‘hercy hinder turbolent eddy farmaton and delay the transition to turbulent ow. although thre isa rend foward. treating this phenomenon. with viscoelastic {echoology, there a mo eatlatactry quantitative analy- Sst the Teme ebennomnan. iletloyredction materials tend to degrade with tine Tf they are ed ina locodycle water nstem, the activity of the elation would have to be contine uly rejuvenated A water porge may aio Ue neces. Sry to prevent th vesaty from gradually increasing. Conclusion the author has attampied to prevent camping of tne pratieal tools svalale Cor sing pevcess ining To do so and yet cover the broed ld of Haid fw, crus nocrary to omit many toafl desexiptions, cor TRtaons and theoretical developments. The reader in- reted in parmung this eobject in greater depth Shon consult the reference “may vem that dae emphasis as boon placed on twophase upow and downdowe penomens. This Wat done intentional Doease the technology as Thang fuportant indore apleations, and becatse {he author has not found a satisfactory treatment of the anbject iewhere “fotably sheen frum this artile ls tho technology cof pavtcuate solide Sn elorey ow and in pneumatic Chareying. Both involve important pipesizing rob fous Ret 1 and 0 cen be od se ah ifroducion to selido ow. raha plaosd on thar sling methods that have ound buses Been ifthe hese in notog sore than Mtpy our eatimets of the eeonofea ine size Sut Hive a tater eeimate Wun would be obtained Fran an arbitrary vloaty aftr giving consideration to posible future expansion, eroston and increasing ‘iain. ‘Acknowledgement The author is grtefel to ie colleagues for their sistance ad to Union Carbide Corp. for permission fe publish eis enteral, References fee Heat tars ae A wt He eats tat oY E stig ests Ske a, ie aera nee jetallc pipi ss industri tn 0 piping ees cake fae Os gin. Puen ee ete ON this eet ea nly wit ‘ste the ear oe erie SNe ae ee aaa iS Ets roan eae Beast ag Hs dita Baw TETAS nny rurnns. oe 38 barat at ‘his ee sates tate aad ‘The watt demand |