Validation of Stress Concentration Factor and

Estimation of Local Stresses Beyond Yield

Point of the Material for Notched Specimens

Vinayak H. Khatawate1, M A. Dharap2, R I K. Moorthy3

1,3
Dept. of Mechanical Engineering, PCE NewPanvel, Navimumbai-410206, India

2
Dept. of Mechanical Engineering, VJTI, Matunga,Mumbai-400019, India


Abstract— In this paper, estimation of stresses at the notch fields in the vicinity of stress concentrations on the basis of
root is reported for the three notched specimens using finite local effects, net section collapse and the reference stress. The
element analysis (FEA). Linear FEA is performed on the notches NSSC rules were used to develop the method. A notch root
to evaluate stress concentration factor (Kt) and validated using
total strains (for monotonic loading) and total strain ranges
Roark’s formulae. For the three specimens, nonlinear finite
element analysis (NLFEA) is performed to evaluate local plastic (for cyclic loading) [13] are investigated using the
stresses at the notch root and compared using notch stress strain intermediate NSSC rules and the results were compared with
conversion (NSSC) rules viz., Neuber’s rule and Equivalent FEA. The hollow tube was used for the analysis. The notch-
strain energy density (ESED) approach. It is found that Neuber’s root stress-strain behavior was [14] measured by means of the
rule predicts better results of plastic stresses for flat specimens, companion specimen method and was compared using both a
whereas ESED predicts better results for cylindrical specimen.
constant (long life) fatigue-notch factor and a variable fatigue-
Index Terms— Neuber’s rule; Equivalent strain energy density notch factor found from fatigue life data. A companion
(ESED); Finite element analysis (FEA); Stress concentration specimen method was used to verify a Neuber based notch
factor; Notch; analysis.
Among these, most popular and commonly used methods
are Neuber’s rule and ESED method proposed by Molski and
I. INTRODUCTION Glinka.
In this paper evaluation of theoretical stress concentration
F ATIGUE life prediction methods in use today are based on
the local stress and strain. Therefore, the accurate
evaluation of inelastic stresses and strains induced at the
factor ( ) at notch root for three different notched geometries
using linear FEA (LFEA) is presented. The results obtained
geometric irregularities is significant from the life prediction using linear FEA is validated using Roark’s formulae. The
point of view. Early researches focused primarily on validated model is used to perform NLFEA. The paper also
determining theoretical stress concentration factor ( ) using focuses on prediction of local stresses on three notched
either elasticity theory or photoelastic analysis [1]. So far, specimens by NSSC (Notch stress-strain conversion) rules viz,
approximate methods [2-8] for assessing local stresses and Neuber’s rule and ESED method, and their predictions are
strains have been developed. Some researchers have used assessed by comparing results with NLFEA.
linear FEA to assess local stress and strain by modifying
material properties such as elastic modulus and Poisson’s ratio II. NSSC RULES
[9-11]. A. Neuber’s Rule
A method is proposed [12] for estimating stress-strain
Theoretical or Linear elastic stress concentration factor ( )
[1] is defined as the ratio of maximum stress at notch root
Vinayak H. Khatawate is with the Mechanical Engineering Department, ( ) to nominal stress ( )
Pillai College of Engineering, Newpanvel, Navimumbai-410206 India (e-
mail: vinayakhk@ gmail.com; [email protected]).
M. A. Dharap, is Professor (Retd.) from Veermata Jijabai Technological
Institute (VJTI), Department of Mechanical Engineering, Matunga, Mumbai-
400019 India (e-mail: [email protected]). Beyond yield point, becomes invalid and two more
R.I.K. Moorthy is with the Mechanical Engineering Department, Pillai concentration factors namely stress concentration factor (
College of Engineering, Newpanvel, Navimumbai-410206 India (e-mail:
[email protected]).
and strain concentration factor ( ), are needed.
International conference on Advances in Thermal Systems, Materials and Design Engineering (ATSMDE2017) 1

Electronic copy available at: https://ssrn.com/abstract=3101362

 Neuber’s rule [2] (Neuber,1961) correlates all the three The relation between theoretical stress concentration factor
concentration factors by equation, and Strain energies ( is given by

(2) (4)
In terms of maximum stresses and strains equation (2)
becomes
Graphical interpretation of and is shown in Figure 2.
(3)
The above relation is used to find notch stress and strain in
plastic region.
Equation (3) when solved with stress strain curve equation
gives the actual local stress-strain state of the notch as shown III. MATERIAL PROPERTY
in Figure 1.
Johnson Cook elastic plastic material model [15] is used for
nonlinear analysis. For laboratory experiment at room
temperature with normal strain rate, the model is given by
equation,

( ) (5)

Johnson Cook material model parameters for Aluminum

6063T7
a= 90.26 MPa, b=223.13 MPa and n = 0.374
Failure Stress for the material, m = 175 MPa.

IV. PROBLEM DEFINITION

TABLE I
GEOMETRIES SELECTED FOR THE ANALYSIS AND STRESS CONCENTRATION
FACTOR BY ROARK’S FORMULAE [16].

Fig. 1. Local stress and strain using Neuber’s rule

Specimen configuration Formuulae

Specimen-1:Flat plate with centre ( )

circular Hole
(D= 40mm, r =5mm,t =5mm) ( )
( )
=2.42

Specimen-2:Flat plate with Semicircular ( )

edge notches.
(D=40mm, r=h=5mm, t=5mm) ( )

( )
= 2.27

Fig. 2. Strain energy density due to and [3].

B. Equivalent Strain Energy Density (ESED) Method Specimen-3:Semicircular groove in ( )

cylindrical bar.
One more frequently used method to find notch stress and (D=40mm, r=h=5mm) ( )
strain is energy density approach given by Molski and Glinka
[3]. Strain energy density is strain energy per unit volume and ( )
is given as W= ∫ and graphically it is area under =2.03
stress-strain curve..
International conference on Advances in Thermal Systems, Materials and Design Engineering (ATSMDE2017) 2

Electronic copy available at: https://ssrn.com/abstract=3101362

 Three specimen geometries are considered, viz., flat plate
with centre circular hole, flat plate with semicircular edge
notches and cylindrical specimen with circumferential groove.
Empirical formulae of for the geometries selected are
given by Roark [16] and are presented in Table I. All the three
specimens are subjected to uniaxial tension. The material is
homogeneous and isotropic. For this analysis the Cartesian
coordinates (x,y,z) are used.

V. FINITE ELEMENT MODEL

For the specimen-1 and 2, the thickness is smaller in
comparison with length and width therefore plane stress
condition is assumed and hence 2D elements (4 node Quad Fig. 4. Meshing of Specimen-1 with 2D elements.
elements) are used, whereas for the specimen-3, 8 node brick
TABLE II
elements are preferred. Case study for specimen-1 (Figure 3) RESULTS FOR LINEAR ANALYSIS.
is presented for convergence of results. The FEA is carried
out using HYPERWORKS finite element program [17]. Specimen % error
(LFEA) (Roark (LFEA)
MPa [16])
Initially both auto mesh and biased mesh option is used to
carry out the convergence of results. In auto mesh option, with
1 63.90 2.42 2.39 1.23
0.2 mm mesh length, the number of elements generated are
2 57.85 2.27 2.16 4.84
24412. In bias option,1 mm mesh size is selected and biasing
is done in critical region. Biasing is one of the useful feature in 3 54.03 2.03 1.91 5.91
commercial FEA software. With this feature even without
increasing the number of elements one could achieve better
results by appropriate arrangments of nodes and elements. VI. RESULTS AND DISCUSSION
Bias intensity in the steps of 0,5,10 and 20 is selected and the
results for specimen-1 as a case study is shown in Figure 3. A. Linear analysis
For specimen-1, with mesh size of 1 mm for all the 4 bias A linear FEA (LFEA) of three specimens with varying
intensities, the number of elements generated are 1100. With notch geometry is performed. For linear analysis the value of
bias intensity of 20, it is found that, the solution approaches will remain same irrespective of magnitude of axial load.
the exact result. For specimen-1 and 2, an axial load of 2000N is applied and a
load of 20000N is applied to specimen-3. The convergence of
value is studied with respect to theoretical Roark’s
3.0
formulae [16] and LFEA. Figure 3 show variation of with
number of elements. As number of elements increased, the
2.5
value of converged to approximately exact value. From the
Figure 3, it is also clear that better convergence is obtained
2.0
with less number of elements when biased meshing is used.
Kt by Roark
Therefore for the analysis of all the specimens, biased
Kt

1.5
Kt by Biased Mesh meshing is used.
Kt by Auto Mesh Figure 4, show meshing of specimen-1 using biased mesh.
1.0
Owing to symmetry only one quarter of all the three models
are meshed. Appropriate boundary condition constraints are
0.5
placed at all planes of symmetry. Figure 5,6 and 7 shows
maximum Von-Mises stresses of 63.90, 57.85 and 54.03 MPa
0.0
0 5000 10000 15000 20000 25000 30000 in specimens-1, 2 and 3 respectively. The corresponding
Number of elements LFEA values of for all the specimens and percentage error
between Roark and LFEA are tabulated in Table II. The
nominal stress ( ) is obtained by ratio of applied load and net
Fig. 3. Convergence of stress concentration factor. cross sectional area.

International conference on Advances in Thermal Systems, Materials and Design Engineering (ATSMDE2017) 3

Electronic copy available at: https://ssrn.com/abstract=3101362

B. Nonlinear analysis
The validated finite element models in the linear analysis is
used to find local stresses using NLFEA. Nonlinear analysis is
carried out by using three methods viz., Neuber’s rule, ESED
method and elastic-plastic FEA (NLFEA). Local stresses for
all the three specimens are calculated for different values of
nominal stress ( ). Since Neuber’s rule is valid when
nominal stress is within and maximum up to elastic limit,
hence is taken up to yield strength of the material for all the
three specimens.

Fig. 7. Maximum stress in Specimen-3

Fig. 5.Maximum stress in specimen-1

Fig. 8. Distribution of local stresses for specimen-1.

Fig. 6. Maximum stress in specimen-2.

Fig. 9. Distribution of local stresses for specimen-2.

International conference on Advances in Thermal Systems, Materials and Design Engineering (ATSMDE2017) 4
 three specimens the results of local stresses by NLFEA lie
between results of Neuber’s and ESED method. Neuber’s rule
give upper bound values i.e., the results are conservative,
whereas ESED method give lower bound values. It is shown
that Neuber’s rule gives better prediction for flat specimens-1
and 2, whereas ESED method gives better prediction for
cylindrical specimen-3, when compared to NLFEA.

APPENDIX
Theoretical stress concentration factor (Peterson’s
stress concentration factor
notch root stress
notch root strain
Nominal stress at net section
Stress concentration factor
Fig. 10. Distribution of local stresses for specimen-3. Strain concentration factor
E Modulus of elasticity
Strain Energy density due to nominal stress
The distribution of maximum stresses by Neuber, ESED Strain Energy density due to notch stress
and NLFEA at the notch root with nominal stress for the three True stress
specimens is shown in Figures 8, 9 and 10. From the figures a Yield stress
(8, 9 and 10), it is clear that for all the three specimens b Hardening parameter
NLFEA results beyond yield point lie between results of p Plastic Strain
Neuber’s and ESED method. Neuber’s rule gives upper bound n Hardening exponent
values i.e., the results are conservative, whereas ESED method  True strain
gives lower bound values. Figure 8 shows variation of e Elastic strain
maximum stress values versus different for specimen-1, in Failure stress
which variation of maximum stress with is approximately
same for Neuber’s, ESED and NLFEA up to yield point. ACKNOWLEDGMENT
Beyond yield point, the results of NLFEA are more closer to
Neuber’s rule compared to ESED method. Similar results can Authors are grateful to Pillai college of engineering,
be seen for specimen-2 as shown in Figure 9. However for NewPanvel for supporting this work. Authors would like to
specimen-3, the results of NLFEA are more close to ESED thank Dr.K.M.Prabhakaran of BARC Mumbai for helpful
method than Neuber’s rule (Figure 10). It is found that, for discussions.
specimen-1, the percentage variation of Neuber and ESED
from NLFEA at = 50, 70 and 90 MPa is -0.686%, -0.9%, - REFERENCES
0.209% and 1.692%, 2.887%, 4.380%. For specimen-2, it is - [1] Pilkey D Walter. Peterson’s stress concentration factors, 3rd ed. John,
Hoboken, New Jersey: Wiley & Sons, Inc, 2008.
0.551%, -1.101%, -0.958% and 1.397%, 2.45%, 3.483%, [2] Neuber H (1961). Theory of Stress Concentration for Shear-Strained
where as for specimen-3, the percentage variations are -3.65%, Prismatic Bodies with Nonlinear Stress-Strain Law. ASME Journal of
-2.04%, -2.265% and -2.274%, 1.154%, 1.830%. From the Applied Mechanics; 28(4): pp.544-550.
[3] Molski K and Glinka G (1981). A Method of Elastic-Plastic Stress
above analysis, it is clear that, Neuber’s rule predicts better and Strain Calculation at a Notch Root. Materials Science and
results of local stresses for flat specimens-1 and 2, whereas Engineering; 50 (1): pp.93–100.
ESED method give better prediction for cylindrical specimen- [4] Glinka G (1985). Calculation of Inelastic Notch-tip Strain-Stress
Histories under Cyclic Loading. Engineering Fracture
3 when compared to NLFEA. Mechanics;22(5):pp.839-854.
[5] Glinka G (1985). Energy density approach to calculation of Inelastic
strain-stress near notches and cracks. Engineering Fracture Mechanics;
VII. CONCLUSIONS 22(3): pp.485-508.
[6] Seeger Timm and Heuler Paul (1980). Generalised Application of
In this paper most commonly used criteria to compute the Neuber’s Rule, Journal of Testing and Evaluation;8(4):pp.199-204.
elastic-plastic stresses for the three notched geometries is [7] Hitham M.Tlilan, Sugawara Yosuke and Majima Tamotsu (2005). Effect
systematically studied. All the three specimens are meshed of notch depth on strain-concentration factor of notched cylindrical bars
under static tension. European Journal of Mechanics A/solids;24(3):
with Auto and Biased meshing. It is found that by using biased pp.406–416.
meshing highest degree of accuracy can be obtained with less [8] Sharpe W., Yang C., and Tregoning R (1992). An Evaluation of the
number of elements compared to Auto meshing. Hence all the Neuber and Glinka Relations for Monotonic Loading. ASME Journal of
three specimens are meshed with Biased meshing and Applied Mechanics; 59(2S):pp.S50-S56.
validated using Roark’s formulae. Non-linear finite element
analysis is carried out on these validated models. For all the
International conference on Advances in Thermal Systems, Materials and Design Engineering (ATSMDE2017) 5
[9] Seshadri R (1991). The Generalized Local Stress Strain GLOSS
Analysis-Theory and Applications. ASME Journal of Pressure Vessel
Technology;113(2):pp.219– 227.
[10] Babu, S., and Iyer,P.K (1998). Inelastic Analysis of Components Using
a Modulus Adjustment Scheme. Journal Pressure Vessel
Technology;pp.120(1):1-5.
[11] Hyde, T. H., Sabesan, R., and Leen, S. B, J (2004). Approximate
Prediction Methods for Notch Stresses and Strains under Elastic-Plastic
and Creep. J.Strain Analysis;39(5): pp.515–527.
[12] Adibi-Asl, and Seshadri, R (2010). Improved Prediction Method for
Estimating Notch Elastic-Plastic Strains. J. Pressure Vessels
Technology;pp.132(1)/011401.
[13] Gowhari-Anaraki, A. R., and Hardy, S. J (1991). Low Cycle Fatigue
Life Predictions for Hollow tubes with Axially Loaded Axisymmetric
Internal Projections. J. Strain Analysis; 26(2): pp.133–146.
[14] Conle, A., and Nowwack, H (1977). Verification of a Neuber based
Notch Analysis by the Companion-Specimen Method. Experimental
Mechanics;17(2):pp.57-63.
[15] Johnson Gordon R, and Cook William H (1983). A constitutive mode
and data for metals subjected to Large strains, high strain rates and high
temperatures. Seventh international symposium on Ballistics, The
Hague, Netherlands April 1983.
[16] Young W C & R G Budynas. Roark's Formulas for Stress and Strain, 7th
ed. NewYork: Mc.Graw Hill,2002.
[17] HyperWorks, Version-11, HyperWorks tutorials, Altair Engineering,
Inc, 2011.

International conference on Advances in Thermal Systems, Materials and Design Engineering (ATSMDE2017) 6