Stress Concentration Factors in T-Head
Stress Concentration Factors in T-Head
CO
AMRA TR 66-36
STRESS CONCENTRATION
FACTORS IN T-HEADS
oc
I
TECHNICAL REPORT
by
FRANCIS I. BARATTA
NOVEMBER 1966
C,
mm
U. S. ARMY MATERIALS RESEARCH AGENCY
WATERTQWN, MASSACHUSETTS 02172
-* srcSEC,I"'"
U
BW
A"
"
n &l.m&
[
ABU
,ST.
ic/miiB;un
GOOES
ttL */ SKCIAlf
DISPOSITION INSTRUCTIONS
Destroy this report when it is no longer needed.
Do not return it to the originator.
T-*""I'Tnti
i ,"|'>t^ '*~-^*~ .-
*mr
by
Francis I. Baratt a
November 1966
U.
S.
WATERTOWN, MASSACHUSETTS
tm*m*mvmm
0 217 2
^J^T^m
ABSTRACT
Two simple formulae are presented which predict stress concentration
factors applicable to a two-dimensional symrr.*trical T-head configuration.
This configuration consists of a deep head joined to a shank by fillet radii.
The independent equations that predict stress concentration factors for the
same geometry are derived for two different leading conditions.
In one
instance, a tensile force is applied to the shank end of the T-shape.
Equilibrium of forces is attained by supporting the bottom edge of the head
section, resulting in the shank section being pulled in tension.
In the
second instance, a compressive load is appliec to the top edge of the head
section while the configuration is again supported at the bottom edge.
Thus, only the head section is stressed and ir,> a compressive manner.
Because the analysis is not exact, the magnitudes of the stress concentration factors resulting from the predictive equations appear to be
overly conservative at some ranges of the geometry parameter ratios.
Therefore, an arbitrary "limit of application", as it is termed in the text, is
recommended when using these equations.
Again, because of the inexactness of the analysis, experimental stress
concentration factors are indirectly obtained for the first loading condition and lirectly obtained for the second loading condition mentioned above.
These data were obtained for several geometric ratios of the T-head configuration aid compared to the corresponding predicted values.
It W4s found that the formulae could be utilized, with engineering accuracy, witiin a certain range of the two pertirent geometry ratios.
Beyond
these ranges, the error became excessive, but conservative.
..'
CONTKNTS
age
ABSTRACT
INTRODUCTION
FORMULATION
.......
LIMIT 07 APPLICATION
............
RESULTS WD DISCUSSION
12
15
ACKNOWLEDGMENT
15
REFERENCES
16
liiin
^_.
INTRODUCTION
A brief literature survey is presented of past work relating to stress
concentration factors in two-dimensional T-head configurations.
The results
of the survey are pertinent to subsequent discussions.
Hetenyi
was one of
the first experimenters to test a two-dimensional T-head configuration.
He
stressed the body in a similar but slightly different manner than that shown
in Figure la.
The results were applicable for geometries which included several h/d ratios but only one d/R ratio.
Twenty years later Hetenyi presented
additional results which were an extension of Reference 1.
Heywood8 further
applied an empirical formula to the initial results of Hetenyi.
His objective was to extrapolate Hetenyi's data by including the effect of d/R for
T-heads of various h/d ratios.
Nishihara and Fujii4 obtained, by elasticity
theory, the stress distribution in a two-dimensional bolt head.
Although the
results of this reference yield the desired stress concentration factor, the
formulae are complicated and applicable over a limited range.
The objective of this report is to obtain an expression for the stress
concentration factor associated with the configuration and loading shown in
Figure la.
The formula is developed because:
a.
the empirical expression given in Reference 3 is based on the
limited data of Reference 1; and
b.
the formulae of Reference 4 are not amenable to simple computations
and are inaccurate at some ratios of d/R of practical interest.
Formulae are obtained by
superposition of various loading cases of Figure 1 which
are guided by experimental
observation.
Two expressions,
which yield yalues of the
stress concentration factors
applicable to Figures la and
lc, result from this analysis.
These formulae increase the
usable range of existing data
and are easy to apply.
However, because of the apparent conservativeness of the
resulting equations at some
ranges of the geometry parameters, a limit of application
is suggested in the text.
Also, since these formulae are
not exact, experimental verification is provided.
.**
an
[Uli
/
-d -
h-t
w w
amJ
L- _.i__i _ J
(a)
*a
(b)
*a
(c
FORMULATION
Photographs of models loaded similarly to those shown in Figures lb and
lc are presented in Figures 2a and 2b.
In particular, note the isochromatic
fringe distribution at the fillets.
An experimental determination of the
location of these fringes at the fillet periphery for both loading cases
revealed that for this model which had a large D/d ratio:
a.
the stress gradient at and near the maximum stress site was
relatively small; and
b.
rrm fTTT
r*
ID
m
a.
b.
where &fjt
CT
fa
of Figures la,
8nc
ic
are
tne
fa
max mum
+ cr
fe-
rn
The stress concentration factors for the loading cases shown in Figure 1
are defined as follows:
rfT
iT
fa
~n
(2)
and
,
kfc s
"fc
'v57d)'
fT~kfa *
fC
(3)
Thus, it is seen that the three concentrators associated with Figure 1 are
directly superposable, if it can be assumed that jaximum stresses are
coincident.
It remains to obtain a relationship between the stress concentrators,
ka and kfc, to reduce the number of variables in Equation 3 to one.
Accomplishment of this is realized by equating the loading state represented
by Figure lb, for which the stress concentration factor kfa is experimentally
well known,5"
to the sum of the two loading states shown in Figures 3b fend
3c. Figure 3b can, in turn, be reduced to the same loading cases as those
shown in Figures lb and lc, in which it is again assumed that the points of
maximum stress of the two cases coincide. Therefore, the stress states represented by the loading cases shown in Figures 3b and 3c are superposable
and result in the following formula:
fa ~ a fb a fc
(4)
where cr^ in the maximum principal stress at the fillet of the configuration
and loading hown in Figure 3b.
iilllWIIIMgBWa
wmimmmmmm
'111'
[III]
I
111
III
t^i; 4 *- 4 -i --*
*a
(a)
* A_i._J.__i
a
b
(>)
LJ.
*c
(c)
(R/d be small).
(5)
Defining the stress concentration factors kfa and kfc in terms of the
applied stress shown in Figures 3a and 3c, we have:
r
fa
'fa
and
'fc
'fc
* crc(D/d).
fc = 1
(6)
'
or
'fa - 1
fc
D/d -
d/R
(7)
- 1
fa W* - dTR)-1
cfT
(8)
D/d -
d7R
- 1
.k*
'faa
% 1 + I
(9)
["(D/d 1) d/R]'65
[.2(2.8 D/d - 2)J
2_
D/d - T75d/R - 1
(D/d -
d/R )- 1
(10)
and
'fc
2
D/d - l
d/R
(ii;
The reader is cautioned that Equations 8 and 10 are invalid for small
values of h/d (see Figure 1) since bending of the flanges of the head is
precluded from the analysis. Hetenyi2 indicates that when h/d is equal to
3.0 or greater, the head of the T can be considered infinitely deep, thus
eliminating the existence of bending stresses at the fillet.
A more useful definition of the stress concentration factor for the configuration shown in Figure 3c is kj or, /a , and from Equation 2 we see
I,
A
fc * *
L2<2-8 /d - 2)J
(12)
^-TR-1
LIMIT OF APPLICATION
The stress concentration factors, kf-r and Kc, described by Equations 10
and 12, are shown plotted as a function of the constant parameter D/d and variable d/R in Figures 4a and 4b. There are distinct minimum values for each curve described by D/d in these figures.
As previously indicated, it would seem that
as d/R is decreased, kfj and kfc should also decrease (which they do up to a
point) and asymptotically approach a minimum value (which they do not).
It is
evident that this behavior is not correct and occurs as a result of inexactness
of the equations which in turn could be cauaeu by:
a.
the approximate nature of the observation that the maximum principal
stresses occur at the same fillet locations; and
b.
the violation of th*
to the width d.
>T
J
T
cr
h/d
4.0
8.0
12.0
28.0
=
=
=
>
stress at filltt
applied stress at shank
applied stress at shoulders
3.0
32.0
d/R
Figure 4a. STRESS C0NCEMTRATI0N FACTOR kfT
rnmrn
0/d = 1.5
1
(7
mi mi
D/d = 1.75
D/d =2.0
D/d = 2.5
D/d =3.0
D/d = 4.0
0/d =5.0
D/d = oo
Jlk Hilft
*t
LEGEND
"fc
-c
stress at fillet
applied stress at head
applied stress at shoulders
3.0
4.0
8.0
12.0 16.0
20.0
24.0 28.0
32.0
d/R
Thus, the accuracy and application of the expressions beyond the minimum points are questionable. The seemingly odd behavior indicated above
wi'l provide convenient practical cut-off points for each curve describing
kfj and kfc. (The cut-off points are experimentally checked for two cases
and are discussed later). These cut-off points can be determined by th
minimum values of the stress concentration factors defined by particular
values of d/R as a function of D/d. These minimum values will provide
limits on the use of Equations 10 and 12 and shall be called "limits of
application".
The limits are analytically determined by simply applying maximumminimum principles to Equations 10 and 12. The limits of application of
Equation 10 describing kfj are given in the following:
d/R >
(13)
2(1 n)
n(D/d-l)
(14)
where n = 0.65. Both limits have been computed and are shown in Figures 4a
and 4b as lines connecting the minimum point in each curve and are labeled
"limit of application".
EXPERIMENTAL STUDY
Because of the approximations used in the previous section it was necessary to establish the validity of the derived equations. Therefore, the primary objective of the experimental program was to determine the order of
magnitude of the inherent error associated with the determination of kff and
k}-c by Equations 10 and 12.
To accomplish the above in the simplest manner, the stress concentration
factors kfa and kj-c were experimentally determined for four cases which had
the various geometry ratios shown in Table I. The stress concentration factor
kj was then obtained from these data for each of the four cases by simple
superposition as indicated by Equation 3.
Table I.
EXPERIMENTAL DATA
Case d/R D/d hf.
>J
kfc
SUPERPOSED DATA
Coabined Factor, Equation 6
k
fT
f.
+ (D d
/ -d7R*mfc
E l 6
'
PREDICTED DATA
kn
Percent
difference -;.
Percent
di fferencel
JA
0.94
2.86
- 0.7
.1.96
- 4.4
IB
1.27
3.79
+ 8.6
-3.10
-'5.0
1.41
5.71
+S1.8
-5.43
+ 100
1.62
3.13
+25.7
-2.65
+ 145
II"
'Obtained
tCoaputed
fCoaputed
**Liaiit of
LLL
!__
CASE
IA
IB
IC
0.079 0079
OXJT*
0.110
BJOOO
2"
2"
2"
2"
7"
3"
5"
5"
1.900 1.129
1.129
/
f.
NOTES:
1. DECIMAL TOLERANCE
IS 0.001"
2. SHADED AREAS SHOW
CARDBOARD PAD
LOCATIONS
*h
(b)
Figure 5. LOADING AND MODEL PARAMETERS
Pads were used st sil contact surfaces to siaulate constantly distributed load.
H t
\
mm -
H
CASE
IA
-100
-200
-300
IB
ZC
l 125
ft
2"
2"
2"
2"
5"
5"
5"
1/2"
-400
NOTES:
1. DECIMAL TOLERANCE
IS t 0.001"
2. SHAOEO AREAS SHOW
CAROeOARD PAD
LOCATIONS
-500
P APPLIED LOAD-LB
Figure 6. FRINGE ORDER Nfa AND Nd AS A FUNCTION OF APPLIED LOAD
10
\
-
CASE
IA
7.0
6.0
19
IC
0.075 04)75
OJOTS
0.110
2"
2"
2"
5"
5"
1/2"
1/2-
2"
h
t
R -d-
in
1/2" 1/2"
Hfc
NOTES:
I. OECIMAL TOLERANCE
IS 0.001"
5.0
r.-
5
UJ
-300
-i*00
-500
-700
P APPLIED LOAD-LB
Figure 7. FRINGE ORDER Nfc AS A FUNCTION OF APPLIED LOAD
I 1
z~. i-jisri
Uli
_Nfc
fc
and
k
fc " kfc(D/d).
The stress concentration factors for each case are shown in Table I.
It is noted that in the formulation of Equation 6, it was necessary to
make use of:
a. an experimental approximation, i.e., it was assumed that the points
of maximum stress of the loading cases shown in Figures lb and lc coincided;
b. an experimental fact, i.e., from Reference 10 it was determined that
under idealized conditions the stress concentrator would be equal to 1.0; and
c. the restriction that the fillet radius R be small compared to the
small width d
If all of the above were true, the left side of Equation 6, which is termed
the Combined Factor, would be equal to unity (1.0). This Combined Factor
could be used as an index, when compared to 1.0, on the relative accuracy of
Equations 10 and 12, which predict kfr and kc. The Combined Factor, for
which the computations were based on the experimentally determined concentrators kfa and kfc, as well as the predicted values of kfp and kjc, and the
percent difference when compared to experimental values are also given in
Table I.
Results given in Table I indicate that when d/R is constant (see cases
IA, IB, and IC), the Combined Factor approaches the idealized value of 1.0
with increasing D/d. Also, when D/d is constant but d/R is increased, as in
12
^F3*7*:
cases II and IB, the Combined Factor appears to approach the idealized value
of 1.0. It is apparent from Table I that as the Combined Factor approaches
1.0, the percent differences for both kfj and kc become small. These differences are based on a comparison of experimental values of krj (and superposed data) and kj-c to those given by the predictive equations 10 and 12.
Further examination of these relative differences reveals that:
1. kf-r determined by Equation 10 is accurate, compared to the experimental value (see superposed data), to within less than 9.0% if D/d 2 2.0
and d/R > 10.0 (compare cases IA and IB to case IC).
2. kc determined by Equation 12 is accurate to within 25.0% if
D/d > 2.0 and d/R . 10.0 when related to the experimental data (compare
cases IA and IB to case IC).
3. The two cases, IC and II, which are at the limit of application for
the concentrator kc, yieJd the maximum differences, 100% and 145%, respectively.
However, these differences are positive and are considered conservative.
Comparisons can also be made to the datt given in References 1 and 2 by
Hetenyi even though the loading configuration used by this author, shown in
Figure 8a, is different from that shown in Figure 8b. The difference between
these two loading states is shown in Figure c. It is readily apparent that
if the fillet radius R is small, the Hetenyi-type stress concentration factor
termed here as kjj is approximately equivalent to kj. The data from References 1 and 2 are presented in Table II as well as the predicted values of
the stress concentration factor kj given by Equation 10, and the percent
error when compared to k|f
Mill
l -fH
*- d -
ittii/
Uiii
(a)
1/
(b)
(O
J3
-'--
'
II
Table II.
\w
fH
d/R * 20.0
k
% rror
d/R * 13.33
k
ffl
fT*
X rror
>>ffl
d/R 10.0
k
fT# % rror
fH
d/R 5.0
k
fT# % error 1
s.o
4.10
4.20
2.4
3. SO
3.48
- 0.6
3.10
3.08
. 1,0
2.52
2.38
5.6
2.5
4.47
4. SO
- 1.0
3.6S
3.73
+ 2.2
3.02
3.30
9.3
2.35
2.58
+ 9.8
2.0
S.00
S.10
2.0
3.90
4.25
9.0
3.30
3.79
+14.8
2.60
3.10
19.2
+15.0
4.90
6.S
+24.0
4.70
5.71
21.5
*Coapted according to Eqeatioa 10 (AH values of kfT aro within the liait of
application fiten by Eqaatioa 13)
NOTE: -y>3.0
A comparison of the results given in Table II indicates that as D/d increases, with d/R remaining constant, the.error generally decreases. This
is because the points of maximum stress for the various loading cases tend
to coincide and the stress gradient becomes small as D/d becomes large.
The results also indicate that, generally, Equation 10 becomes more accurate
as d/R increases, with D/d remaining constant. This is due to:
a.
the restriction that the fillet radius R be small compared to the
width d in the analysis; and
b.
to kfj.
14
\!
\
D/d 2.325
tu
mi lii
%)
LEGEND
<7fT
cr
h/d >
I
4.0
8.0
12.0
16.0
20.0 24.0
d/R
HtK
strsts at filUt
applied straas at hank
applied strass at shouldara
3.0
28.0
32.0
SUMMARY OF CONCLUSIONS
.
1.
The difference between the experimentally determined concentiaLors,
kfj and kr, and the equations used to predict them become quite small when
either:
(a) the ratio of d/R is increased, while retaining D/d as a constant
parameter; or (b) the ratio of D/d is increfsed while retaining d/R as a
constant parameter.
2.
When compared to the experimental value, if D/d > 2.0 and d/R>10.0,
kfj can be determined by Equation 10 within a difference of 9% or less.
Equation 10 can be used in other ranges with a corresponding increase in the
difference, which appears to be conservative.
Alternatively, the formula of
Reference 4 can be used in those regions.
3.
The prediction of kfc by Equation 12 could be useful as a first-order
approximation, and it is considered probable that the error in predicting this
concentrator will be conservative.
4.
The Hetenyi-type concentrator kfjj can be determined by Equation 10
within an error of 10% or less if:
(a) D/d > 2.5 with d/R between 5.0 and
20.0; and (b) D/d > 2.0 with d/R between 13.33 and 20.0.
ACKNOWLEDGMENT
The author gratefully acknowledges the constructive criticisms and
suggestions given by Mr. J. Adachi, which materially aided in realizing the
final copy.
15
REFERENCES
1.
2.
3.
HEYWOOD, R. B.
1952, p. 178.
4.
5.
6.
7.
8.
9.
Mechanical
Trans.
10.
PETERSON, R. E.
1953.
11.
16
4
Designing by Photoelasticity.
J. Appl. Mech.,
AMRA TR 66-36
November 1966
28 N
No. Of
Copies
1
20
2
2
1
To
1
1
1
5
1
1
1
1
48090
wm-
No. of
Copies
mnn
To
1
1
1
ATTN:
1
1
19137
07801
20438
AMXDO, Library
12189
20360
.***-
T-V
No. of
Copies
To
1
1
93555
22448
1
1
1
_!!'-
s
-
- -,.
~-
NO. Of
Copies
To
1
1
5
1
1
1
1
90
" HZLhSSIHEL-
2 6 GROUP
Unclassified
3. REPORT TITLE
Baratta, Francis I.
. REPORT DATE
t,
PROJECT NO.
11
D/A 1N542718D387
-1_
November 1966
AMRA TR 66-36
f 6. OTHER REPORT NOfS) (Any otttmr numbmrm thmt mmv bm mmmimyimd
thim rmpott)
"SuhtasV ^5475
u. ABSTRACTv.Two
12
SPCKSOPING
MILITARY ACTIVITY
Picatinny Arsenal
Dover, New Jersey
07801
simple formulae are presented which" predict stress concentration ractors applicable to a two-dimensional symmetrical T-head configuration. This configuration corsists of a deep head joined to a shank by fillet radii. The independent
equations that predict stress concentration factors for the same geometry are derived for two different loading conditions. In one instance, a tensile force is
applied to the shank end of the T-shape. Equilibrium of forces is attained by
supporting the bottom edge of the head section, resulting in the shank section being
pulled in tension. In the second instance, a conpressive load is applied to the top
edge of the head section while the configuration is again supported at the bottom
edge. Thus, only the head section is stressed and in a compressive manner.
Because the analysis is not exact, the magnitudes of
factors resulting from the predictive equations appear to
some ranges of the geometry parameter ratios. Therefore,
application", as it is termed in the text, is recommended
Again, because of the inexactness of the analysis, experimental stress concentration factors are indirectly obtained for the first loading condition and directly
obtained for the second loading condition mentioned above. These data were obtainec
for several geometric ratios jyf^the T-head configuration and compared to the corresponding predicted values.
It was found that the
within a certain range of the
the error becomes excessive,
DD ,'Jffl 1473
UNCLASSIFIED
Security Classification
TT
LINK A
KEY WOROS
LINK B
LINK C
Experimental mechanics
Stresses
Stress concentrations
Stress intensity
Expefimental stress analysis
Photoelasticity
INSTRUCTIONS
1. ORIGINATING ACTIVITY: Enter the name and address
10. AVAILABILITY/LIMITATION NOTICES: Enter any limof the contractor, subcontractor, grantee. Department of Deitations on further disst mination of the report, other than those
fense activity or other organization (corporate author) issuing
imposed by security classification, using standard statements
the report.
such as:
2a. REPORT SECURITY CLASSIFICATION: Enter the over(1) "Qualified requesters .may .obtain copies of this
all security classification of the report. Indicate whether
report from DDC"
"Restricted Data" is included. Marking is to be in accord(2) "Foreign announcement and dissemination of this
ance with appropriate security regulations.
report by DDC is not authorized."
2b. GROUP Automatic downgrading is specified in DoD Di(3) "U. S. Government agencies may obtain copies of
rective 5200.10 and Armed Forces Industrial Manual. Enter
this report directly from DDC. Other qualified DDC
the group number. Also, whe r applicable, show that optional,
users shall request through
markings hsve been used f
jup 3 and Group 4 as authorized.
ii
3. REPORT TITLE: Enter t complete report title in all
capital letters. Titles in all cases should be unclassified.
If a meaningful title cannot be selected without classification, show title classification in all capitals in parenthesis
immediately following the title.
4. DESCRIPTIVE NOTES: If appropriate, enter the type of
report, e.g., interim, progress, summary, annual, or final.
Give the inclusive dates when a specific reporting period is
covered.
5. AUTHOR(S): Enter the nanve(s) of authors) as shown on
or in the report. Enter last name, first name, middle initial.
If military, show rank and branch of service. The name of
the principal author is an absolute minimum requirement.
6. REPORT DATE Enter the date of the report as day,
month, y?ar; or month, year. If more than one date appears
on the report, use date of publication.
la. TOTAL NUMBER OF PAGES: The total page count
should follow normal pagination procedures, i.e., enter the
number of pages containing information.
7b. NUMBER OF REFERENCES: Enter the total number of
references cited in the report.
8a. CONTRACT OR GRANT NUMBER: If appropriate, enter
the applicable number of the contract or grant under which,
the report was written.
8b, 8c, & 8d. PROJECT NUMBER: Enter the appropriate
military department identification, such as project number,
subproject number, system numbers, task number, etc.
9a. ORIGINATOR'S REPORT NUMBER(S): Enter the official report number by which the document will be identified
and controlled by the originating activity. This number must
be unique to this report.
9b. OTHER REPORT NUMBER(S): If the report has been
assigned any other report numbers (either by the originator
ot by the sponsor), also enter this numbers).
(4)
(5)
"All distribution of this report is controlled. Qualified DDC users shall request through
UNCLASSIFIED
Security Classification
'
*