CEN-525

FEM IN GEOTECHNICAL ENGINEERING

INTRODUCTION TO PLASTICITY

Instructor- Dr. AKANKSHA TYAGI

[email protected]
Introduction to Elastic-Plastic Behaviour
• Soil is highly non linear
• Concrete and steel behaviour is also non linear (strictly
speaking)
• Although in many cases, it is possible to assume linear
elastic behaviour of concrete and steel for typical load levels
under working conditions

2
Introduction to Elastic-Plastic Behaviour

• Elasticity - Strains and deformations are completely

recoverable upon unloading recoverable upon unloading

• Plasticity - Part of the strains and deformations are

irrecoverable upon unloading. Goes beyond nonlinear
elasticity

3
• New “rules” are needed to describe
behaviour the material in the plastic regime.

• Existing “rules” governing elastic behaviour

(such as the generalised Hooke’s Law) fail to
give an adequate description of plastic
behaviour.
Any similarities with e-log 𝝈’ ?
4
Basic Ingredients of a Plasticity Model

• Yield Criterion

• Flow Rule

• Hardening Rule

5
Yield Criterion
• The criterion which defines the conditions under plastic behaviour will
be initiated is known as a yield criterion.
• In a 1- dimensional tensile test on a steel specimen, this criterion is
often simply defined as the stress above which yielding occurs, also
known as the yield point.

Can be a single stress component or parameter 𝝈 > 𝝈y or yield function

F= 𝝈 - 𝝈y
6
Yield Criterion

• In more complex situations : 2-D or 3-D, single stress component or

parameter is often insufficient to characterize the conditions governing
the onset of yielding
• Yield criterion is often defined as a combination of stresses or stress
parameters
• Yield criterion can be defined in terms of the 3-D stress components
such as 𝝈x , 𝝈y , 𝝈z etc
• Quite common to define yield criterion in terms of stress invariants p’
and q
𝝈1 +𝝈2 +𝝈3 𝝈′1 +𝝈′2 +𝝈′3
Cambridge Stress path parameters, 𝑝 = , 𝑝′ = ,
3 3

𝑞 = 𝑞 ′ = 𝝈1 − 𝝈3 = 𝝈′1 − 𝝈′3

7
Yield Criterion

• In more complex situations : 2-D or 3-D

• a 2-parameter yield locus, (e.g. in terms of p’ and q) can be plotted as a
curve or yield locus in stress space

8
Yield Criterion

• If yield criterion is a function of 3 or more parameters, then it can be

represented by a yield surface in 3 or n dimensional stress space.

• Inside the yield surface, the stress state of the material point is elastic.

• When the stress state of the material point moves until it reaches the
yield surface, then the point starts to yield.

• When the point continues to yield, it will move along the yield surface.

• The stress state of the point cannot move outside the yield surface.

• Stress states outside the yield surface are impossible to attain!

9
Yield Criterion

Mathematically,

• general form of yield function: F(p’,q’,….)=0

• if F(p’, q ) = 0, then the stress state of the F(p’, q, …..) = 0

• if F(p , q, …) = 0, then the stress state of the soil is such that the material
is yielding.

• if F(p , q, ) , ’, q, …) < 0, then the stress state of the soil is such that the
material is elastic (inside the yield surface).

• F(p’, q, …) > 0 is not defined, as it implies that the stress state of the soil
is outside the yield surface.

10
Flow Rule

• The word ‘flow’ implies motion similar to that of a fluid.

• When a material yields or turns plastic When a material yields or turns
plastic, it tends to ‘flow’ more like a fluid. It no longer follows the
generalized Hooke’s Law like an elastic material.
• A ‘flow rule’ is used to describe how a material deforms when it is
yielding.
∆𝜀𝑣 ∆𝑝′
• For elastic behaviour, ∆𝜀𝑣 ∝ ∆𝑝′ ∝
∆𝜀𝑠 ∆𝑞
∆𝜀𝑠 ∝ ∆𝑞 ′
• Above behavior (in terms of incremental strains Δ𝜀 and incremental
stresses Δ𝝈) is not necessarily true once the material yields and
becomes plastic

11
Flow Rule

• For many materials, e.g. metals, the ratio of strain increments in the
plastic regime is not dependent on the stress increments.
• For plastic behaviour,
∆𝜀𝑣 𝑝′
∝
∆𝜀𝑠 𝑞

• The ratio of strain increments in the plastic regime tends to depend on

the stresses (or stress states) of the material (during plastic straining)
• A convenient way to describe the plastic strain increments is by means
of a potential function

12
 Flow Rule
• In mathematics, a potential is a scalar function that may be defined in
terms of several variables
• Recall from fluid mechanics the idea of several variables a potential
function (x,y) to describe fluid flow.
• The velocity field v(x,y) is obtained as the gradient of the scalar
potential function φ.

Orthogonality of stream lines and equi-potential lines

13
http://www-mdp.eng.cam.ac.uk/
Flow Rule
• In the same way, once we have defined the flow potential function G for
the plastic behavior as a function of say, the stresses p’ and q,
• we can obtain the ratio of the plastic flow increments as the ratio of the
partial derivatives of G with respect to p’ and q.

𝑑𝜀𝑠 = 𝜆 𝛛𝐺ൗ𝛛𝑞 𝑑𝜀𝑣 = 𝜆 𝛛𝐺ൗ

𝛛𝑝′

14
Flow Rule

• For many materials, the yield function F and the plastic potential G
appear to be coincident or completely associated with each other.
• In this case, the plastic strain increment vector is normal to the yield
locus F and the condition of “normality” is said to hold.
• Alternatively, this situation is sometimes termed as one of “associated”
flow.

15
Flow Rule

• When
• G ≠ F, we get the ‘non-associated’ flow.

16
Flow Rule
Flow rule relates only to plastic component of strains

Total strain increment

The elastic strain increment is computed in the same way as that in the
elastic regime, using the generalized Hooke’s Law. 17
Hardening Rule
• Soils are capable to resisting increasing stresses even after yielding
(other examples are steel and copper)
• Even on the virgin compression line, effective overburden stress
increases as the void ratio is decreased, as the soil is compressed. This
phenomenon is known as strain-hardening.
• In plasticity theory, hardening is represented by allowing the yield
surface to expand (i.e. the yield function to change) after yielding. This
is normally implemented by changing the values of parameters which
control the size of the yield surface, or hardening parameters

18
The Mohr-Coulomb Yield Criterion
• Probably the most widely used yield criterion
Mohr circle in σ’- τ space

In terms of effective stress

τ = 𝑐′ + 𝝈′ 𝑡𝑎𝑛′ Eq 1

• However, as a yield criterion, above equation is difficult to utilize since

𝝈’ and τ depend upon the direction. It is therefore much better to
express the Mohr-Coulomb yield criterion in terms of stress invariants.
𝝈′1 − 𝝈′3 𝝈′1 + 𝝈′3
= 𝑠𝑖𝑛′ + 𝑐 ′𝑐𝑜𝑡′ Eq 2
2 2
19
 The Mohr-Coulomb Yield Criterion
𝝈′1 − 𝝈′3 ′ 𝝈′1 + 𝝈′3
= 𝑠𝑖𝑛 + 𝑐 ′𝑐𝑜𝑡′ Eq 2
2 2
Substituting Eq 3 and 4 in Eq 2

In triaxial stress state, 6𝑠𝑖𝑛′ 6𝑐 ′ 𝑐𝑜𝑠′

𝑞′ = 𝑝′ 3−𝑠𝑖𝑛′
+ 3−𝑠𝑖𝑛′
𝝈′1 +𝝈′2 +𝝈′3 Eq 3
𝑝′ =
3

′
𝑞′ = 𝑀𝑝′+𝐶
𝑞 = 𝝈′1 − 𝝈′3 Eq 4

Under triaxial conditions, the Mohr-Coulomb criterion can also be

represented by a straight line in p’-q stress space.

6𝑠𝑖𝑛′ 6𝑐 ′ 𝑐𝑜𝑠′
𝑀= 𝐶=
3 − 𝑠𝑖𝑛′ 3 − 𝑠𝑖𝑛′

20
The Mohr-Coulomb Yield Criterion : 3-D Representation

• Under general stress conditions, the Mohr-Coulomb is more

complicated because there are six possible permutations of the
magnitudes of 𝝈1’, 𝝈 2’ and 𝝈 3’.

This requires six planes to be

generated in 3- dimensional principal
stress space (𝝈1’ - 𝝈 2’ - 𝝈 3’).

Irregular hexagonal pyramid in

principal effective stress space

21
The Mohr-Coulomb Yield Criterion : 3-D Representation

• Under general stress conditions, the Mohr-Coulomb is more complicated because

there are six possible permutations of the magnitudes of 𝝈1’, 𝝈 2’ and 𝝈 3’.

Note that, along the space diagonal of this

figure, the three principal stresses 𝝈a’= 𝝈b’ =
𝝈c’ are all equal and therefore q = 0.
Thus, the space diagonal is co-aligned with
the p’ axis

• A plane can be defined at right angle to the p’-axis, this

plane is termed as π-plane.
• Since the q-axis is normal to the p’-axis, it is also normal
to the space diagonal, and therefore lies on π-plane
• There are an infinite possible number of q axes
depending upon the loading.

22
The Mohr-Coulomb Yield Criterion : 3-D Representation

Extension

Compression

• Triaxial compression loading is represented in above Figure by the three

outermost apexes of three conical cross-section

• The other three apexes represent triaxial extension type of loading.

23
Mohr Coulomb Yield Criterion- Major Drawbacks

1. Yielding can only occur if the deviator stress exceeds a

certain value in relation to the normal stress
• Under a condition of isotropic compression, q = 0 while p’ increases and
the Mohr-Coulomb yield criterion would predict no yielding at all.
• This is incorrect since, as yielding can also occur under isotropic
compression as the p’ increases
• The volumetric yielding behaviour of soils cannot be modelled by the
Mohr-Coulomb criterion.
• In situations where the soil is normally or lightly over-consolidated (e.g.
soft marine clays), the application of loads may lead to volumetric
yielding and the use of the Mohr-Coulomb criterion can result in quite
significant errors.

24
 Mohr Coulomb Yield Criterion- Major Drawbacks

Suppose K0 = 0.6 and ɸ’ = 24°, then cu ~ 0.35σv’. If K0 = 1.0, then cu ~ 0.47σv’, which
is also far higher than the experimental range of cu ~ 0.2σv’ – 0.25σv’.

In other words, the use of the Mohr-Coulomb model with effective strength
parameters in undrained analyses over-estimated the implied undrained shear
strength of the soil and therefore errors on the unconservative side.
For this reason, its use is not recommended for normally consolidated and light
25
over-consolidated soils.
Mohr Coulomb Yield Criterion- Two Major Drawbacks

2. Shear-Induced Dilation does not Terminate

• If the associated flow rule is assumed i.e. (dilation angle= friction angle),
then the Mohr-Coulomb criterion will always predict dilation whereas
the soil may either compress or dilate depending on whether it is
normally consolidated or heavily over-consolidated.
• Even for heavily overconsolidated soils, the Mohr-Coulomb criterion
tends to over-predict the amount of dilation if the soil is subjected to
prolonged shearing. This is because the Mohr-Coulomb model does not
have a mechanism by which the dilation can be terminated.

26
Mohr-Coulomb model with Associated Flow Rule (𝜓 = ’)

p’ → Plastic volumetric strain increment - d𝜀vp

q → Plastic shear strain increment - d𝜀sp

On the yield surface,

δq/δp’ = M

From associated flow rule

In a drained test, plastic volumetric expansion will occur as

long as plastic shear strain continues to increase. This is
obviously in contradiction to real soil behaviour which
normally shows volumetric dilation tailing off as the
ultimate state is reached.
27
Mohr-Coulomb model with Associated Flow Rule (𝜓 = ’)

In an undrained test, volumetric change cannot occur i.e. δ𝜀v = 0

Stress-strain curve of Mohr-Coulomb

model with associated flow rule under
undrained loading.

Stress will increase continuously with strain at a constant rate and

failure will not occur. This does not agree with real soil behaviour
at large strains close to ultimate failure.
28
• Above equation shows that dilation will continue indefinitely.
• The angle of friction for soils typically ranges from about 22° to about
33°. Hence, the value of M in triaxial compression will range
correspondingly from ~0.86 to ~1.3.
• In other words, the dilative volumetric strain increment is typically 0.86
to 1.3 times the plastic shear strain increment, the Mohr-Coulomb
model with associated flow rule.
• This is significantly higher than the maximum rate of dilation observed
on most soils.
• Hence, the use of Mohr-Coulomb model with associated flow rule is
usually not recommended.

29
 Mohr-Coulomb model with Non-Associated Flow Rule (𝜓 ≠ ’)

• The problem of overestimation of dilation can be mitigated by using the

Mohr-Coulomb model with non-associated flow rule.
• Angle of dilation is difficult to measure and often not measured in
laboratory

On the yield surface, δq/δp’ = M

From non-associated flow rule: δ𝜀vp = -N δ𝜀sp

The drained test behaviour is similar to that for associated flow rule, with plastic
volumetric expansion occurring as long as plastic shear strain continues to increase.
This is obviously in contradiction to real soil behaviour which normally shows volumetric
dilation tailing off as the ultimate state is reached 30
Mohr-Coulomb model with Non-Associated Flow Rule (𝜓 ≠ ’)
Let us consider a case of the undrained triaxial test i.e. δ𝜀v = 0

We also note that when ψ = 0, then N = 0 and H = 0.

In other words, if the angle of dilation is zero, then the post-yield undrained
stress-strain curve of the Mohr-Coulomb model will be a horizontal straight line.
31
 Effect of dilation angle on stress-strain curve
Plastic modulus is given by the
gradient of the predicted post-
yield stress-strain curve.
The lower the angle of dilation,
the lower is the plastic modulus
and the predicted curve
intersects measured curve at a
large shear strain.

With a lower angle of dilation, the Mohr-Coulomb model will under-predict the stress
for a given shear strain for a wide range of strain, and is therefore more conservative
in the respect.

Hence, one approach to selecting an appropriate angle of dilation is to first pick a

limiting strain level which is unlikely to be exceeded, then by the type of analysis
above, select an angle of dilation which does not cause the predicted curve to rise
above the measured curve for all strain levels below the limiting value.

After the analysis, if the limiting strain level is indeed not exceeded, then the limiting
32
strain level will have to be raised and the angle of dilation lowered.
References

• ‘Geotechnical Constitutive Modelling’ -Lecture notes by Professor Lee

Fook Hou, National University of Singapore.
• ‘Geotechnical Analysis’ - Lecture notes by Professor Goh Siang Huat,
National University of Singapore.
• ‘Geotechnical Analysis’ - Lecture notes by Professor Harry Tan, National
University of Singapore.
• Brinkgreve, R.B.J., 2005. Selection of soil models and parameters for
geotechnical engineering application. Evaluation, Selection, and
Calibration, ASCE 2005.
• David M Potts and Lidija Zdravkovic, “Finite Element Analysis in
Geotechnical Engineering: Theory”, Thomas Telford.

33