Stress Constrate Factor Plastic
Stress Constrate Factor Plastic
Stress Constrate Factor Plastic
DSpace Repository
1996-03
Long, Kenneth S.
Monterey, California. Naval Postgraduate School
http://hdl.handle.net/10945/32178
THESIS
by
Kenneth S. Long
March, 1996
19960806 014
REPORT DOCUMENTATION PAGE Form Approved OMB No. 0704-0188
Public reporting burden for this collection of information is estimated to average I hour per response, including the time for reviewing instruction, searching existing
data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate
or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information
Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork
Reduction Project (0704-0188) Washington DC 20503.
1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED
March 1996 Master's Thesis
4. TITLE AND SUBTITLE EXPERIMENTAL STUDY OF INELASTIC 5. FUNDING NUMBERS
STRESS CONCENTRATION AROUND A CIRCULAR NOTCH
6. AUTHOR(S) Long, Kenneth S.
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION
Naval Postgraduate School REPORT NUMBER
Monterey CA 93943-5000
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING/MONITORING
AGENCY REPORT NUMBER
11. SUPPLEMENTARY NOTES The views expressed in this thesis are those of the author and do not reflect the
official policy or position of the Department of Defense or the U.S. Government.
12a. DISTRIBUTION/AVAJLABILITY STA'IEMENT 12b. DISTRIBUTION CODE
Approved for public release; distribution is unlimited.
13. ABSTRACT
Experiments were conducted to determine whether energy -density relations can be used to predict
elastic-plastic stresses and strains near a circular notch for 7075-T6 aluminum alloys and ARALL 4
composites. The loading conditions were tension and four-point bending. Glinka and Neuber have
developed relations that predict local inelastic strain response based on the stress-strain solution for
small plastic zone sizes. It has been shown that these relations are appropriate for ·simple tension and in-
plane bending, where stress and strain are uniform through the thickness. This study investigates the
application of the Glinka and Neuber relations to samples where stress/strain is not constant though the
thickness. Nonconstant stresses/strains are the result of out-of-plane bending and laminate
characteristics.
14. SUBJECT TERMS Stress, strain, stress concentration factor, ARALL, 15. NUMBER OF
composite, tension, bending, plasticity PAGES 90
16. PRICE CODE
17. SECURITY CLASSIFI- 18. SECURITY CLASSIFI- 19. SECURITY CLASSIFI- 20. LIMITATION OF
CATION OF REPORT CATION OF THIS CATION OF ABSTRACT
Unclassified PAGE ABSTRACT UL
Unclassified Unclassified
NSN 7540-01-280-5500 Standard Form 298 (Rev. 2-89)
Prescribed by ANSI Std. 239-18 298-102
ii
Approved for public release; distribution is unlimited.
Kenneth S. Long
Lieutenant, United States Navy
B.S. Petroleum Engineering, University of Texas at Austin, 1989
from the
Author:
Approved by:
R. MeNelley, Chc:tiJfllrtan
Uepal~t!Jl~ntof Mechanical Engineering
iii
iv
ABSTRACT
circular notch for 7075-T6 aluminum alloys and ARALL 4 composites. The
loading conditions were tension and four-point bending. Glinka and Neuber
have developed relations that predict local inelastic strain response based on
the stress-strain solution for small plastic zone sizes. It has been shown that
these relations are appropriate for simple tension and in-plane bending,
where stress and strain are uniform through the thickness. This study
characteristics.
v
vi
TABLE OF CONTENTS
I. INTRODUCTION .......................................................................................... 1
II. LITERATURE SURVEY ............................................................................... 3
III. MATERIAL PROPERTY VERIFICATION ................................................ 7
A. 7075-T6 ALUMINUM ALLOY ......................................................... 7
1. Material .................................................................................... 7
2. Experimental Apparatus ...................................................... 7
3. Results ...................................................................................... 9
a. Young's Modulus ....................................................... 12
b. Poisson's Ratio ............................................................ 13
c. Nonlinear Coefficients ..... .... .. ... ..... .. .. .. ............ ...... .. 13
4. Statistical Analysis ................................................................. 14
B. ARAMID ALUMINUM LAMINATE ........................................... 16
1. Material .................................................................................... 16
2. Experimental Apparatus ...................................................... 16
3. Results ...................................................................................... 19
a. Young's Modulus ....................................................... 19
b. Bending Modulus ...................................................... 24
c. Poisson's Ratio ............................................................ 25
IV. TENSIONTEST .............................................................................................. 31
A. EXPERIMENTAL APPARATUS .................................................... 31
B. THEORY OF ENERGY-DENSITY METHOD ................................ 35
1. Based on Ramberg-Osgood Relation ................................. 35
2. Based on the Bilinear Relation ........................................... 38
C. METHOD OF COMPUTATIONS .................................................... 39
1. 7075-T6 Aluminum Alloy ......................................... .. ........ 39
2. ARALL 4 Composite ............................................................. 39
D. RESULTS ............................................................................................. 40
1. 7075-T6 Aluminum Alloy ................. ... ........... .................... 40
a. Theoretical Circular Notch Strain ................ ...... .... 40
b. Stress Concentration Factors ..................... .............. 43
2. ARALL 4 Composite ............................................................. 46
vii
V. FLEXURAL TEST ........................................................................................... 57
A. EXPERIMENTAL APPARATUS .................................................... 57
B. BEND TEST PROCEDURE ............................................................... 60
C. THEORY ............................................................................................... 60
1. 7075-T6 Aluminum Alloy ................................................... 61
2. ARALL 4 Composite ............................................................. 61
D. RESULTS ............................................................................................. 62
1. 7075-T6 Aluminum Alloy ................................................... 62
2. ARALL 4 Composite ............................................................. 66
VI. CONCLUSIONS AND RECOMMENDATIONS ..................................... 73
LIST OF REFERENCES ............................................................................................. 77
INITIAL DISTRIBUTION LIST .............................................................................. 81
viii
I. INTRODUCTION
notched sample is the relationship between the load, the nominal elastic
stress, and the elastic-plastic notch tip strain. Material failure is often initiated
Because solutions for local stresses and strains in the inelastic region are
cyclic loads. Approximate methods have been developed that generate good
inelastic notch tip stresses and strains. The foundation of this method is that
the energy density in the elastic range is nearly the same as the energy density
in the inelastic range. Previous authors have applied this method to several
(e.g., circular, elliptical, V, keyhole, etc.). They also considered tension, in-
plane bending, and cyclic loading. The stress and strain in these tests were
1
In this study, experimental testing was conducted on 7075-T6
nonconstant stress and strain through the thickness was created. To analyze
• For small scale localized plastic yielding around a notch, the inelastic
• A uniaxial stress state was considered at the vicinity of the notch tip
direction.
both tension and out-of-plane bending. The material properties for each
material will be verified and/or established for this study. For each loading
type, the experimental setup will be described and the theory used to predict
2
II. LITERATURE SURVEY
NOTCHES
for the stresses in a semi-infinite plate, with a circular hole mid-way between
infinite series form that were used to plot the stress variation around the
hole. Each term of the series was based on the ratio of hole diameter to plate
In 1939, Dumont [2] tested a large plate containing an open circular hole
by subjecting the plate to uniform bending normal to the plane of the plate.
He found that stress concentrations occurred at the edge of the hole. Further,
stress became constant at a distance of 2.5 times the hole's diameter from the
In 1952, Hill [3] stated that in practice the radius of curvature of the
notch root must be two or three times the plate thickness if the theory of
numerous plate geometries. Related to this study was the plot of stress
containing a circular hole. He did not plot the stress gradient ahead of a hole,
3
but rather the stress concentration factor at the edge of the hole for ratios of
In 1990, Prasad and Shuart [5] developed a closed form solution for the
stresses and strains near notches was developed by Neuber [6] in 1961. More
recent studies have shown that Neuber's rule can lead to considerable over-
In 1968, Hutchinson [8] showed that the strain energy in the plastic
zone ahead of a crack tip is the same as that calculated based on the linear
strain behavior. In other words, the energy density at the notch root does not
elastic material.
In 1981, Molski and Glinka [9] used Hutchinson's energy density theory
4
and made the assumption that the theory could be applied to notches and
In 1985, Glinka [10] applied the energy density method for a uniaxial
inelastic notch tip strains for several notches and materials. This study uses
the uniaxial stress state with monotonic loading and applies it to tension
density method by adding a factor Cp that accounts for the increase in plastic
zone size due to local yielding. This factor becomes important when strains
estimating elastic-plastic notch stresses and strains for a multiaxial stress state
in the vicinity of the notch tip. This method applied to both monotonic and
cyclic loading.
In 1988, Glinka, Ott and Nowack [13] used the equivalent strain energy
under a plane strain condition. Results were compared with FEM solutions.
5
In 1992, Sharpe, Yang and Tregoning [14] conducted an in-depth review
of the Neuber and Glinka relations for monotonic loading. They determined
that in general, the Neuber method was best suited for a plane stress
provided an upper bound of stress. For large plastic deformation regions, the
In 1995, Lee, Chang and Wong [15] improved Hoffmann's and Seeger's
notch stresses and strains were estimated bas-ed on elastic notch stress
solutions for in-phase or out-of-phase loading. The authors also reported that
in general, the Glinka and Neuber methods underestimate the notch strain in
the high yield deformation region and overestimate the local strain in the
6
III. MATERIAL PROPERTY VERIFICATION
This chapter describes the preliminary tests and results that were
and establish material properties for the material provided because that
1. Material
The aluminum alloy used in this study was cut from one sheet of
7075-T6 aluminum. This aluminum sheet metal was also identified by the
dimensions of the large sheet were: 54 in. (137.2 em) x 46 in. (116.8 em) x 0.125
2. Experimental Apparatus
Tension tests were conducted on four test coupons to verify the 7075-T6
design and the testing methods used were governed by American Society for
Testing of Materials (ASTM) standard test methods [16]. Figure 1 shows the
dimensions of the tension test coupons used for material verification. The
methods listed in [16] were used to determine the material properties of yield
7
strength (offset=0.2%) and ultimate tensile strength. The testing methods
used to determine Young's modulus and Poisson's ratio were listed in [17]
L
A
~
'"
~[' w
, '\.I/
v '
'\.
G "
/
Digital Micrometer, accurate to 0.00005 in. (0.001 mm). The 90 degree strain
gage rosettes used were Measurements Group Inc. type CEA-13-125UT, 120
ohm, gage factor 2.085 + 0.5% (longitudinal) and 2.115 + 0.5% (transverse).
The Instron dynamic extensiometer used had a 3.5 in. (8.9 em) gage length
and a + 0.2 in. (0.51 em) range. The Instron Model 4507 tension/ compression
8
test machine was used with a 200 KN load cell. Load and displacement data
were obtained using the Instron Model 4500 data acquisition tower with the
gage readings were acquired through the Measurements Group P-3500 strain
indicator /SB-100 switch and balance unit combination. These two units were
used to acquire all the strain gage readings in following tests. The equipment
set up used for this and subsequent tests is pictured in Figure 2. All data
recorded from the tests was manually entered into MATLAB programs that
3. Results
material properties with published material properties [19] for bare 7075-T6
data and calculations from three of the four test coupons. Three test coupons
strengths and ductility in the short transverse direction, can be less than those
in the longitudinal direction [20], one test coupon was cut from the sheet in a
9
Figure 2. Test Equipment Setup.
10
----·~~~~------------------------------------------------------------------
11
a. Young's Modulus
tested three times in the elastic range by cyclic loading, as per Reference 21.
The modulus values calculated for each of the three runs were averaged,
resulting in an average modulus value for that test coupon. The Young's
modulus used in this study was the average modulus value of the three test
During the tensile tests, strain data was collected from two
sources. One source was a strain gage rosette mounted on the geometric
center of the test coupon. The other source was an extensiometer. Young's
modulii were calculated using stress and both strain data sources. As before,
modulus values calculated for each run were averaged to result in a modulus
value for each test coupon. Averaging the longitudinally oriented test
of 10,269 ksi (70.6 GPa). The average longitudinal Young's modulus based on
strain gage data was 10,121 ksi (69.6 GPa), a 1.5% difference. Reference 22
describes the standard method of recording strain data with the use of
subjecting them to impact forces that occur at ultimate failure, they were
removed just prior to sample failure. The only means of obtaining strain
12
b. Poisson's Ratio
strain gage based data was also used for determining Poisson's ratio. The
Instron data acquisition system and automated testing software computed the
ultimate tensile strength. Yield strengths (0.2% offset) were taken from stress-
c. Nonlinear Coefficients
(j (j ...!..
E=- +(-)" (3.1)
E K
n were experimentally determined from the results of the tension tests used
for material property verification. For most metals, a log-log plot of true
stress versus true plastic strain is modeled as a straight line [25]. However, the
relationship often deviates from linearity at low strains and/ or high strains
[26]. The slope of the line is nand the intercept with the log true stress axis is
K. Figure 3 shows the plot of log true stress versus log true plastic strain for
sample 2. Similar plots were seen with samples 3 and 4. K and n values were
calculated for samples 2, 3, and 4 and averaged. It is obvious that the data is
13
not linear throughout the plastic range, but it is satisfactory to calculate K and
n for the strain range over which the log-log plot is linear. Therein lies the
question of which region of the plot should be considered linear. Two cases
of log true plastic strain between -4 and -2.4. Case 2 considers the data points
of log true plastic strain between -3.2 and -2.4. A least squares approximation
of the data for both cases is shown in Figure 3. The results were K = 110.13 ksi,
n = 0.0790 for case 1 and K = 114.88 ksi, n = 0.0946 for case 2. It will be shown
later that the case 2 linear approximation for K and n will provide slightly
better results. Thus, the average K and n values used in this study are 114.88
4. Statistical Analysis
Correlation coefficients (r) were calculated for each run and averaged for each
test coupon. The test coupons' averaged coefficients were then averaged to
reliability of the curve fit to the data and a perfect curve fit has a coefficient
stress-strain plots in the elastic region was 0.99979. When the data is basically
linear, as was the case in these tensile tests, the correlation coefficient is not a
very sensitive indicator of the precision of the data. McClintock found that
14
11.45.-------.--------.-------,,-------.-------~------~
11.4
11.35
(/)
(/)
~ 11.3
Ci5
Q)
2
l-
g>11.25
-'
11.2
0
f - - Case 1 Linear Approximation
- Case 2 Linear Approximation
0
0
11.15 0
0
0
0
11.1U,L________ L ._ _ _ _ _ __ . J ._ _ _ _ _ _ _ _J....__ _ _ _ _ __J,__ _ _ _ _ ____L.._ _ _ _ _ __ _
-5 -4.5 -4 -3.5 -3 -2.5 -2
Log True Plastic Strain
Figure 3. Log of True Plastic Stress-Strain Behavior for Sample 2. Strength
Coefficient K and Strain Hardening Index n were determined using Case 2.
15
I
I
(l-r
2
)
2
is a better indicator of the curve fit quality [27]. Here, (l-r 2 ) 2 =
0.0146. This value indicates that the vertical standard deviation of the data is
only 1.46 percent of the total vertical variation caused by the straight-line
1. Material
configuratio n. There were five layers of 2024-T8 aluminum alloy and four
laminate was symmetric and balanced with respect to the test direction. The
2024-T8 aluminum lamina were isotropic and all the aramid laminas were
2. Experimental Apparatus
Four samples were initially cut from the sheets provided. Two
aluminum samples used for the material verification tensile tests (see Figure
1). One of these samples had the fibers oriented parallel to the loading
direction and the other had the fibers oriented transverse to the loading
16
transverse modulii. The other two samples were machine finished to a
rectangular shape of 1 in. (2.54 em) by 9 in. (22.86 em). These samples had
fibers oriented only in the longitudinal direction and were used in flexural
The equipment used for the laminate material property tensile tests
was the same as that used for the 7075-T6 aluminum material property tests.
Figure 4 shows an ARALL sample undergoing a tension test. For the flexural
was the load configuration, which is a typical loading condition for high-
modulus materials. The support span was 5 5/8 in. (143 mm) and the load
span was 2 13/16 in. (71.5 mm). The sample length was 7 1/4 in. (188 mm)
and laminate thickness was 3/32 in. (2.4 mm). Sample length was based on
gage, accurate to 0.001 in. The strain gages used were Measurements Group
Inc. type CEA-13-250UN, 350 ohm, gage factor 2.12 + 0.5%. For both tensile
and flexural tests, the Instron test machine was used in the manual load
17
Figure 4. Tension Test of an ARALL 4 Sample.
18
Load
Load Span
Support Span
3. Results
The reference used for the tensile test to determine Young's modulus
and Poisson's ratio was ASTM 3039 [29] and the reference used for the flexural
a. Young's Modulus
modulus. This was the modulus for the laminate rather than a modulus for
each of the different lamina. One sample was used to calculate E 1 (Young's
modulus in the fiber direction) and the other was used to calculate E 2
(Young's modulus transverse to the fiber direction). Two strain gages were
19
placed on each sample. One gage was oriented in the fiber direction and the
other gage was oriented transverse to the fiber direction. Although each
sample was tested to ultimate failure, only data in the linear elastic range was
used to calculate modulus values. For the sample with fibers oriented
curves in the longitudinal and transverse directions for the case of fibers
oriented transverse to the loading direction. Figure 8 shows the plot of stress
versus strain for longitudinal and transverse directions in the linear elastic
range. A MATLAB program was written to solve for the modulii in a least
squares sense. The best fit approximation of the slope of the data in the elastic
region for Figures 6 and 7 was the modulus value. For the longitudinal
modulus, E 1 = 9,100 ksi (62.6 GPa) and for the transverse modulus, E 2 = 6,600
ksi (45.4 GPa). These modulii show that a sample with fibers oriented
modulii (tension) for 2024 aluminum and Kevlar 49 fibers is made. Because
was not available. The published elastic modulus (tension) for 2024
aluminum is 10,500 ksi (72.4 GPa) [31] and for Kevlar 49 fibers is 18,100 ksi (125
20
80~----------------.-----------------~----------------~
/
/
70 /
/
I
/
/
60 /
/
/
I
I
50 I
,......,
·u;
~ I
I
~ 40
...
-Q) I
I
C/)
I
30 I
I
I - - Longitudinal Direction
20 - - Transverse Direction
I
I
10 I
I
QI
0 0.005 0.01 0.015
Strain
Figure 6. Stress-Strain Behavior in Longitudinal and Transverse Directions
for Fibers Oriented Longitudinal to Loading Direction.
21
50
45
I
I
40 I
I
35 I
I
I
30
~
~25
-
._
Q)
( /)
20
15
10 I
I I
I I - Longitudinal Direction
- - Transverse Direction
5 I
I ,
I
0~----L-----L-----L-----L-----L-----L---~L---~L---~
0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018
Strain
Figure 7. Stress-Strain Behavior in Longitudinal and Transverse Directions
for Fibers Oriented Transverse to Loading Direcion.
22
45r--------.---------.--------.-------~---------.------~
40
Longitudinal
35
30
........
~25
en
en
Q)
.; 20
(J)
15
0~------~~------~--------~--------~--------~--------~
0 1 2 3 4 5
Strain
23
GPa) [32]. Although the elastic modulus for Kevlar 49 is high, the lamina did
not consist of a 100 percent volume fraction of Kevlar 49 fibers. These lamina
volume fraction of Kevlar 49 fibers much less than 100 percent. The low
modulus values greatly depend on the volume fraction of the fibers within
the lamina. It follows that E1 = 9,100 ksi (62.6 GPa) is a reasonable laminate
elastic modulus.
b. Bending Modulus
The flexural test method was used to determine a smeared bending modulus
repeatability of the data. Since the results for each sample were nearly
identical, only one sample's data will be presented. Figure 9 shows the plot of
load versus deflection in the elastic range. For the test, E 8 = 11,200 ksi (77.0
GPa). The bending modulus was calculated using
24
(3.2)
where L =length of support span [in. (mm)], b =with of beam [in. (mm)], t =
the different stiffnesses of the lamina. The 2024 aluminum lamina are stiffer
than the aramid fiber lamina. The outermost lamina of the composite were
the 2024 aluminum lamina. The stiffer outermost lamina results in higher
c. Poisson's Ratio
The strain data taken from each sample used in the Young's
modulus section was used to determine Poisson's ratio. Two Poisson's ratios
longitudinally to the loading direction, while v 21 is for the sample with fibers
25
140~----~----~------~----~----~------~----~----~
120
100
80
0
40
-20L-----~----~------L------L----~------L------L----~
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
Deflection [in.]
Figure 9. Load-Deflection Response for Flexural Test in Elastic Range used to
Determine Bending Modulus.
26
v 12, plotted versus load in the elastic region. The average value is drawn
vavg = (3.3)
n
where Eli is the longitudinal strain, Eti is the transverse strain, and n is the
total number of data points in the elastic region. This calculation was used to
aluminum. This compares nicely with the calculations for v 12 . Since the
transverse stiffness, the laminate Poisson's ratio was essentially due to the
The relation
= (3.4)
can be used to predict v 21 , since the other values are known. The calculation
Poisson's ratio, v 21 , versus load in the elastic region is shown in Figure 11.
Averaging the ratios results in v 21 = 0.270, which is 15 percent higher than the
27
predicted value. This difference can be attributed several factors. One factor is
related to gage factor errors and load readout errors. Another factor is that the
However, the predicted v value does fall within the range of data point
21
expected, Figures 10 and 11 also show that Poisson's ratio is higher for fibers
28
0.35~--------~--------~---------.--------~--------~
0.345
0.34
0.335
0 0.33
~
cr:
0
§ 0.325 0 0 0 0 0
en
en o 0 0 00 0
oo oo 00 oOoo
0 0
·c; 0 0
a.. 0.32
0.315
0 Avg Poisson Ratio = 0.323
0.31
0.305
0.3~--------~--------~--------~--------~--------~
0 500 1000 1500 2000 2500
Load [lbs]
Figure 10. Poisson's Ratio in the Elastic Range for Fibers Oriented
Longitudinal to Loading Direction.
29
0.29~---,r----.----~----~----~----~----~----~----~
0.28
o 0 oo
O oo
0 0
0
0 0 0 0
0 0 0 0 0
0.27 0 0 0
o o0
0
0 0
0
~0.26
a:
c:
0 0
en
en
·o
c..
o.2s
0.23
0.22~0~--~----~----~----~----~----~----~~----~--~
0 200 400 600 800 1000 1200 1400 1600 1800
Load [lbs]
Figure 11. Poisson's Ratio in the Elastic Range for Fibers Oriented Transverse
to Loading Direction.
30
IV. TENSION TEST
Tension tests on circular notched samples were conducted for both the
purpose of the tests on the 7075-T6 aluminum was to verify Glinka's previous
discusses the equipment apparatus used for tension testing, the theory of the
computations for each material, and the results of the comparison of Glinka's
experimental stresses and strains. Test results indicated that the Glinka and
sample but not for the ARALL 4 sample. Good theoretical inelastic strain
methods.
A. EXPERIMENTAL APPARATUS
samples used for the tension tests. The length of the 7075-T6 aluminum
sample was 15 in. and the length of the ARALL 4 sample was 12 in. Both
31
TOP VIEW
~ I [J
single element strain gage (AI only)
L . . . . - -_ _ _ _ _ _ _ _ .
~
~118 in- 7075-T6 AI
3/32 in - ARALL 4 SIDE VIEW
dia=5/8in ~ 0
BOTTOM VIEW
Figure 12. Circular Notched Sample Dimensio ns and Strain Gage Locations.
32
sample types had machine finished edges. The circular notch was milled
through in the 7075-T6 aluminum sample and drilled through in the ARALL
sample. The circular notch was located in the geometric center of the sample.
Also shown in Figure 12 is the location and orientation of the strain gages.
The 10-element strain gages used were Measurements Group Inc. type CEA-
09-020PF, 120 ohm, gage factor 2.00 + 3.0%. The single element strain gages
used were Measurements Group Inc. type CEA-13-032UW, 120 ohm, gage
factor 2.11 ± 1.0%. The two single element strain gages were positioned back-
to-back only on the 7075-T6 aluminum sample. Having the two single gages
positioned back-to-back and midway between the outer edge of the hole and
edge of the sample permitted the comparison of strains on both sides of the
sample. Greater strains on one side of the sample would indicate the
tests were conducted on the Material Test System (MTS) Model 810
tension/ compression test machine. The load cell cartridge was rated at 25
kips and the maximum deflection cartridge was rated at 1 inch. The MTS test
machine was selected for testing because the grip width was 2 in., which was
wide enough to extend across the entire width of the circular notched
Figure 13. For the 7075-T6 aluminum sample, the length of the sample
gripped by the MTS machine was 2 in. on each end. The ARALL 4 sample
gripped length was 1.25 in. on each end. The MTS machine was manually
33
Figure 13. Tension Test of a Circular Notched Sample.
34
load controlled. Using a 10,000 sec/in. loading speed, the loading was
periodically paused to allow recording of load and strain data. Samples were
E=E+E
e p = (4.1)
calculated as
e_
!
ws = I sijcteij
0
(4.2)
where sq__ is the local elastic stress tensor and e q.. is the local elastic strain
tensor. For localized small-scale plastic yielding, the energy density in the
plastic zone is nearly equal to that in the elastic zone. Thus, the energy
elastic solution
35
e. E.
w s = (J
'
w or f sijdeij = f' (jijdcij (4.3)
0 0
where cr .. is the local inelastic-plastic stress tensor and E .. is the local elastic-
~ ~
(4.4)
where Sy is the local elastic stress component in they direction, Kt is the stress
concentratio n factor, and Sn is the nominal stress. For uniaxial stress at the
(4.5)
This states that elastic strain energy density W s at the notch tip is equal
to the product of the strain energy density wsn due to the nominal stress sn
(4.6)
With localized yielding at the notch tip, the energy density should be
(j2 (j (j .2..
_Y +-y-(-y) ". (4.7)
2E n+l K
When the nominal stress Sn has exceeded the proportiona l limit, equation
36
(4.6) takes the form
(j2 (j (j ..!..
_Y +-y-(-y)". (4.8)
2E n+1 K
Equations (4.7) and (4.8) enable one to calculate the inelastic stress cr y
and strain E at the notch tip for given nominal stress S and stress
y n
(4.9)
the sample.
stress crN. In the elastic region, the energy-density method and Neuber's rule
are the same. However, in the inelastic region, Neuber's rule takes the form
(4.10)
but the right hand side does not represent energy density. crN is the Neuber
37
2. Based on the Bilinear Relation
inelastic notch strains for the ARALL 4 composite are unsatisfactory when the
Glinka and Neuber methods are used. The Glinka and Neuber equations can
composite.
becomes
where Et is the modulus in the upper linear region [ksi] and cr1 is the stress at
the intersection of each linear region [ksi]. Theoretical inelastic notch strains
(4.12)
where £
1
is the strain at the intersection of each linear region.
(4.13)
where crN represents the stress from Neuber's method [ksi]. Theoretical
38
inelastic notch strains for the modified Neuber's method are also found using
(4.12).
C. METHOD OF COMPUTATIONS
For nominal stresses lower than the yield stress, theoretical elastic
strains were calculated using Hooke's Law. Equations (4.8) and (4.10) were
the nominal stresses calculated in (4.9). Because (4.8) and (4.10) are nonlinear,
the bisection numerical technique was employed to solve for the theoretical
the theoretical inelastic stresses from (4.8) and (4.10) into (4.1).
2. ARALL 4 Composite
In the elastic region, strains were calculated using Hooke's Law and
results were obtained when Glinka's and Neuber's methods were modified in
(4.11) and (4.13). Inelastic stresses were calculated from (4.11) and (4.13) and
39
D. RESULTS
strength coefficient K and strain hardening index n can affect the results.
theoretical approximati ons, cases 1 and 2, are shown. In Case 1, K = 110.13 ksi
mentioned in Chapter III and now apparent, case 2 K and n material property
behavior, the experimenta l, theoretical, and stress-strain curves are the same.
The deviation of the experimenta l and theoretical curves from the stress-
range, the results confirm Glinka's results [10] in that, "The difference
between calculated and measured notch strains were smaller than 10 percent
" For the 7075-T6 aluminum sample, the maximum calculated and
measured strain difference was 11.8 percent in the inelastic region. This
experiment confirmed the Lee et al. [15] claim that Glinka's rule
40
100
0
90
0
80
70
..........
60
!(/)
(/) 50
-(])
....
(/')
40
o Experimental
30 - Ramberg-Osgood Approx
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Strain
41
120.----.----.----.----.---~----.---~----.---~--~
0
_,.·
·"" 0
/
/. / 0
/·_,. / 0
100 // 0
/./
/./ 0
///0
//0
/G
80 g·/ -------------
0.7 -----------
c
C/)
60
r
0/
r
40
I
( o
·- ·-
Experimental
Case 1 Theoretical
- - Case 2 Theroetical
- Stress-Strain
20
o~/----~----L----J-----L-----L----~--~~--~-----L--~
0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 0.02
Strain
42
underestimates the notch strain in the high yield deformation region and
Sharpe et al. [14] reports that the Neuber rule is the best single model for
Neuber rule is more conservative than the Glinka rule. For large plastic
deformation regions, the Glinka plane stress model gives an upper bound of
strain. The sample reached ultimate failure at a tensile load of 13,270 lbs; the
factor Kt. The elastic stress concentration factor gradient ahead of the notch
was calculated using Howland's method of coefficients [1]. The ratio of hole
for the ratios of A = 0.1, 0.2, 0.3, 0.4 and 0.5. Thus, the coefficients used in this
concentration factor gradient away from the notch is shown in Figure 17. The
gradient is steepest near the notch and decreases away from the notch.
43
120.---------.---------.---------~--------~--------~
0
·"' 0
100
80
-~
-c
(/)
60
..
40
o Experimental
-·Glinka
- - Neuber
- Stress-Strain
20
44
3~----------------~----------------~--------------~
2.5
~
-1.5
52
0.5
QL-----------------L---------------~~--------------~
1 1.5 2 2.5
x/a
Figure 17. Stress Concentration Factor Gradient Away from the Circular
Notch for 7075-T6 Aluminum Alloy. Individual Strain Gage Locations on
the 10-Element Strip are Indicated by Horizontal Error Bars.
45
Superimposed on the gradient curve are the locations of individual strain
gages (indicated by horizontal error bars) from the 10-element strip gage. The
average stress concentration factor over the strain gage range is indicated by a
circle. Next to the hole, Kt = 2.26 using Howland's method. Integrating the
gradient over the range of the first strain gage's measuring range to find the
average, Kt = 2.04. The stress concentration factor used for the 7075-T6
aluminum tension test results was Kt = 1.90. Although this Kt fell within the
range of the first strain gage, it was slightly lower than the average because the
first gage was located a small distance away from the edge of the notch.
circular notch for two nominal stresses in the elastic range: sn = 34.0 ksi and
Sn = 19.2 ksi. In [10], Glinka's application of the energy-density method
gradient plots. This study found that theoretical inelastic strains based on
shows strain gradients ahead of the circular notch for several nominal
stress is increased, the gradient next to the edge of the notch greatly increases.
2. ARALL 4 Composite
Both Glinka's equation (4.8) and Neuber's equation (4.10), that are used
46
0.012 r-------,-------.------r------r-----...,,-------.
0.01
o Experimental
- Theoretical
0.008 0
~ Sn=34.0 ksi
~0.006
-~
Ci5
0.002
0------0-------:;:;:-----------
v o ----o-·-·-o-··--o---(_)--··-
o~------~--------~------~--------~------~------~
1 1.2 1.4 1.6 1.8 2 2.2
X/P
47
0.02.-------.-------.--------.-------r------~------~~
0.018
0.016
0.014
0.012
''
''
s::
·e ''
-
( J)
0.01
''
''
0.008. ' '
Sn [ksi]
'
0.006
-- - 61.0
......... 53.5
0.004
-·- - -·- ·-·-·-·-·-·-·-· 41.9
34.0
19.2
0~------~------~------~--------~------~------~~
1 1.2 1.4 1.6 1.8 2 2.2
X/P
Figure 19. Strain Gradients Ahead of the Circular Notch for Several Nominal
Stresses in 7075-T6 Aluminum.
48
and the strain hardening index, n. Assuming the stress-strain behavior for
(4.1), K and n must be calculated and applied to (4.8) and (4.10). It is evident
bilinear. Figure 20 shows the plot of log true plastic stress versus log true
plastic strain. The method for calculating K and n for the ARALL 4 composite
was similar to that for the 7075-T6 aluminum. The straight line in the upper
linear region of the curve was used to calculate K = 889 ksi and n = 0.570.
Figure 21 shows the comparison of experimental and theoretical notch
inelastic notch strains. Neuber's method was not shown because it produced
nearly the same overestimation for inelastic strains that Glinka's method did.
Labeled A-B-C is a strain discontinuity for the transition from the linear
elastic region to the inelastic region. Inelastic strains were calculated using
(4.8). If elastic strains are calculated using Hooke's Law, the jump in strain
going from elastic to inelastic behavior is shown as the line A-C. Using (4.7)
in the elastic region, the corresponding jump in strain is shown as line B-C.
The strain discontinuity between (4.7) and (4.8) can be attributed to the
49
Glinka equations produced unsatisfactory results. The stress concentration
factor used in the ARALL tension test results was 2.10. The sample reached
ultimate failure at a tensile load of 8,550 lbs; the corresponding nominal stress
The strain gradients ahead of the notch for several nominal stresses are
shown in Figure 23. The behavior of the gradients as the stress was increased
improved by using the modified Glinka (4.11) and modified Neuber (4.13)
methods for the ARALL 4 composite in tension is shown in Figure 24. These
50
11.5.--------.--------.,--------.---------.--------.------~
11
10.5
C/)
-
C/)
Q)
,_
( /)
Q)
:::J
,_ 10
1- +
C) +
0
...J
+
+
9.5 + K = 889 ksi and n = 0.570
+
+
+
9 7'
8.5~------~--------~------~---------L--------~------~
-7 -6.5 -6 -5.5 -5 -4.5 -4
Log True Plastic Strain
Figure 20. Log of True Plastic Stress-Strain Behavior for ARALL 4 Composite.
The Linear Curve Fit Shows where K and n were Determined.
51
100 /
/
/
0 /
/
90 0
/
/
/
/
0 /
80 /
/
0 /
/
0 /
70 /
/
/
0 /
-c.
"(i)
60
0 /
/
/
/
/
c: 50 /
(/) /
-
"
:::.::::
40
A
/P _.I
.f : B
1/
f!; _.·
/c
' .
'I
If : 0 Experimental
30 / ..
. " ..
/
- Stress-Strain
!/ ... -- Glinka (Eqn. 4.8)
·I,
1/ · ..... Glinka (Eqn. 4.7)
20
·!."
/i ·-·-Hooke's Law
1/
10 .¢
1/
/
0 0.005 0.01 0.015 0.02 0.025 0.03
Strain
52
80~--------~--------~--------~~--------~--------~
70
60
50
30
20 c Experimental
- Ramberg-Osgood Approx
10
0~--------~----------~--------~----------~--------~
0 0.005 0.01 0.015 0.02 0.025
Strain
53
0.012 .----.-------.-------.------,...---~----.....-
0.01
\
0.008 \
\
\
c:
~ 0.006.
C/)
Sn [ksi]
39.2
... '•,
34.1
25.6
·-·- - -·-·- ·-·-·-·- 17.2
8.5
o~------~------~--------L-------~------~------~~
1 1.2 1.4 1.6 1.8 2 2.2
X/P
Figure 23. Strain Gradients Ahead of the Circular Notch for Several Nominal
Stresses in ARALL 4 Composite.
54
100
90
80
70
-
-
'Ci)
~
r:::::
(/)
60
-
"'
:::.:::::
50
40
30 o Experimental
- Stress/Strain
·- ·- Modified Glinka
20 - - Modified Neuber
10
0
/ 0.002 0.004 0.006 0.008
Strain
0.01 0.012 0.014 0.016
55
56
V. FLEXURAL TEST
notch. Flexural testing was conducted to determine whether the Glinka and
Neuber relations established for tensile loading could be used predict local
apparatus used in conducting the tests, the flexural test procedure, the theory
used to determine inelastic bending stresses for both materials, and the test
A. EXPERIMENTAL APPARATUS
The design of the circular notched samples used in the flexural tests
were the same as that used in the tension tests. Sample sizes and dimensions
listed in Figure 12 also apply to flexural test samples. The same 10-element
strip gages were used but the single back-to-back strain gages were omitted.
Also, the Instron test machine used for material property verification was
used for flexural tests. However, in order to take incremental load and strain
data, the Instron was operated in the manual load control mode. A four-
specifications [30] and the dimensions are shown in Figure 22. The
57
Load
58
Figure 26. Flexural Test of Circular Notched Sample Subjected to
Four-Point Loading.
59
B. BEND TEST PROCEDURE
A sample was placed on the base of the beam test apparatus and
symmetrically aligned. Each strain gage of the 10-elernent strip gage was
zeroed. The top of the beam test apparatus, which weighed 9.6 lbs, was placed
on the sample and symmetrically aligned. This condition accounted for the
first data point. Incremental loads were then applied to the sample to obtain
the remaining data points. For the 7075-T6 aluminum sample, the test was
concluded when the sample deflected approx. 2.1 inches, almost touching the
base of the beam test apparatus. At this point, the corresponding load was 365
lbs. The maximum load reached before the ARALL 4 sample deflected into
the base of the bend test apparatus was 410 lbs. Because the ARALL 4 sample
C THEORY
yielding first occurs at the edge of the notch and progresses away from the
notch, towards the edge of the sample. At any location away from the notch,
the amount of plastic yielding at that distance will be the same through the
sample thickness. Just as in tensile loading, the plastic yielding for bending
begins at the edge of the circular notch and progresses outward. However, at a
given location away from the notch, as loading is increased plastic yielding
begins on the surface of the sample and progresses inward, through the
thickness, towards the neutral axis. A small region of plastic yielding is often
60
referred to as localized plastic yielding. For small localized plastic regions,
where nominal bending stress Sn exceeds yield stress cry, the energy density
yielding at the notch tip under tensile loading can be used for bending
(4.7)
The Neuber relation for bending is similar to (4.7) but the n! I in the
(5.1)
6M
s = ---- (5.2)
n e(D-2d)
where M is the bending moment [lbf.in], tis the sample thickness [in.], D is
the plate width [in.] and dis the notch diameter [in.].
2. ARALL 4 Composite
Chapter IV can be also used for bending. The modified Glinka equation is
61
(a~- cr~)
+-~-- for (a <a). (4.11)
2E 2Et y 1
(4.13)
D. RESULTS
Glinka approximation was much closer to the experimental strains than the
Neuber approximation. For low and moderate local inelastic strains, the
Glinka method was an overestimation but in the high yield region, the
curve was shown to identify the departure from linearity. Figure 29 shows
several strain gradients ahead of the circular notch. These gradients are
slightly different when compared to those for the 7075-T6 aluminum samples
in tension loading. The slope of the gradient next to the notch was lower for
bending than tension. As in the tension test, the strain gradients in bending
62
140.---------.----------.---------.----------.---------~
' 0
(y'
a/ . .--·
120 0/1'
0 / . .--·
0 / /
0 /
0 /
100 0 //
0 /
0 /
-
~ 80
0
0 /
0/ /
/
/
/
------
r
0 / / ------------
c
(/)
..
52 60
40 fZJ
I - - Glinka
·-·-Neuber
!JI -
o Experimental
Stress-Strain
20
I!J
;()
OL---------~----------~----------~--------~--------~
0 0.005 0.01 0.015 0.02 0.025
Strain
Figure 27. Comparison of Glinka and Neuber Methods for Determining
Theoretical and Experimental Circular Notch Strain for 7075-T6 Aluminum
in Out-of-Phase Bending.
63
300
250
Cii'200
g
-g
0
...J 150
100
50
0~--------~--------~----------~--------~--------~
0 0.5 1 1.5 2 2.5
Deflection [in.]
64
Sn [ksi]
0.01
.....
..... 95.1
.....
-....
"(ti
en 0.006 66.6
- -·- ·-·-.
·-·-·- -·-·-·- -·-
0.004
-·-·-·-·-·-· 47.6
0.002t_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
19.0
0~------~------~------~------~------~------~~
1 1.2 1.4 1.6 1.8 2 2.2
X/P
Figure 29. Strain Gradient Ahead of the Circular Notch for Several Nominal
Stresses for 7075-T6 Aluminum in Out-of-Plane Bending.
65
Peterson has plotted several curves for stress concentration factors subject to
experimental circular notch diameter (a) to sample width (W) ratio was 0.34.
Peterson plotted the stress concentration factor curves for ~ from 0 to 0.2. If
Chapter IV, the strain gage was located small distance from the edge of the
notch. The strain gage averaged the stress concentration factor gradient over
centering of the circular notch by 0.03 in. (0.76 mm) during the machining
process.
2. ARALL 4 Composite
the elastic-plastic region to predict theoretical notch strains. This was justified
by Glinka's statement in Reference 10, " ... within some range of the nominal
bending stress Sn > cry, only localized plastic yielding can occur in the analyzed
section, whereas under tensile loading plastic yielding of the whole section
would occur." The discontinuity line A-B shows the transition from elastic
66
strains calculated using Hooke's Law and inelastic strains calculated using
low inelastic yielding, the two methods predicted a strain response that
dependence on the strength coefficient (K) and strain hardening index (n)
bending. The stress concentration factor used in the ARALL 4 tests was 1.05.
67
100
90
80
0
70
-::=.
·c;; 60
c:
en 50
"
52 40
30 o Experimental
- Stress-Strain
- - Glinka
20 ·-·-Neuber
·· · · · Hooke s
10
68
100
0 0
0
90
80
70
........ 60
·u;
=..
c: 50
(/)
.. /
.1.1
52
40
30 o Experimental
- Stress/Strain
·- ·- Modified Glinka
20 - - Modified Neuber
10
Figure 31. Comparison of Modified Glinka and Modified Neuber Methods for
ARALL 4 Composite in Out-of-Plane Bending.
69
120
100
en 80
=:
-g
.3 60
40
20
70
-- --------~~~~~~~--~~~~~~~~~~~~~~~~~~~~~~~~~~-
0.012.-------.-------~------~------~------~------~~
1
0.0 ~-------------
0.008
... '
Sn [ksi]
78.5
0.004~------------------ 38.9
0.002
o~------~------~------~--------L-------~------~--
1 1.2 1.4 1.6 1.8 2 2.2
X/P
Figure 33. Strain Gradient Ahead of the Circular Notch for Several Nominal
Stresses for ARALL 4 Composite in Out-of-Plane Bending.
71
72
VI. CONCLUSIONS AND RECOMMENDATIONS
notch strain produced good results for the 7075-T6 aluminum in tension.
the elastic and inelastic region, a strain discontinuity was created in the
73
stress-strain curve and Glinka's and Neuber's method are based on the
notch strains was also observed for the 7075-T6 aluminum sample in tension.
better than samples in out-of-plane bending. This was true for the standard
predictions for theoretical inelastic notch strains. For the ARALL 4 composite
74
Future research in the inelastic stress distribution around notches
obtain. Precision could be improved by using larger sample widths and notch
diameters with small strain gages. However, the problems associated with
increased sample lengths and grip configuration at the ends for tension tests
requires careful planning. To record large inelastic strains, the strain gage grid
must be as large as possible to avoid internal gage failure. The trade off is that
strain gages tend to average the strain over the area covered by the grid. For a
sample and notch geometries could be modeled for different sample loadings.
75
76
LIST OF REFERENCES
4. Peterson, R. E., Stress Concentration Factors, John Wiley and Sons, 1974.
9. Molski,K. and Glinka, G., "A Method of Elastic-Plastic Stress and Strain
Calculation at a Notch Root," Material Science Engineering, Vol. 50, 1981, pp.
93-100.
77
pp. 839-854.
12. Hoffman, M. and Seeger, T., "A Generalized Method for Estimating
Multiaxial Elastic-Plastic Notch Stresses and Strains, Part 1: Theory," ASME
Journal of Engineering Materials and Technology, Vol. 107, No. 4, Oct. 1985,
pp. 250-254.
13. Glinka, G., Ott, W. and Nowack, H., "Elastoplastic Plane Stress Analysis of
Stresses and Strains at the Notch Root," ASME Journal of Engineering
Materials and Technology, Vol. 110, July 1988, pp. 195-204.
14. Sharpe, W., Yang, C., and Tregoning, R., "An Evaluation of the Neuber
and Glinka Relations for Monotomic Loading," ASME Journal of Applied
Mechanics, Vol. 59, Oct. 1992, pp. S50-S56.
15. Lee, Y., Chang, Y., and Wong, H., "A Constitutive Model for Estimating
Multiaxial Notch Strains," ASME Journal of Engineering Materials and
Technology, Vol. 117, No. 1, Jan. 1995, pp. 33-40.
16. ASTM E8-95a: "Standard Test Methods for Tension Testing of Metallic
Materials," American Society for Testing of Materials, ASTM Standards, 1995.
17. ASTM E111-82: "Standard Test Method for Young's Modulus, Tangent
Modulus, and Chord Modulus," American Society for Testing of Materials,
ASTM Standards, 1988.
18. ASTM E132-86: "Standard Test Method for Poisson's Ratio at Room
Temperature," American Society for Testing of Materials, ASTM Standards,
1992.
19. American Society for Metals, Metals Handbook, ASM International, Vol.
2, Table 108, 1985, pp. 115-117.
20. American Society for Metals, Metals Handbook, ASM International, Vol.
2, 1985, p. 116.
21. ASTM E111-82: "Standard Test Method for Young's Modulus, Tangent
Modulus, and Chord Modulus," American Society for Testing of Materials,
ASTM Standards, 1988, p. 224.
22. ASTM E132-86: "Standard Test Method for Poisson's Ratio at Room
Temperature," American Society for Testing of Materials, ASTM Standards,
1992.
78
23. ASTM E111-82: "Standard Test Method for Young's Modulus, Tangent
Modulus, and Chord Modulus," American Society for Testing of Materials,
ASTM Standards, 1988.
25. Bannantine, J., Comer, J., and Handrock J., Fundamentals of Metal Fatigue
Analysis, Prentice Hall, 1990, p. 45.
31. American Society for Metals, Metals Handbook, ASM International, Vol.
2, 1985, p. 70.
35. Peterson, R. E., Stress Concentration Factors, John Wiley and Sons, Fig.
160, 1974, p. 231.
79
80
INITIAL DISTRIBUTION LIST
No. Copies
81