I

AMMRC TR 70-2 AD

~ PLASTIC CONCEkNTRATiON FACTORS

0 IN FLAT NOTCHED SPECIMENS
I:
*OF AISI 4340 STEEL

RALNt PAPIRNO r
THO•RIICAL & APPLIED MECHANk.S RESEARCH LABORATORY AR2
• ," . APRI x

February 1970 j

docuenwt ha be
MThi approved fo# public rilem and sale; Its distribution is unlimited.

Reprwiuccd by Oin
C LEAR
c N GiH 0 UISE
Fol Federal ScelnfIc & Tech nical
Intarm-!:in Springfield Va. 22151

ARMY MATERIALS AND MECHANICS RESEARCH CENTER

Wtertown, Massachusetts 02172
AMMRC TR 70-2

PLASTIC CONCENTRATION FACTORS IN FLAT

NOTCHED SPECIMENS OF AISI 4340 STEEL

Technical Report by

RALPH PAPIRNO

February 1970

D/A Project 1T061102B32A

AMCMS Code 501B.11.855
Research in Materials
Agency Accession No. DA 0C4686

This document ha been approved for public relee and sale; Its d~.tribution is unlimited,

THEORETICAL & APPLIED MECHANICS RESEARCH LABORATORY

ARMY MATERIALS AND MECHANICS RESEARCH CENTER
Watertown, Mamchunetts 02172
 ARMY MATERIALS AND MECHANICS RESEARCH CENTER

PLASTIC CONCENTRATION FACTORS IN FLAT NOTCHED SPECIMENS

uF AISI 434u
.. TEEL

ABSTRACT
Various authors have proposed methods for predicting the plastic behavior at
the root of a notch under monotonic loading. Among thcse itia method by Neuber,
which was originally developed for shear but which haL been empirically applied,
at Neuber's suggestion, to tension and compression loading. There has been only
a limited confirmation of Neuber's method in tests of notched specimens. Addi-
tional confirmation is given in this report for a range of notch geometry.

The basis of the Neuber approach is the suggested rule that the geometric
.an of the stress and strain concentration factors, when the root of the notch
is plastic, is given by the theoretical elastic concentration factor:

(KKE )1/2 = Kt.

The Neuber rule is evaluated using an appropriate analytic representation

of the stress-strain curve of AISI 4340 steel and predictions of maximum notch
strain versus nominal net section stress are developed, The theoretical results,
when compared with test data from flat notched specimens of the same material
with a range of initial elastic concentration factors, show agreement within 5%.
It is shown that the limitations of the strain gages in measuring the notch root
strains can account for a major part of the discrepancy.
 CONTENTS

ARWR•ArT

LIST OF SYMBOLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

INTRODUCTION ................. ............................... .I...

NEUBER FORMUIATION . .......................... 2

ANALYfIC STRESS-STRAIN RELATIONS FOR AISI 4340 :TECL ...... ........... 3

-NOTCH STRESS AND STRAIN ANALYSIS ............... .................... 3

A. Analytic Formulatinns .................. ...................... 4
B. Calculation of Theoretical Results ........... ............... 8

EXPERIMENTAL PROCEDURE ............... ............................. 9

A. Material ........................ ............................ 9
B. Stress-Strain Tests ....................................... 10

C. Notch Specimen Preparation and Testing .... ............. .... 10

EXPERIMENTAL AND THEORETICAL RESULTS ........ ................... .... 12

A. Mechanical Property Data ................ .................... 12
B. Notch Specimen Data ........... ....................... ... 13
C. Experimental Accuracy ................. ...................... 13
CONCLUSIONS .................... ............................... ... 14

ACKNOWLEDGMENTS .................. ............................. ... 14

LITERATURE CITED ............... ............................. .... 15

APPENDIX A. COMPUTER PROGRAMS ................. ...................... 16

 LIST OF SYMBOLS
in.2
Anet - not section area,
a w exponent
b - notch •pecP• half Eross-widthe b srw/2, in-
C general coefficient, strain-stress
g strain power law
Cp
P coefficient; C : Cp in plastic region
Ct coefficient; C Ct in transition region
D - coefficient, strain-stress power law
d a notch specimen net-section width, in.
E a modulus of elasticity, psi
Kt v theoretical elastic concentration factor
KF a strain concentration factor, elastic or plastic
Ka 0 stress concentration factor, elastic
or plastic
notched semi-infinite piate
K. - elastic concentration factor for a
L - notch length, in.
m w exponent; p w m in transition region
n - exponent; p - n in plastic region
P a applied load, lb.
power law
p - general exponent, strain-stress x strain
q a exponent, strain-stress power law
r - notch root radius, in.
t P specimen thickness, tn.
w a notch specimen gross width, in.
a a exponent
8 - exponent
c a strain, in./in.
c - elastic strain, in./in.
in./in.in./in.
strain,strain,
C0 nominal notch
Cn m maximum net-section
: • p plastIC strain, in./in. (See Eq 8)
•Pproportional. limit strain, in./in.

t "transit.onal strain, in./in. (See Eq 7)

aay 0.1% offset yield strain, in./in.
a a 3Lress, psi
Ce a elastic stress, psi
psi
on a maximum notch stress,
co - nominal net-section stress, psi; o0 P/Anet
op - plastic stress, psi (See Eq 8)
ot - transitional stress, psi (See Eq 7)

ii

II I n I I- - - -" -T
-- I I .
"I
"VL-- - • "•
 INTRODUCTION
Metheds nf pvAirtni the plastic behavior in notches and other discontinu-
ities under cyclic loading have been developed by adopting static prediction
methods for low cycle fatigue. A study by Stowell, 1 of the plastic concentration
factors around a hole in a puit- un.... static loading, gwar
... eby !-iAv-th
and Ohman 2 to include various other geometric discontinuities and then applied by
Crews and Hardrath 3 to cyclic loading. Kuhn and Figge,4 analogously building on
earlier work of Nouber, 5 were also able to arrive at a scheme for predicting the
strength of notched parts under cyclic loading. As another example of this pritc-
ess, Wetzel, 6 using a later formulation of Neuber 7 on plastic concentration
factors under monotonic loading, was able to develop a method for relating the
conditions in a smooth specimen to those in a notched specimen under cyclic load-
ing. The aforementioned examples are not meant to be an exhaustive survey of the
field but have been presented as illustrative examples of a particular approach
to cyclic, plastic behavior of elements with stress concentrations. These ap-
proaches have in comnon a particular sequence of analysis:

a, Modification of the conventional, monotonic loading elastic concentration

factor to take into account plastic behavior in the notch or other discontinuity.

b. Experimental verification of the derived plastic concentration factor

under monotonic loading conditions.
c. Modification of the plastic concentration factor for cyclic loading to
obtain fatigue concentration factors.

d. Experimental verification of the values of the fatigue reduction factors.

A crucial step in this described sequence is the development of a well-

founded method of predicting plastic concentration factors under monotonic loading
for later application to low cycle fatigue. In monotonic loading, the plastic
concentration factor formulation is evaluated using the virgin stress-strain curve
of the material. This formulation then becomes the basis for fatigue behavior
predictions when the cyclic stress-strain curve is substituted for the virgin
stress-strain curve to develop the analytic results.

The method developed by Neuber 7 for predicting plastic concentration factors

is attractive since it can easily be adapted to machine computation. Because it
has had only limited experimental confirmation, the study described here was
undertaken to assess its predictive value for monotonic loading, prior to applying
the theory in low cycle fatigue. This report describes a combined analytical and
experimental investigation with the followinig major objectives:

a. To refine the procedure of plastic concentration factor prediction for

monotonic loading using an appropriate analytic representation of the virgin
stre'--strain curve of AISI 4340 steel.

To perform experiments on notched tension specimens of AISI 4340 steel

with a range of initial elastic concentration factors for comparison with theoret-
ical prediction of the plastic concentration factors resulting from monotonic
loading.
 NEUBER FORMULATION
The basis of Neuber's approach is the suggested rule that the geometric
mean of the stress and strain concentration factors is the theoretical elastic
concentration factor:

(K K)/2 1(1)

where K•0 on
t 0 nd k Cn/uo for plane stress.

Although the original formulation of Neuber's rule was developed for mono-
tonic loading in shear, Neuber has suggested and there has been some experimental
evidence to show that it may also apply to tension or compression loading.
Krempl 8 presented plastic strain and stress concentration factors for notched
specimens (nominal Kt - 3) of carbon steel, 2.5 Cr-l Mo alloy steel, and type 304
stainless steel from which geometric means could be computed. The calculated
discrepancy between the geometric mean of the plastic stress and strain concen-
tration factors and the theoretical elastic concentration factor averaged less
than approximately '± 5%. These results should not be considered conclusive since
the precise values of Kt for each of the individual specimens was not reported
and data points were taken from charts. Nevertheless, Krempl's results were suf-
ficiently good to warrant further pursuit of the Neuber approach.
It is possible to rewrite Equation 1 by applying the definition of strain
and stress concentration factors referred to the nominal net section stress and
strain as

(CYnEn/ C0 C) (/2 t (2
(2)
mmn 0 0 t
or (a0 )Kt 2 (ynen). (3)

"The left-hand side of Equation 3 refers to the nominal net section stress and
strain (subscript 0) while the right-hand side refers to conditions at the root
of the notch (subscript n). In order to determine K ow K from Equation 3, it
is necessary that the equation be expressed in terms of stress for the former or
in terms of strain for the latter. This is most easily done by the use of a
power law of stress-strain behavior such as

E = Doq. (4)

It is shown later in this report that calculations ure facilitated if Equation 4

is transformed into

G C(os)P (5)

where C
(l(q + 1)] and
D p = q/(q + 1).
In the next section of this report, the stress-strain properties of heat-
treated AISI 4340 steel are given in the form of Equation 5 by using a curve-
fitting procedure on test data.
 i
I
ANALYTIC STRESS-STRAIN RELATIONS FOR AISI 4340 STEEL
.. , -I... -- . fcrn tcatzdAh s iateg
rithmic coordinates reveals two linear regions in addition to the elastic region
as shown schematically in the upper graph of Figure 1. The linear region between
the proportional limit strain '1) and the 0.1% offset yield strain t~y has been
designated here as the trwnwritionaZ region and corresponds to the knee of the con-
ventional stress-strain curve shown in the lower graph of Figure 1. The region
where strains are in excess of this yield strain has been designated here as the
pZaaeti region.

] f. .
•plplastic Plastic
logeE Strews a Conventional

EPI
- ----- Transitional

Elastic .
log (Q.e0 I(Lool--
-- Ev
Strain
Figure 1. Log stress-log strew-strain and strew-strain curves for AISI 4340 steel (schematic)

All three regions can be represented by equations of the form given in (5):

Elastic (1/E)1/2 (ae)1/2 (6)

Transitional Ct
W C a
(aEt)m (7)

n
Plastic = C (o F) (8)
p p pp
A typical log c- versus log (oc) plot for one heat of heat-treated AISI steel
is shown in Figure 2. An automated data reduction procedure using a least-squares
analysis was applied to obtain the coefficients and exponents shown in the figure.
This procedure involved a number of steps: (1) autographic recording of the engi-
neering stress-strain curve; (2) automatic analog to digital conversion of the
data on punched tape; (3) tape to card conversion; (4) computer data reduction
using a specially written program which included linear and logarithmic least-
squares analyses.

NOTCH STRESS AND STRAIN ANALYSIS

Referring to Equation 3, it should first be noted that the G0 and E0 are nom-
inal values referring to the net section. The net section conditions are defined
as follows: o0 = P/Anet and E0 is given by an equation of the form of Equation 5.

]3
 /I

Data Points in Transitional

and Plastic Regions

•;0.01

Plastic Region
e- 7.52 x 10-6a E 0.96-2

"Transitional Region
4 42
E= 1.39 x 1O- (Ce)°'s

- .. "'I e =30.0Region
x 10'
e 1.82 x 10-4 (OC)°'
5 00
SElastic
0.0 0l 1 1 t1
1 1 1 _L
3 104
102 10
Stress x Strain, (oi), psi
Figure 2. Log strons-long stress-strain data for one heat of heat treated AISI 4340 steel
"showingmaterial constants for three regions
The net section strain conditions will depend upon whether the net section
stress is elastic, transitional, or plastic. Referring to (3), it is possible to
recognize six cases of net section (right-hand side) and notch (left-hand side)
conditions for any given notch geometry as determined by the value of the elastic
concentration factor Kt. These are enumerated below:

Case Net Section Notch A. Analytic Formulations

No, Condition Condition
A given specimen with a predetermined value
12 Elastic
Elastic Elastic
Traiisitional of the elastic concentration factor will pro-
gress through a number of cases as it is loaded
3 Elastic Plastic monotunically to failure. The actual proares-
4
4 Transitional
Transitional Transitional
Tansitio "
sion will depend on the magnitude of Kt and on
5 Transitiol Plastic. _ the toughness of the test material. A partic-
ular specimen need not progress through Ll the
cases as it is loaded to fracture. For example, a specimen with high Kt manufac-
tured of a material with only moderate toughness may progress through cases 1, 2,
and 3 only. With higher toughness the progress may be through 1, 2, 3, 5, and 6.
Other combinations of toughness and Kt would lead to different progressions.

4
 It is assumed in the following development that plane stress conditiors ii,
ISJLI
aA A 'D VL..A A jJL CV .. .A.
L L
AI La i~fl
AA ii~ t t LL&Sl Lttig L~j1i1i[jig uu r

Case 1: The elastic conditions of case I follow the familiar stress concentration
factor relations where:

a) K. Kf . K
a r t

b) u = Kto and n Kt
K

In the subsoqucnt development of cases 2 through 6, the notch strain and

stress values will be obtained by suitable substitutions of relations (6), (7),
or (8), into (3). Then the appropriate plastic strain and stress concentration
factors will be given. Each of the cases is considered separately below.

Case 2: Elastic Net Section - Transitional Notch (c <E ; En )

a) Notch strain: Since the nominal net section is elastic the righL-hand
side of (3) may be expressed in terms of strain and the elastic modulus:

UanEn (n2E)K2 (9)

Now (9) can be combined with (7) and solved for c

E = C (02K2E)m• (10)
n t 0t
b) Notch Stress: The right-hand side of (3) may be expressed in terms of
stress and the elastic modulus since:

- a~ /E (1
0 0
then
a E (o 2 /E)K 2 (12)
nn 0 t,
The right-hand side of (10) may also be expressed in terms of stress

En Ct(og/E)M Kt2m. (13)

Now (13) is substituted into (12) and solved for i

a "n (1/ct 2/E)(1-m)Kt(2-2m). (14)

c) Strain Concentration Factor:

By definition the strain concentration factor is given by

 KE = niloU•-

ý,ubstitutioij of (10) inLu (15) i,•ult6 is

K mCCK
% (2m-l)
KF. = Ct(KtE)M~ (16
(16)

d) Stress Concnntration Factor:

Analogously to (15) the stress concentration form is

K an/a. (17)

Substitution of (14) into (17) results In

C8
K, - (1/Ct) (IK2E) ('-m) (y )3(1-2M)

Case 3: Elastic Net Secion - Plastic Notch (E <PI ; y>cy)

This case is directly analagous to case 2; however, the material properties

fur the notch are described by (8). The equations can be written by inspection
using (10), (14), (16), and (18) as guides.
a) Notch strain (analogous to (10)):
C a•nn
C (E2 K E)n( )

b) Notch Stress (analogous to (14)):

a = (1/Cp)(a2/E)(l-n)K (2-2n). (20)

c) Strain Concentration Factor (analogous to (16)):

C(K2E)(n), (2n-1) (21)

d) Stress Concentration Factor (analogous to (18)):

KE a (/C p )(K 2t0
/E) (i-n), (0 )(1-2n). (22)

Case 4: Transitional Net Section - Transitional Notch (EP <E0<cn-<y)

a) Notch Strain: This is quite simply obtained by substitution of (7) into

both sides of (3), resulting in

(C /C )1/m . K2(r /C )1/m (23)

6
whirth can he Teduced to

C K~m E0 (24)
|
H t

b) Notch Stress: Substitutions of (24) into (3) re.eults .n an expresston

for notch stress:
SK2-2-2M, (25)
n t 0

c) Strairn Concentration Factor: This is obtirnod ,1irectly 'rom (24):

KC K2m (26)

d) Stress Concentration Factor: This is obtained directly from (25):

K K2-2m. (27)
a t
Case 5: Transitional Net Section - Plastic Notch (cE
p<E0 <E y; Cy <En

a) Notch Strain: Determination of notch strain in this case is made by

substitution of (7) into the right-hand side of (3) and (8) into the left-hand
side. The result, after simplification is:

En .KK2nCp(E0/Ct) n/m (28)

b) Notch Stress:

By suitable algebraic manipulation of (7) and (8) and subsequent sub-

stitution into (28), the strain factors can be transformed to stress with the
following result:
a, K2 .n(i/C )Ca (29)
n t p to0

where a = (l-n)/(1-m).

c) Strain Concentration Factor: Substitution of (28) into (1S) results in

Kr K2 nC (1/c )n/m (n-m)/m. (30)

C t p t E0

d) Stress Concentration Factor: Substitution of (29) into (17) results in

(2-2n) N ((-I
K -K (i/C )(C
(t)(- (31)
St oto
Case 6: Plastic Net Section -Plastic Notch ( E<

Y
14

This case is directly analogous to case 4 and the various relations can be

Srange.

a) Notch Strain:
: - K2n E (32)
A n t

b) Notch Stress:

'-V .• K -2n, . (33)

n t
c) Strain Concentration Factor:

K - K•2 (34)
t.
d) Stress Concentration Factor:

0 M K 2 - 2 n. (35) S
B. Calculation of Theoretical Results

Laborious computations are required to evaluate the stres:: and strain history
of a particular specimen as it is monotonically loaded and progresses through the
various cases. A computer program was developed to perform the calculations for
any specimens of a given Kt value, which properly discriminates the correct pro-
gression through the various cases and eliminates those which are unnecessary.
The program yields the notch stress and strain values as a function of net sec-
* tion stress and strain, the plastic concentration factors, and the required loads
for a given net section area. A listing of the program and a typical printout
are given in Appendix A.

As an added convenience, a program for computing the elastic concentration

factor Kt can also be developed and combined with the plastic program. Such a
program for a flat tension specimen with semicircular notch ends was developed
using separate formulations for deep and shallow notches given subsequently in
this paper in (43-46). The combined program is also given in Appendix A.

In the experiments, described in the next section of this report, notched

tension specimens were loaded and an autographic record of net section stress ver-
sus maximum notch strain was obtained. The theoretical values of the two param-
eters were obtained from the computer program directly without the necessity of
using an explicit relation between notch strain and net section stress of the
form:

-0 f(cn). (36)
For completeness, however, the explicit relations between net section stress
and notch strain have been developed for each of the six cases and they are listed
below in (37-42).

m•8
tCasel I a0 = (1/Kt>)v (37)

case 2: Go= Ei~L(Cn /CtK )I/Zm for r ,l< <.y (38)

n/ 'KttfrF 1
p En ,y (9

Case 4:CO= 1/ n p n pin

for yl<f Fn (40)

Case 4: ("O= (El/K,% 2mC /m)Lflm)/m for

F <E <F~
L 1 (40)

where 8-[l-n(l--mfl/m(1-m)

Case 6: c0 (I/Kt(2)((1f/Ci/ for EF -r Fr (42)

EXPERIMENTAL PROCEDURE

Externally notched flat tension specimens with the notch configurations shown
in Figure 3 were fabricated from two lots of AISI 4340 steel plate. These were
K, gsiven identical heat treatments before
~ ~.L. pecimen fabrication. Electrical resist-
ance strain gages were installed at the
0.6001,000 1.50 roots of the notches and the specimens
were monotonically loaded to fracture, The
nominal net section stress (converted from
0.39
0002applied
± load) and the notch maximum strain
R 0002were
0.95 autographically recorded up to approx-
* imately 2% strain. Monotonic loading was
continued until fracture and the fracture
load was recorded. Stress-strain proper-
ties were obtained from standard flat
tension specimens using strain gages and
0.800 1.000 1,59 clip-on e'ctensometers. The stress-strain
data were recorded autographically and the
antographic records were analyzed Using an
R =0,32 ±0.002 automiated data reduction process. The de-
tails of the procedure are described in
this section; comparison of the experimen-
tal results with the theoretical predic-
tions is given in the next section.

O6U1.000 2.00 A. Material

Material properties specimiens and
R - 0.159 ±0.002 notched specimens were fabricated from two
separate heats of AISI 4340 steel plate,
Figure 3. Notch confifi irstions und notch dimensions received in the annealed condition. Lot
of 0.1-inch thick flat notiwed specimens. No. 1, used for one notched specimen and
one smooth tension specimen, was reculveu ua C.5-•,h-t. .....
2 was received as 0.75-inch-thick plate. Th6 chcmical analyses nf both lots are
grvpr hvInw together with the heat treatment details:

Chemical Analysers (wt %)

C Mn P S Si Ni Cr Me Fe
Lot No. •, 0.40 0.74 0.003 0.004 0.22 1.88 0.87 0.25 Remainder
Heat 3931362
Lot No. 2, 0.39 0.80 0.005 0.006 0.23 1.77 0.78 0.26 Remainder
Heat 3830298

Heat Treatment (Applied Mechanics Research Laboratory Designation: A-16)

Austenitize at 2300 F, 1 Hr; Furnace Cool to 1550 F

Oil Quench to R. T.; Hold 15 min.
Double Normalize in Salt at 1650 F, 1 hr; Air Cool
Reaustenitize in Salt at 1550 F, 1 hr
Oil Quench to R.T.; Hold 15 min.
Quench to Liquid Nitrogen Temperature
Temper in Salt at 920 F, 1 hr
Water Quench to R.T.

The as-received material was cut into blanks from which one or more specimens
could later be prepared and heat treated in its full thickness. The thickness was
then reduced to 0.10 inch for specimen preparation.
B. Stress-Strain Tests

Standard flat + tension specimens (2-inch gage length, 0.50-inch wide, and
0.10-inch thick) were tested to obtain stress-strain properties of the heat
treated material. These were loaded in a Tinius Olsen hydraulic testing machine
and strains were measured either by electrical resistance strain gages or by a
clip-on extensometer. The data were recorded autographically on a X-Y recorder
whose axes were calibrated for each specimen to read strain and stregs directly
(rather than load and extension). The resulting stress-strain data w-re auto-
matically converted to digital form and subsequently were analyzed by computer
using least-squares analyses: Linear, in the elastic region to obtain V; and
linear-logarithmic in the nonelastic regions to obtain the material constants and
exponents required for the experimental approximation of the stress-strain
properties.
C. Notch Specimen Preparation and Testing

For the design of the notched specimen shown in Figure 3 with Kt values of
1.5, 1.59, and 2.00, the following relations, empirically derived by Heywood, 9
were fnployed:

Kt = [(Z/r)/(l.55(w/d]-l,3)] (43)

10
 7I

where

./A
14 I q Q/,.r11/211r(wl/d
to _ I + f(/r)/2 (44)

(See Figure 4 for Identification of notch parameters.)

These relations were incorporated into a computer program. In the interest

of completeness, a formulation for specimens where Kt>2 was also included in the
program. Baratta and Nea1 1 0 using a prior formulation of Bowiele showed that the
following relations are appropriate for deeper U notches:

Kt = [ + 0.182(t/b) 1.071(t/b) 2 + l.727(Z/b) 3

][I (k/b)]K. (45)

where

K = 0.775 + 2.243(P/r)1/2 (46)

After manufacture, the specimens were carefully measured and the actual
stress concentration factors for the notches were re-evaluated using the appro-
priate formulas. Because of manufacturing tolerances the values of Kt computed
from actual dimensions could depart from the nominal values by several percent.

The basic notch specimen design is

shown in Figuzi 5 for one notch configura-
tion. This basic design was used for all
the notches shown in Figure 3. Specimen
blanks were cut from the as-received mate- td ,w-2b
rial, heat treated, and then reduced in
thickness to 0.10-inch by machining equal 9
amounts of material from each surface.
The various holes and contours were ma-
chined into the final thickness blank. Figure 4. Identification of notch parameters

±0.010
8.00 -- -
±1132
±0.005
±1 1/3 . .,
tO.001

F 2-1/8 - -
400001020

R- Se 35D/8.70

.2.500
FiguJre 5. Notch specimen design and dimensions.
Notch configuration dimension given in Figure 3.

11
±000
 Strain gages, .i'ype .L. ^40-
00.. 1ing andA n n.-inch wide. were
cemented at the notch roots, one in each notch in individual specimens. These
were eletri•nA1v connected in series so that bending com'1ponents of strain were
eliminated and so that the longitudinal strain reading obtained was Ohi avriage
for the two notches in each specimen. The specimen gages formed one arm of a
Wheatstone bridge with compensating gages on a dummy specimen forming an opposite
arm. The remainder of the bridge consisted of precision resistors. The unbalance
bridge voltage was recorded on a Hewlett-Packard X-Y recorder calibrated to read
'O.01-in./in. strain per one-half inch of pen displacement along the recorder
X-axis.
The byidge energizing voltage was held to approximately two vults. This lim
ited the power dissipation of the gages to less than 5 watts/sq in., a sufficient-
ly small value so that excessive heating of the specimen in the notch roqt was
avoided.
Loads were recorded on the Y-axis of the recorder. By taking the specimen
area into account it was possible to calibrate the recorder to read net section
stress directly on a scale where 10 ksi was the equivalent of one-half inch of
pen displacement. The resulting autographic recording showed net section stress
as a function of notch strain.

EXPERIMENTAL AND THEORETICAL RESULTS

In a test of a notched specimen it is possible to measure the maximum strain

at the root of the notch and to determine the net section stress from the net sec-
tion area and the applied load. Theoretical values of these two parameters can
also be developed. It is assumed that a comparison of the theoretical and experi-
mental values of the two parameters will constitute a valid test of Neuber's
hypothesis although the hypothesis itself is stated in slightly different terms.
A. Mechanical Property Data
The reduced material property data for the two separate heats of material
used for specimen manufacture are given in Table I. The constants given for tran-
sition and plastic regions of the stress-strain curve are those applicable to
Equations 7 and 8.
Table I. STRESS-STRAIN DATA FOR TWO HEATS OF HEAT-TREATED AISI 4340
STEEL OBTAINED FROM A LEAST-SQUARES, CURVE-FITTING ANALYSIS
Elastic
Heat Modulus, Epl Ey Transition-Region Plastic-Region
No. 106 psi m
inCt Cp n
1 30.0 0.47 0.60 1.39x10_4 0.542 7.52xi0- 6 0.962
2 29.6 0.53 0.65 0.65x10- 4 0.655 7.S0xlO- 6 0.962
The values of Epl and E were calculated from the fitted curves. The propor-
tional limit strain a l has been taken at the intersection of the elastic line
with the curve repzes nting the transition region. The yield strain value, Ey,
represents the intersection of the transition region curve with the plastic region
curve.

12
 drr

a IZ " a4

al

InI
uIE I~A11
ZD-P

ILL_
Z- a.~_
~S tjjg

2-
z
r

En to

C)~ z 0

a LA

20 z

-1 r

z ~-J

u CC

o I-

VI~1j
Is
iid 20W
Ij, tth
t
I~gL
13 -5 o
B. Notch Specimen Data -

Experimental and theoretical net Ipimun i2

section stress and notch strain data for aK
The figure shows fairly good agreement be- 1o- Limit
tween experiment and theory up to approxi- j ximum 0.014 in./ln.
mately 1.4% strain which is typical for Theory. Discrepancy
all specimens tested. A summary of the +2.5%
(omparison between experiment and theory Experiment
for all the tests is given in Table II
below. The agreement in the elastic re-
gion was within two percent.
z
In each case the experimental values
of notch stress for a given notch strain
were greater -:han the theoretically pre- 10J
dicted values. Beyond the indicated 10s 10-2
strain limit values shown in the table Maximum Notch Strain, en, ln./in.
there was a much larger discrepancy which
is interpreted as an effect of mu].tiaxial Figure 6. Typical experim•ntal aid theoretical data
stress and resulting notch strengthening. forKt"1.5
It is not altogether clear whether the
discrepancies shown in Table II are also a result of multiaxial stress effects
or result from inevitable variations in heat treating in the separate batches
which were used in the program. With the exception of one specimen, the theory
appears to be conservative by approximately 5% up to about 1.5% notch strain.
However, certain possible errors, discussed below, could account for the
discrepancy.

C. Experimental Accuracy

Load errors (and hence net section stress erors) were neglible since the
testing machine was calibrated just prior to the testing program using proving
rings whose own calibration was traceable to the Bureau of Standards.

Table II. COMPARISON BETWEEN EXPERIMENTAL AND

PREDICTED NET SECTIUN STRESS-NOTCH STRAIN DATA

Specimen Heat Kt Discrepancy Strain Limit

No. No. Exp. vs Theory in./in.
A-1-1 1 1.59 <1% 0.0200
DI-21-4 2 1.50 +4.0% 0.0140
DI-21-2 2 1.50 +2.5% 0.0140
D13-12-9 2 1.50 +4.8% 0.0140
D1-31-.1 2 2.00 +6.5% 0.0140
D13-15-5 2 2.0 +5.5% 0.0160
D13-13-9 2 2.0 +4.5% 0.015

13
 ,ii

J Major sources of error in the strain measurements resulted from the following:

1. The manufacturer's stated ±3% uncertainty in the value of the gage factor.
1

2. The presence of a strain gradient at the notch root with the maximum
strain value confined in an area smaller than that of the strain gage.

3. An estimated possible ±0.01 inch deviation of the position of the center

line of the gage from the center of the notch root.

The magnitude of the latter two sources of error is not known, however, both
would tend to produce strain readings which were less than the actual maximum
strain in the notch.
Some uncertainty in the results arises from the fact that it was not poss1ble
to heat treat the stress-strain specimens and the notch specimens all in the same
batch because of the limited capacity of the heat-treating facilities. Tests indi-
cated that there could be a variation in the computed notch strains of ±1% based
upon scatter of the material properties.

CONCLUSIONS

The major conclusions of this study are as follows:

1. Predictions of plastic notch strain and notch stress can be developed

using Neuber's rule and an analytic representation of the stress-strain curve in
three regions: elastic, transitional, and plastic.

2. Theoretical predictions of net section stress versus plastic notch maxi-

mum strain are within S%, on the average, of experimentally observed values for
notched specimens of heat-treated AISI 4340 steel up to a maximum notch strain
value of approximately 0.015 in./in.

ACKNOWLEDGMENTS
A number of staff members of the Theoretical and Applied Mechanics Research
Laboratory contributed to the investigation described herein. Mr. John Campo per-
formed the majority of the laboratory tests; Mr. John Hannon delicately applied
the strain gages in the notch Toots and contributed to the development of the
automated data analysis program; Mr. Joseph Wong, a summer student aide, assisted
with the computer programming and with the calculations. I am grateful to all who
contributed and am happy to acknowledge their contributions.

14

2|~~~
 LITERATURF CITED

Infinite Plate. NACA TN 2073, February 1950.

. iiAKI)RATH, 11. F., and OHMAN, L. A .YdUy of" ,'/4rtia• rnd PTanti;ie ,trinao Gion c(n-
tration Factoro Due to Notcher and Fillnts in Flat lI'aten. NACA Report 1117,
1953.

3. CREWS, J. H., Jr., and HARDRATH, H. F. A Study of Cycle PiKatic St•,es at

a Notch Root. Experimental Mechanics, v. 6, no. 6, June 1966, p. 313-320.

4. KIIHN, P., and FIGGE, I. E. Unified Notch-Strength Analynie for Wrought Alu-
minwn Alloys. NASA TN D-1259, May 1962.

5. NEUBER, H. Theory of Notch StreseHs: Principles for Exact Strese Calcula-

tion. J. W. Edwards, Ann Arbor, Michigan, 1946.

6. WETZEL, R. M. Smooth Specimen Simulation of Fatigue Behavior of Notches.

T & AM Report No. 295, Department of Theoretical and Applied Mechanics, Uni-
versity of Illinois, Urbana, Illinois, May 1967.

7. NEUBER, II. Theory of Stress Concentration for Shear Strained Pri.-matioal

Bodies &zithArbitrary Non-Linear Stress-Strain Law. Trans. ASME, .1. Appl
Mech, December 1961, p. 544-550.

8. KREMPL, E. Low-Cycle Fatigue Strength Reduction in Notched Flate Plates.

ASME, Preprint 67-Met-13.

9. HEYNOOD, R. B. Designing by PhotoeZasticity. Chapman and Hall, Ltd, London,

1952, p. 162-166.

10. BARATTA, F. I., and 1-"AT, D. M. Streae-ConcentrationFactors in U-Shaped and

Semi-Elliptical Shaped L4d.' Notches. AMRC TR 70-1, January 1970.

11. BOWIE, 0. L. Rectangular Tensile Sheet with Synvnetric Edge Cracks. Army
Materials and Mechanics Research Center, AMRA TR 63-22, October 1963.

15
 AEDIMATY A -. CQM.fPUTEr PRaORAMS

The programs are written as Fortran I

1. Program No. 1. Plastic stress aud strain in notched tension specimens with a
given value of the elastic concentration factor and given material properties.
The input data required are given below:
Parameter Fortran Designation
Elastic Concentration Factoz, Kt X
Modulus of Elasticity, E, psi E
Transitional Exponent, m Q1
Plastic Exponent, n Q2
Transitional Coefficient, Ct Cl
Plastic Coeffi,:ient, CP C2
Proportional Limit Strain, c.1, percent EPL
0.1% Offset Yield Strain, cy, percent EY
Program Cut-Off Notch Strain, percent EMAX
Specimen Net-Section Area, sq in. AREA
Specimen number is entered using up to 11 alphanumeric characters.
A listing of the program is given on page 17.

2. Typical Output of Program No. 1. A typical output run of Prograff, No. 1 for
a Lpecimen of AISI 4340 steel with and elastic concentration factor, Ktml.S, is
given on page 19.

3. Program No. 2. Elastic concentration factor and plastic stress and strain
in notched tension specimens with given notch dimensions and given material
properties.
This program combines a program for computing the elastic concentration
factox using either of two methods (Heywood method for shallow semicircular
notches or Bowie method for deep semicircular notches) with Program No. 1.
In addition to the material property parameters listed as input for P-ogram
No. 1, the following input data are required for the combined program:
Parameter Fortran Designation
Specimen Width, w, in. W
Length of Notch No. 1, Ri, in. Al
Length of Notch No. 2, Z2, in. A2
Radius of Notch No. 1, rI, in. R1
Radius of Notch No. 2, r 2 , in. R2
Specimen Thickneps, t, in. T
The method of computation is designated by the entry of NN in the 4th state-
ment of the progra, lists given on page 20. For NN = 00, the Heywood method is
used; for NN u 01, the Bowie method is used. The yords "Heywood" or "Bowie" are
entered in the aeme statement. The specimen amber is entered as described pre-
viously for Program No. 1.

1r
 B0OCOLUMN PRINTOUT OF PROGRAM NO. 1

2~ PRINT il. !1ES1w.Pi p4,P5.p6.P7,pl.pl,8.PIOP1I

2 A
PRINT 102,ELPL.L T.1MAX
PRINT 103.CI.O1,C2,02
PRINT II2.X.ARIA
PRINT 13
PRINT IN4
PRINT 15
PRINT 16
7
PRINtI 202

17 Fo:-SO/F

lF IFN -EP L1 14S. I1,300

P * So.ARE A
SOP-sO01000.
S Npop.SNm/
I1U00.
E NP- ER 1U00
-- _ _ PAJ5-j P~jE I WL.$OP , f P 4NP
50-504 10000.
GO TC' 17
30o PRINT 102

EN-C 1l('X'*2)IEO"*2 i'EI"01

N; Fj-EY IliI I19,00
19 SN.(X0R214(tOO#2141/EN

XE-EN/EQ
P..50AR(A

SNP-sN/1000:

E0-E0~ 0. 000 1
IFIEO-EPLI 301.301.500
400 PRINT_402L.. _______
401 50:E0*E
"IFEN-CMAX 20.2.
2c IE4I
S1N-IIJXGE0I"so20E

XE-EN #00

50$ -SOft 000w

EOPsE 04100.
SNiP.SN/LOGD
PNI E ?212020..EOP.SOP.ENP.SIIP.XS.XE
-
E .1OtEO+0000 1
FIEIO-EPLI 4Q1.401.600
Soo PAIN4T .502
F 501 SO-f boicnv**t1./OR 1.900
E%-EOXKe'12.#QI1I
2
EF(EN-EVI 1,21,A00
21 SN-I(EOCSOISFX#II2I/EN

XE -EN/CO_____________

CU a/$O0003

IFIeo-Efl 501,501.700

17
 600 pRINT 602
601 1N.C24XS? vVt)IO( 'IJ/I
IF IF'f-EMAXI 22.22.1
e~ It IE11
I

It-tN/Eu
P. SO*AM LA
DIP-S0O'1 Q0U.
[P.N': :100.

PRINT Ž22sP.EOP.SOP,.j;FPIýSNP,.KS.XC
(0-ECO+0.,000 2
lPF(o-EyI60l .601.700
10 PNINT 703

IP(LN-P.MiX) 23.23,1
23 5O"ItEO/C2#*'I I./02fl lEG

,
Sn
50E
IS.

SLW.S0/IOOO.

SIP-SN/1000.

PRIT?Z2P.,E0P.SO$'.ENP.SN4P.XS.KE

13 TORMATI2f+X26HPRLDICTEG NOTCH PROPERTILS,#/l

14 CflMAII TX RAPL.L,EMSlIN;.7X3HIiJH.Il0ýdJ4Ag.TiAlh.A,.BXIHK,
1 711HKMI
15. rfORMAT I&XAHLOAV,5MISPINA6STESHSTSTAI.X6S MESS.

16 r04MATIX6NPaUNO5,&,X7HPERCENT.3X3HKSI,7KX7HPENaI(4T.3X3K3.I.gA1Hs,
101 FONM4ATIZF9.bi
1* FONNATi //.5K,8HLS.P&.2.6IVLe1066,3K .AHEPL..FI.&,IXIHtEY. ,PB.6.lX,
I S4MEAA..FI.b.i/)

__ rw112mz fll .fl.)AIk\. .773.Sx.ThiliIflr.MxTh"E..77.5,

202 POAN4ATI1.,2OXISNELASTIC-LLAStIC./I
282 FORtMAT(712.6 .F1O.5,F9.1,F1O4.P.7. 1 3XFJ.4.2X.P7.4i)
302 FONNATI /.ZOKZ0hELASTIC-TRAANSITIONAL,/I
402 PONMAYI/,2OK1SHELASkTIC-PL.ASTIC,/)
5 02 FONM4ATI/,92OKZSHTIANSITIZONAL-TRANSITIONAL,/I

702 V011t4AfI/.ZOX1SHPLASTIC-PLA.STIC./I
777 PORNATiI1.11AII
77@ PORNATI//.IXASHELASTIC. TRANSITIONt, AND PLASTIC ANALYSIS OF
I 11HS'ECI'QEN NOvl1A1I/I)
069 FORMAT I 2FT.S,2E11.S,3F7.4.Ff&.2 I

999 STOP
END

18
 TYPICAL OUTrPUT OF PROGRAM NO. I

CS.
.100-0 ~ owli~r .5000 o" 2 11. *33fiI? ) .92%

AWL III oiw o o - wit "AN-.

6 0 S l

Pou..s PSSO?4? 965791 0

It All I t-lA'?IC

50.80on 1013 11.0 .3070140.

300S.f0000
Doa 51
a 0. 70.0::al,
.21 "
,
3520.0,0,8 21 50.0 .3047 9.71
42994 o00 .256 .3, 7 110.1
S31%, 119191 1 04001 407.0 .4 1 331.0

It ASTIr- 7*4140.311051

59003947 .3962 23. .6050 356&4.0 I 1 30

751.799 .4035 33.9 .52m.9 300.4 .35314
1110. 39 .42352 12135.01 " I34 3.406511 i.%L)S

1467.9915 .%7167 121.4 .01,74 1 134.1 3.350 3.5094

055n1.91 I .43552 A327.5 1..k79 3147.6 1 53 1.54" a

5032.34134 .S95?16.2 .03394 2157.! .2 11

9263.9`18% .0335? 30. .10371.3 3.31%o2 3.9S26

19*411 313903A-PLAt00I7f

?
S3047 1317 505
62.06 3.034 300?11.1 13.,l344 "011.

9414.9,041 .3435 2 20.. 2.21374 376.5. mTo

2.05 5as
308?651 .6311 35. 3.3577 379I.4 .13230

905"433
. 1, .539 74.2 1.5101q,5 50. 3. I,1311 2.33
IM1010.2911 .54351 514.5 3.00893 160.5 3.0I321231

5044.477 .335 v31.0 3.941:s 3s.0.0 1.0133 .33

5005.553
.0052 30.21.53463 35.0 nI3I
I.33 7

10142.0950 .53352 310S. 12.7l3IIl.7 153 .0331 0.3137

31003.7101 .63316 31. 3.13 55.4 .
305 2..3I537
30051.57 .643I5a2 31.15.80 ,,,.3837363. 3.33
IDII
30002.344 ::,1,
835 3.0a 36. 3.21 2.33
305.520 ...... 315.4 .47-6. .33333

1061,1.14 1.0.995 377.2

. .5 6. .33253

3251'.0331. 5.035. 37..220"X6. 113 233

20604.1 3.3332 323.9 6. s.77 .11233

506831.92041 3.53352 3181.0 2.15886 3613.16 203 233
 80 COLUMN PRINTOUT OF PROGRAM NO. 2

eran 71

s" vI Y.

'I, 14-Al
I 0
11? 1 -D Y .f

flI.

[. N7A A0 I I01.1 -O 4

11:14 Ia. AI2AIl0.l, ~ 1TARI..3Ill.-AIIIl

~ f t ol.A Il -6

air jo0. lrm

WRIN i-i.irr;.A2OZ

RINfT
Pon 307

11 lr
10 *'61 In* Fr

1%P%. oi so 1.1.11

S: P .Ut,0I90

10:1[00IfJ

400 r[tiI loopl

IF1 11.1)

CIO.11111II

slop t".4F1 0

[V0410 . 0."UOI

firgro-tvis 48Stl too~U

200 Pll O
 Smol, I.,I nfaVII f1 0 1la

60 i rad
f:ir 1.S
I 3.1.2 111ann
; 1 -Z
re I -,) . ;-w
* i ae7

xr -reic o

$OP:-%aIfin MD
fnr:(O-Inn.
%MP: S4eI 10DO.

PO INTI Ij.PI(O.OP.[r,(op. S11. Is.tfr

fO:rgUj* rw'5

;'jO PfT'S! SO?7

If If(i-CpAII 220's.1

P'0: I)
%O.A1
(N0:fqpo. I rr.

(PowfliW

:N/"
Inc 11,11141mTo Q

I' :I[O-CYIOI.&(hl. TO

14!' S-uAlM j.fp-

FAIR:(O. In''.

Tfl:F.O.1. '4'0
604T1 70 T

1A"Vl0.N-AVO :V&.S
I?1 I1OR"&TI P1 AI0I A944417 Af(A.lotSFIggNIOS11.111M
AO Ws

IVTINS)1/
, 1!N 54LOA.%ICM-LASS Ib IC,
16Z ifOPPl1 I£A f¶SNNTIIN'36S1

jo? VOxWATI/.1nIII(~ICf*TCV

go? r Lf%
vwjr,.lI 4g C-0% FCTR11te/l V ME

IT'S FOWNITI/A-.11ft44f4?f INI?6.&OPATI NL SO

r 4D5(I~b 0IS,/

Ats rOw~IItI
Ill (4.14*?4172.
 Iis.
rDOCUMENT
'tipi I I-, 11¼.C...... II, 1-d. dv h.f ac d
Pbt~fI
CONTROL DATA
eun~,~rallcl
R&D
-- ff1 he oniI.,d whe,,ten
110, REPORT
- repelr 1.
IRCUF1"I TVY C L4101IF~f
. fI.
1.e
A.T1111
I (JNIGINATIC.G .ACTIVITY (Corporaff -tih-i)

IArmy Mater i als and Mcchanics Research Center

Wterinwn, Maqqavhuasetts 02172 [2b.
Ithcl as si fied
G04OUP

1 r PON r Y IT L'W'
A1A W N E -- T
.R O .... O.. 11 F'TlK"I 0 1' A~'
AIS 1 340 STEELT

4 fllRS NIP TIV; NOY S (r'Yp* of report and Inch..i- J4104)

9 AU THORMS (Firefllname, middle WHOII~. 1-1nio-e')

Ralph Papirno

70. TOTAL NO. OF PAGES IL. No. o- NKs

ma
I REPONT
February 1970
OATS
____24 I i
i.CON TfqAC T ON GRANT NO. 5.OIIAOIRPR UURS

6. ,PqojrrT HO. Ti/A 1T061102B32A A1MNRC TR 70-2

AWNIS Code 501B.11.855 Wb. OTHER REPORT #N0(s)(AroWG~he' nsu0nN1& Sal may be asaiudro

Id- Agency Ascession Number DA 0C4686 ____________

10l OI&TRINUTION STATEMENT

This document has been approved for public release ariJ sale; its distribution IS
unlimited.
II SPPLEENTARY NOTES 12.. SPONSORING MIL11AIR' ACTIVITY

U. S. Army Materiel Command

Washington, D. C. 20315
Various authors have proposed methods for predicting the plastic behavi,.r
at the root of a notch under monotonic loading. Among these is v method by Neuber,
which was originally developed for shear but which has been empiý:ically applied, at
Netiber's suggestion, to tension and compression loading. There has been only a lim-
ited confirmation of Neuber's method in tests of notched specimens. Additional
confirmation is given i~n this report for a range of notch geometry..

Trho basis of the Neuber approach is the suggested rule that the geometric
mean of the stress and strain concentration factors, when the root of the notch is
plastic, is given by the theoretical elas-ic concentration factor: (K K )1/2 = K

The Neuber rule is evaluated using an appropriate analytic represoec:.ation

of the stress-sl'ra-'k curve of AISI 4340 steel and predictions of maximum notch strain
versus nominal iset section stress are developed. The theoretical results, when com-
pared with test data from flat a range of notched specimens of the same material with
initial elastic concentration factors, show agreement within 5%. It is shown that
the limitations of the straini gages iii measuring the notch root strains can account
for a major part of the discrepancy. (Author)

DD NO .1473MYUS. UNCLASSIFIED
• : S e c u r tiq , .I mia fir ~ti on
.= N¢'Y~~Kf
WORf|• l L f?"4 N . .LIN4N
. ..a I-NK r7
MI. TE WT M0OL.E MT 0 LE
K 11

Plast.,. deformation
Stress concentration
Mechanical properties

iUNCLASSI FIED
Seeurity Clamtuffcatoion