You are on page 1of 25

5 ELASTO-PLASTIC MODEL

5.1 Introduction

Metals and alloys exhibit a greater variety of mechanical behavior


under different loading conditions. The knowledge of mechanical
behavior of materials plays a vital role in engineering stress analysis
and design. The elastic and plastic behavior of materials is discussed
in the subsequent paragraphs.

Plastic deformation is characterized by a permanent set. When


the load applied is beyond the elastic limit the material begins to flow
plastically. Applying suitable yield criteria one can identify the onset of
plastic deformation. The plastic deformation is essentially irreversible
and time independent. In the plastic region the material behaves in a
non-linear fashion. Elasto-plastic material behavior is characterized by
an initial elastic behavior on to which a plastic deformation is
superimposed. Elasto-Plastic material behavior is not dependent on
time. Many practical applications such as sheet metal forming, metal
cutting process, behavior of adhesive material in a bonded joint, brazed
and soldered material at the joint, all behave like elasto-plastic material.
In many practical applications, elasto-plastic material model is sufficient
to study the behavior of metal flow. In the elasto-plastic material
model, material may behave as elastic, perfectly plastic or elastic-
plastic with isotropic strain hardening and strain rate sensitivity,
depending on the work hardening parameter V. Work hardening
parameter is determined from the non-linear characteristic of the stress-
strain relationship. Mathematically • an axial stress (a) is given by
KSn, where *• is the strength coefficient and V is the strain-hardening
exponent. In this case, a and s denote the true stress and true strain.
Figure 5.1 represents the stress-strain curve of metals in simple
tension. Initially the relation between stress and strain is essentially a
straight line of slope E, which is known as young’s modulus. Point A
represents the proportional limit, at which the linear relationship
between the stress and strain ceases to hold. I.e., on further straining,
the relation between the stress and strain is no longer linear, but the
material is still elastic and upon release of the load, the specimen
reverts its original length. This property of the material is known as
elastic material behaviour. The maximum load that can be applied
without causing permanent deformation defines the elastic limit.
Usually, there is little difference between the proportionality limit and
the elastic limit.

Point B marks the end of purely elastic straining and the point of
initiation of plastic deformation. It is known as the upper yield point. At
the upper yield point, a laminar plastic zone known as Luder’s band,
inclined at approximately 45° to the tensile axis appears at a local
stress concentration. During the subsequent elongation under constant
stress, several Luder’s bands appear and gradually spread over the
entire specimen.

A sudden drop in load and then extension at approximately


constant load accompanies further straining. The lower yield stress is
defined as the load at CD divided by the original cross sectional area of
the bar. After D, the load increases with further strain. This effect of
the material being able to withstand a greater load despite the uniform
reduction in cross sectional area is known as strain hardening or work
hardening.

97
Figure 5.1 Stress-Strain curve of metals in simple tension for ductile material

/ /
/ /
/ /

Figure 5.2Stress- strain curve of metals with effect of unloading and reverse loading

98
At E, the rate of work hardening is unable to keep pace with the

rate of reduction in the cross sectional area and a maximum work

hardening occurs in the load followed by local straining leading to

fracture at F.

Beyond the yield point, the stress continually increases with

further plastic strain, while the slope of the stress-strain curve,

representing the rate of strain hardening, steadily decreases with

increasing stress. If the specimen is stressed to some point C

(See figure 5.2) in the plastic range and the load is subsequently

released, there is an elastic recovery following the path CD that is very

nearly a straight line of slope E. The permanent strain that remains on

complete unloading is equal to OE. On reapplication of the load, the

specimen deforms elastically until a new yield point F is reached. The

curve EFG may be regarded as the stress-strain curve of the metal

when pre-strained by an amount OE. The greater the degree of pre-


tht
strain, the higher the new yield point and^flatter the strain-hardening

curve. For a heavily pre-strained metal, the rate of strain hardening is

so small that the material may be regarded as approximately non­

hardening or ideally plastic.

5.2 Application Of Finite Element Method For Elasto- Plastic


Material Model

Most of the metal deformation processes are modeled with

elasto-plastic metal behaviour using finite element method. Earlier they

were analyzed using either upper bound theory or slip line field theory.
Although analytical and numerical studies of metal deformation

processes are extensive, the deformation characteristics of machining

gg
process are not fully understood. An accurate predictive model is yet to
be developed. The common plane strain model, which has been
frequently used, depends on the machining condition. The deformation
process is complex with large strains and strain rate as well as wide
range of temperatures creating special difficulties in developing reliable
process models. Many predictive models are based on idealized shear
plane and rely on experimental inputs.

However, the finite element method provides an effective means


of treating the foregoing complexities and is used in the study to
determine the distribution of strain, stress, strain rate and temperature.
Usually the predictions are compared with results obtained with
experiments. Many investigators have adopted the finite element
method to gain a better understanding of machining process.

Through FEM analysis various quantities can be numerically


calculated including spatial distribution of the strain, stress, temperature
and strain rate, each of which is difficult to measure during the process.
However, to obtain a meaningful result, which reflects the physical
mechanism of processes, it is essential to have satisfactory material
constitutive equations, which describe the deformation behaviour.
Different material constitutive equations have been used in the
literature, thus resulting in substantial variations among the simulation
result. Since none of the studies has been provided thorough
comparison of predicted quantities with experimental result, a
constitutive model has to be formed depending on the application of the
process and a study is required to link the uncertainties in obtaining the
strain, strain rate and temperature.

100
5.3 Concept Of Plasticity And Finite Element Formulation

5.3.1 The Yield Criterion

The yield criterionjDwen and Hinton, [62] determines the stress


level at which plastic deformation begins and its general form is
f(oriJ)=k(K) (5.1)

Where f is some function and k a material parameter to be determined


experimentally. The term k is a function of hardening parameter k. On
physical grounds, any yield criterion should be independent of the
orientation of the coordinate system employed and therefore it is a
function of the three stress invariants only

'2 <Jij<Jjka'ki (5.2)

Experimental observations noted by Bridgeman [62], indicate that


plastic deformation of metals is essentially independent of hydrostatic
pressure. Consequently the yield function is of the form
f(j'2,j;) = k(K) (5.3)

Where j' and J\ are the second and third invariants of the deviatoric
stresses,
l, (5.4)
=e// 2 Ojj&kk

Von-Moises suggested that yielding occurs when J2' reaches a critical


value, or

=*(*■) (5.5)
in which k is a material parameter to be determined. The second
deviatoric stress invariant, J'2 can be explicitly written as:

101
a, - a

— —1 fJCT 12 x + CT ;2 v + (J #2 z 1J4" T«2 xy + T 2 >z +. T 2 .re


. , , . h u
(5.6)

for yield criterion (Equation 5.5) is further written as

(5.7)

Where (5.8)

And a is termed the effective stress, generalized stress or equivalent


stress.

5.3.2 Work Or Strain Hardening

After initial yielding, the stress level at which further plastic


deformation occurs may be dependent on the current degree of plastic
straining. Such a phenomenon is termed as work hardening or strain
hardening. Thus the yield surface will vary at each stage of the plastic
deformation, with the subsequent yield surfaces being dependent on
the plastic strains in some way. In a perfectly plastic material the yield
stress does not depend in any way on the degree of plastification. If
the subsequent yield surfaces are a uniform expansion of the original
yield curve, without translation the strain-hardening model is said to be
isotropic. On the other hand, if the subsequent yield surfaces preserve
their shape and orientation but translate in the space as a rigid body,
kinematic hardening is said to take place. Such a hardening model
gives rise to the experimentally observed Bauchinger effect on cyclic
loading.

For some materials, notably soils, the yield surface may not strain
harden but strain soften instead, so that the yield stress level at a point

102
decreases with increasing plastic deformation. Therefore, an isotropic
model, the yield curve contracts progressively without translation.
Consequently yielding implies local failure and the yield surface
becomes a failure criterion.

The progressive development of the yield surface is defined by


relating the yield stress k to the plastic deformation by means of the
hardening parameter k. This can be done in two ways. Firstly, the
degree of work hardening is postulated as a.function of the total plastic
work (wp) only. Then

k = wp (5.9)

Where
Wp = \a,I(dsll)p (5.10)

in which (dejj)p are the plastic components of strains occurring during a


strain increment. Alternatively k is related to a measure of the total
plastic deformation termed as the effective, generalized or equivalent
plastic strain which is defined incremental as
< = Vp72){(*s),(*s),}^ (5.11)

For situations where the assumption that yielding is independent


of any hydrostatic stress is valid, (ds jj)p = 0 and hence (ds' jj)p = (de ij)p.
Consequently equation (5.11) can be written as
dsp=V(372){(rf4)f(d4)pi^ (5.12)

then the hardening parameter k is assumed to be defined as


*= (5.13)
Where ep\s the result of integration of dspover the strain path. This

behavior is termed stain hardening. Only an isotropic hardening model


is considered in this analysis.

103
Stress states for which f<x,; = k represent plastic states, while

elastic behavior is characterized by f < k. At a plastic state, fatJ= k, the

incremental change in the yield function due to an incremental stress


change is
df = -^-da„ (5.14)
d<JtJ

Then if, df< o elastic unloading occurs (elastic behavior and the
stress point returns inside the yield surface; df = o neutral loading
(plastic behavior for perfectly plastic material) and the stress point
remains on the yield surface and df > o plastic loading (plastic behavior
of a strain hardening material) and the stress point remains on the
expanding yield surface.

5.3.3 Elasto-Plastic Stress/Strain Relation

After initial yielding the material behaviour will be partly elastic


and partly plastic. During any increment of stress, the changes of strain
are assumed to be divisible into elastic and plastic components, so that
dsiJ=(dsiJ)e+(dsil)p (5.15)

Decomposing the stress terms into their deviatoric and


hydrostatic components
dv'u (l-2v)
(d£ij)e + dijdcykk (5.16)
2 fi E

Where e and v are respectively the elastic modulus and Poisson’s


ratio of the material.

In order to derive the relationship between the plastic strain


component and the stress increment a further assumption on the
material behaviour must be made. In particular it will be assumed that

104
the plastic strain increment is proportional to the stress gradient of
quantity termed as the plastic potential Q, so that

(<&,)„(5.17)
day

Where dX is a proportionality constant termed as the plastic


multiplier. Equation (5.17) is termed the flow rule, since it governs the
plastic flow after yielding. The potential Q is a function of J'2and J'3 but
it cannot be determined in its most general form. However, the relation
Q has a special significance in the theory of plasticity, since for
this case certain variational principles and uniqueness theorems can be
formulated. The identity Q is a valid one, since it has been
postulated that both are functions of j\ and J'3 and such an assumption
gives rise to an associated theory of plasticity. In this case, equation
(5.17) becomes
(dst)p=<a-%- (5.18)
day

and it is termed as the normality condition, since df/Scjg is a vector


directed normal to the yield surface at the stress point under
consideration. It is seen that the components of the plastic strain
increment are required to combine vectorial in n-dimensional space to
give a vector, which is normal to the yield surface. For the particular
case of f = j\ one have
df _ dJ'2
(5.19)
day day

then equation (5.18) becomes


{dsij)p=dXa'iJ (5.20)
which are known as the Pradnti-Reuss equations and have been
extensively employed in theoretical work. Experimental observation

105
indicates that the normality condition is an acceptable assumption for
metals. Thus on use of equations (5.15), (5.16) and (5.18) the
complete incremental relationship between stress and strain for elasto-
plastic deformation is found to be
^ + £-^) ^ (5.21)

Considering the uniaxial testing of an elasto-plastic material that


produces the stress-strain curve, the behaviour is initially characterized
by an elastic modulus e until yielding commences at the uniaxial yield
stress ay. Thereafter, the material response is elasto-plastic with the

local tangent to the curve continually varying and is termed the elasto-
plastic tangent modulus, et . The hardening law k = k(ic) could just as
easily be expressed in terms of the effective stress, a (since it is
proportional to J2) to give, forthe strain hardening hypothesis
equation (5.13)
5= = //(£„) (5.22)

^ = H’(ep) (5.23)
dsp

For uniaxial case under consideration o-, = a, a2=a3=o and thus faorm-
equation (5.8)
cr=4$l2){(j'iJa'ij}yi=<y (5.24)

If the plastic strain increment in the direction of loading is dsp, then


(dsx)p =dspand since plastic straining is assumed to be incompressible,

Poisson’s ratio is effectively 0.5 and (ds2)p=-±dep and (de3)p=-±dep.

Then from equation (5.12) the plastic strain becomes

dep = (5-25)

106
Equations (5.24) and (5.25) explain the apparent arbitrary constants
employed in the definition of a and sp, since these terms are required

to become actual stress and strain for uniaxial yielding. Using


equations (5.24) and (5.25) then equation (5.23) becomes.
da da Er
H'(sp) H (5.26)
de„ ds-d„ ds dse E.T
da da E

thus the hardening function H’ can be determined experimentally


from a simple uniaxial yield test.

5.3.4 Matrix Formulation

The theoretical expressions developed in the previous section will


now be converted to matrix form. The yield function, first defined in
equation (5.1), can be rewritten as:
§{a) = k(K) (5.27)
Where a is the stress vector and k is the hardening parameter
which governs the expansion of the yield surface. In particular, from
equations (5.9) and (5.10), die = aTdsp for the work hardening

hypothesis and from equation (5.13) dK = dse for the strain hardening
hypothesis.
Rearranging equation (5.27) we get
F(a,ie) = f(a)-k(ie) = 0 (5.28)
By differentiating equation (5.28) one have
nf? dF, dF , .
(5.29)
da die

or
aTda - AdX = 0 (5.30)
with the definitions

107
dF_ dF dF dF dF dF dF
aT (5.31)
da dax day da, dayx da,x ’ daxy
’ ’ ’ ’

and
1 dF
A=- dK (5.32)
dydrc

The vector ‘a’ is termed the flow vector. Equation (5.21) can be
immediately rewritten as
ds = [DYlda + dl^- (5.33)
da

where D is the usual matrix of elastic constants. Pre-multiplying


both sides of equation (5.33) by dTd=aTD and eliminating aTda by use
of equation (5.33), one obtainjthe plastic multiplier dX to be
dX = -—-XF—-aTdDds (5.34)
[A + a Da]

or substituting equation (5.34) into equation (5.33), one obtain the


complete elasto-plastic incremental stress-strain relation to be
da = Depds (5.35)

where D-A^L_ d D (5 36)


A + dTa ° ° V ’

It now remains to determine the explicit form of the scalar term, A.


Therefore,
dK = arded (5.37)
Equation (5.28) can be rewritten in the form
F(a,k) = f (a) - aY (tc) =0 (5.38)
Since the uniaxial yield stress, aT=J(3)k. Thus from the equation
(5.32)
WF 1 dar
dK = dK (5.39)
dXdic dX dK

108
Note that the full differential may be employed in the last term
since ay is a function of k only. Employing the normality condition in

equation (5.37) to express dsp as

d.K - crTdep - a7 dXa = dAa Ta (5.40)

or,1 for the uniaxial case o- = a = o-vand


y
ds=dsB
p p
where a and sBp

are respectively the effective stress and strain. Thus equation (5.40)
becomes
dK = cry dsp = dAa7a (5.41)

Also, from the equation (5.23) we have


do' ^ do> ^
(5.42)
dsp dsp

Using Euler’s theorem applicable to all homogeneous functions of order


one, we can write from equation (5.38)

= (5.43)

or from equation (5.31)


aTar = <7y (5.44)

Substituting equation (5.42) and equation (5.44) into equation (5.41)


and equation (5.49) we obtain
dZ = dep .

A = H' (5.45)
Thus A is obtained to be the local slope of the uniaxial stress/plastic
strain curve and can be determined experimentally from equation (5.26)
Numerical computation of the flow vector ‘a’

aT = J JL JUL (5.46)

\dax do)y drxy daz ’ ’
a = tfjCj + a-fCj + a3 C3 (5.47)

109
Where a! = {1,1, 0,1}

al ——— —a'a’ 2 a',}


2(./'2) ; A 9 ‘
r, , j 2i F
’ . , T'2l f2
a. L
O’ O’, +----
3 J L <7 0. +—
3 J ■ 2(j' (t' - vxy \ o'o' - r%, + (5.48)

Constants defining the yield surface criteria are;


'J3sin9
Tresca C,=0 C2 = 2 cos <9(1 +tan 0 tan 3$) C,
J[ cos3$

Von Mises c, = o, C, =V3 C3 =0,

5.4 Program Development And Implementation Details

This section describes the development of the program along with


the necessary details for the complete understanding of the elasto-
plastic program.

The typical finite element program is developed in C++ and it is


composed of a series of common modules with elastic model, which
may have different uses in different context. Typical modules are data
input, element stiffness, and equation solving procedures, stress effect
calculation and output display.
The program developed has composed of mainly the following
parts.
Master Controlling Program
1. Function init
2. Function input
3. Function loadps
4. Function band
5. Function zreo
6. Function stiffp

110
7. Function ioadcon
8. Function solve
9. Function bmatps
10. Function modps
11. Function dbe
12. Function linear
13. Function output
14. Function increm
15. Function algor
16. Function residu
17. Function conver
18. Function invar
19. Function yieldf
20. Function flowpl

5.4.1 Master Controlling Program


The master controlling program controls the calling of the other
subroutines in order. The program segment also controls the iterative
process and also the incrementing of the applied load in appropriate
way. A flow chart of the program is given in the Appendix A.4
This program reads the input data, loads, and initializes various
arrays to zero. The main loop is for the load increment and inside the
main loop there is one more loop for the maximum permissible number
of iterations. Within the iteration loop it calls for the formulation of the
element stiffness solves for the unknown displacement and reactions.
It checks for the convergence if the solution has converged then
outputs the solution. If the solution is not converging it stops the
program. Depending on number of load increments the outer loop is
repeated many number of times.

m
5.4.2 Function Input
The role of this subroutine is to accept most of the input data
required for analysis of elasto-plastic problem. This subroutine follows
UntK
closely^ that of the function of input in the elastic problem. Some
changes are as below.
‘nalgo’ is the parameter controlling the nonlinear solution
algorithm
nalgo = 1 Initial stiffness method. The element stiffness are
computed at the beginning of the analysis and remains unchanged
thereafter
nalgo = 2 Tangential stiffness method. The element stiffness are
recomputed during each iteration of each load increment
nalgo = 3 Combined algorithm. The element stiffness are recomputed
for the first iteration of each load increment only.
Nalgo = 4 Combine algorithm. The element stiffness are
recomputed for the second iteration of each load increment
The yield criterion to be employed ‘ncrit’
Ncrit = 1 Tresca, ncrit = 2 Von Mise’s
‘nstre’ - total number of increments in which the final loading is to be
applied ‘nstre’ number of independent stress components for the
application
3 - Plane stress/strain
4 - Axial symmetry
The list of material properties for the elasto-plastic applications is
elastic modulus E, poisson’s ration, material thickness, strain hardening
component, yield stress value of the material,

112
5.4.3 Function Increment
The role of the subroutine is to increment the applied loading or
any proscribed displacements according to the load factors specified as
input. This subroutine is accessed on the first iteration of each load
increment. The total load can be applied in various load increment
factors. This subroutine reads the factor for the current load increment
and adds this to the variable ‘tfact’ which gives the total load that is
being applied on the structure. Depending on the variable ‘facto’ the
load factor for current increment, the rload array is multiplied by this
factor and is stored in the array ‘eload’, the load for current increment.
It inserts appropriate values in the fixity array to control any prescribed
displacements.

5.4.4 Function Algor


The role of this subroutine is to control the solution process
according to the value of. > — j . \ the parameter ‘nalgo’ in the input
file. This sets the value of the variable ‘krsel’, depending on which we
go far either the initial stiffness method or the tangential stiffness
method. If ‘kresel’ = 1 indicates reformation of the element stiffness
accompanied by a full equation solution. If “kresel’ = 2 indicates
element stiffness l;'is not to be modified and consequently only
equation resolution takes place.

5.4.5 Function Invar


This computes the stress invariants from the total elemental
stress it calculates first, second and third invariant of the stress.
Depending on the yield criteria being employed it calculates the
effective stress to predict whether the element has yielded and crossed
the elastic limit.

113
5.4.6 Function Yieldf
The function of the subroutine is to calculate the flow vector ‘a’ as
per equation (5.46), (5.47) and (5.48)

5.4.7 Function Flowpi


This subroutine calculates the vector dD, which is required in the
elasto-plastic matrix Dep as per the equation (3.36).

5.4.8 Function Linear


This subroutine is merely to determine the stresses from the
given displacements assuing the linear elastic material behaviour.

5.4.9 Function Loadcon


This initializes ‘asids’ array by ‘eload’ array prior to solving the
stiffness matrix for any load increment or any iteration.

5.4.10 Function Solve


This solves the resulting simultaneous equations by using Gauss
elimination approach. After solving the equation the displacements are
stored in the array ‘asdis’. This displacement is the incremental
displacement and it is added to the ‘tdisp’ array to get the total
displacement after the current increment. Also the reactions are
calculated and stored in the array ‘fixed’. This is the incremental
reaction and the total reaction is obtained by adding this to the ‘treac’
array, which gives the total reaction after a specified number of load
increments.

5.4.11 Function Residu


This subroutine calculates the residual forces on each nodal point
after a particular load increment is applied. For any load increment, it is

114
necessary to determine what proportion is elastic and which part
produces plastic deformation and then adjust the stress and strain
terms until the yield criterion and the constitutive laws are satisfied.
This subroutine in turn calls ‘bmats’, ‘modps’, ‘linear’, ‘invar’, yieldf’,
and ‘flowpl’ subroutines. The residual forces obtained after a particular
iteration is applied as the nodal forces for the next iteration of the
particular load increment. The airm is to make the residual forces on
each node to zero by means of which convergence can be achieved for
the nonlinear numerical solution.

5.4.12 Function Conver


This subroutine monitors convergence of the nonlinear solution
iteration process. It calculates the norm of the total applied load and
the norm of the residual force. Depending on the ratior of the norm of
the residual forces to the norm of the applied load it gives information
regarding the state of convergence. In this the variable ‘nchek’ = 0, if
the solution has converged and ‘nchek’ = 1 the solution is yet to
converge and if ‘nchek’ = 999 then the solution is diverging and the
program will terminate.
The remaining functions in this program do the same operation as
in the elastic program.

5.5 Spring Back Effect In Sheet Metal Forming

In plasticity and in creep problem it is well known hypothesis that


the partition of deformation is given by
e=ee+em (5.49)
where e, is the total deformation

115
eg is the elastic reversible deformation and is the inelastic
irreversible permanent deformation. In case of elasto-plastic modeling
em=ep i where epis the plastic deformation. In case of creep, the
externally imposed variable may be temperature, variation, humidity
etc, ejn=ecwhere ecis the creep deformation. The difference between
the total deformation and plastic deformation is the main cause for
spring back effect, also it may not be predominant as the variation of
temperature, humidity etc.

The developed elasto-plastic model is used to determine the


elastic deformation as well as plastic deformation for different loading
conditions.
Figure 5.3 illustrates in X-direction and Y-direction deformation for
the axisymmetric loading without friction.

The loads applied are 4000 Newton, 6200 Newton and 6500
Newton respectively. These plots are obtained for elasto-plastic
material model for the comparison. Figure 5.4 illustrates X and Y
deformation considering the reversible elastic strain.

For elasto-plastic modeling maximum displacement in X-direction


is 0.15mm for the load 6000 Newton and it is increased to 0.175mm
when load is increased to 6500 Newton. In reversible elastic strain, it is
observed 0.14mm and 0.125mm for the load 6000 and 6500 Newton
respectively. The displacement obtained considering elasto-plastic
material is slightly higher when contain^elastic stress and elastic strain.

Similarly the stress in X and Y direction and shear stress obtained by


elastoplastic modeling are more compared to elastic-material

116
behaviour. After the load is released, the residual stress due to elastic
strain also relived producing the spring back effect. The difference
between total deformation and plastic strain gives the spring back
effect. This is illustrated in the figure 5.5. In this case the value of the
spring back varies from 0.025mm to 0.035mm. It is also observed the
spring back effect is not appreciably in X direction deformation compar-
to the spring back effect in Y direction. This clearly indicates that the
to
spring back effect is not only affect deformation in Y-direction but also
/\

there is a possibility of element buckling due to the deformation change


in Y-direction.

117
Displacement in X-Direction
0 9 9 9 9 9 9 0 9
i
------- ------- ------- ------- ------- ------- ------- 1-
1
(iutu) juaiusoeiedsia-X

1
11
1
1

1.00 4.00 6.00 8.00 9.00


8

Nodal Point

(a)

Figure 5.3 x-direction and y-direction deformation

118
Displacement in X-Direction
X-Displacement (mm)

Nodal Point

(a)

Displacement in Y-Direction
Y-Displacement (mm)

0.00 1.00 2.00 4.00 6.00 8.00 9.00


Noadal Point

(b)

Figure 5.4 Deformation with reversible elastic strain

119
Figure 5.5 Elastic and plastic strain to show the spring back

120

You might also like