120 | Section 1.

Elastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Section 1 Elastic Material Properties

Elastic materials return to their original state after the external load is removed without any permanent
deformation. FEA NX includes various linear elastic and nonlinear elastic material models. The properties of
each material model are explained in this chapter. Table 4.1.1 lists the available elastic materials for each
element.

Table 4.1.1 Available elastic

materials for each element type
Element type

Axisymmetric Solid
Truss/Elastic Link

Plane Stress

Plane strain
Failure condition

Interface

Geogrid
Beam

Shell

Solid
Linear Elastic Isotropic ∨ ∨ ∨ ∨ ∨ ∨ ∨ ∨

Linear Elastic
∨ ∨ ∨
2D Orthotropic
Linear Elastic
∨ ∨ ∨ ∨ ∨
Transversely Isotropic

Interface Elastic ∨

Nonlinear Elastic (1D) ∨

Jardine ∨ ∨ ∨

D-Min ∨ ∨ ∨

Hyperbolic
∨ ∨ ∨
(Duncan-Chang)

Isotropic materials have the same properties in any arbitrary direction. Linear elastic isotropic materials based
1.1 on Hooke's law can be used on all elements, excluding some special elements. Using the modulus of elasticity
Isotropic Materials E , Poisson's ratio  and coefficient of thermal expansion  , the stress-strain relationship for 3D isotropic
materials can be expressed as follows:
ANALYSIS REFERENCE Chapter 4. Materials

 E (1  )  E (1  )  E (1  ) 
 (1   )(1  2 ) 0 0 0 
(1   )(1  2 ) (1   )(1  2 )
 
 E (1  )  E (1  ) 
0 0 0 
 xx   (1   )(1  2 ) (1   )(1  2 )   T 
     xx 
 yy   E (1  )   yy  T 
 
0 0 0 
 zz 
 (1   )(1  2 )  zz  T 

      (4.1.1)
 xy   E  
symmetric 0 0  xy 
 yz   2(1   )   yz 
   
  
  zx 
  E
0  

 zx

 2(1   ) 
 E 
 
 2(1   ) 

For 2D analysis,  yz   zx   yz   zx  0 and particularly for plane strain analysis,  zz  0 .

 E E 
 0 
 xx  1  1  2
2
  xx  T 
   E E  
 yy    0   yy  T  (4.1.2)
   1   1  2
2
   xy 
 xy   E 
0 0
 2(1   ) 

As  approaches 0.5, the 1  2  2 term approaches‘0 (zero)’, and this can cause numerical errors. Hence, the
range of Poisson's ratio for isotropic materials is restricted as follows:

1.0   0.5 (4.1.3)

Increase in modulus of elasticity with height

The change in modulus of elasticity with height can be simulated. If the change is ‘0 (zero)’, a constant modulus
of elasticity is used, and if it is not ‘0 (zero)’, the modulus of elasticity with reference to a certain height can be
calculated as follows:

E  Eref   yref  y  Einc y  y 

ref
(4.1.4)
E  Eref y  y 
ref

Eref : Input modulus of elasticity

Einc : Incremental slope of modulus of elasticity

yref : Depth where Eref is measured

Section 1. Elastic Material Properties | 121

122 | Section 1. Elastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Figure 4.1.1 Conceptual diagram

for incremental modulus of depth
elasticity

yref elastic
modulus
Eref

Einc

The y in equation (4.1.4) represents the position of the integral point where the element calculation occurs
for the current finite element method. If the integral point y is positioned higher than yref , the modulus of
elasticity can have a negative(-) value and so, the Eref is used as the minimum value for modulus of elasticity
E.

Orthotropic material is one that has different material properties or strengths in different orthogonal directions.
1.2 The structure is geometrically orthotropic with significant different stiffness in horizontal and vertical direction.
Orthotropic Materials It is known that the axial stiffness in vertical direction is larger than the effective stiffness in horizontal direction.
The stress-strain relationship for 3D orthotropic material can be expressed as follows:

1   23 32  21   31 23  31   21 32 

 EE 0 0 0 
E2 E3 E2 E3
 11     11  11T 
2 3

       32   12 31

0   22   22 T 
1 13 31
 22  
0 0
E E  E1E3  
 33   1 3     T 
   1   12 21   33 33 

  12   E1E2 
0 0 0   12 
 23   
  (4.1.5)
   symmetric G12 0 0  23

  31   G23 0 
   31 
 
 G31 
1   12 21   23 32   31 13  2 21 32 13

E1E2 E3
 ANALYSIS REFERENCE Chapter 4. Materials

The stress-strain relationship for 2D orthotropic material is as follows:

 E1  21E1 
1    0 
1   12 21
 11   12 21   11  11T 
    12 E2 E2  
 22    0   22   22 T  (4.1.6)
   1   12 21 1   12 21   12 
 12    
0 0 G12 
 
 

The stress-strain relationship for shear in horizontal direction is as follows:

 31  G31 0   31 
    (4.1.7)
 23   0 G23   23 

In case of orthotropic material, the following properties should be satisfied.

E2 E E E E E
 21
2
 ,  122  1 ,  32
2
 3 ,  23
2
 2 ,  132  1 ,  31
2
 3 (4.1.8)
E1 E2 E2 E3 E3 E1
1 12 21  23 32  3113  2 21 3213  0 (4.1.9)

Transversely isotropic materials are material models defined by the isotropic cross-section and the axis
1.3 perpendicular to it. The axis perpendicular to the cross-section displays symmetric physical behavior. Hence,
Transversely Isotropic the physical properties are the same within the cross-section and different in the perpendicular direction.
Materials Transversely isotropic materials display different physical properties (Modulus of elasticity, Poisson's ratio,
Shear modulus) in each perpendicular direction.

► out-of-plane cross-sectional properties : E1 , 12   13  , G12   G13 

E2   E3   23 G23
► in-plane cross-sectional properties : , ,

Here, E1 is the modulus of elasticity of the perpendicular axis to the cross-section, and  12 ,  13 and G12 , G13

are the Poisson's ratio and shear modulus respectively in the plane created by the perpendicular axis and other
axes of the cross-section. However, because the physical properties are axisymmetric about the perpendicular
axis to the cross-section, 12  13 , G12  G13 . E2 and E3 are the modulus of elasticity for each axis of the
cross-section,  23 is the Poisson's ratio and G23 is the shear modulus. Likewise, because the material is isotropic
in the horizontal direction, E2  E3 .

Section 1. Elastic Material Properties | 123

124 | Section 1. Elastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

On the other hand, transversely isotropic materials and the out-of-plane Poisson's ratio have the following
relationship: 12 / E1  21 / E2 , 13 / E1   31 / E2 . In other words,  12 and  21 need to be distinguished when
defining the transversely isotropic material properties.

MCS / local coordinate system

FEA NX has a local coordinate system defined by the dip angle within the MCS to simulate the slope of the
transversely isotropic material model. However, whilst the behavior of transversely isotropic material models is
defined by the local coordinate system, the constitutive equation is expressed in the MCS. Hence, coordinate
system conversion is needed for the constitutive equation between the local coordinate system and MCS using
the dip angle (Detailed information on this is continued in the Constitutive equation and coordinate system
conversion and Defining interface direction).

Figure 4.1.2 displays the 2D model where the MCS  Z axis and the x axis (tangent direction of the local
coordinate system) have an angle  .

Figure 4.1.2 2D transversely MCS  Z

x
isotropic material model
E1 ,  12 , G12

E2 ,  23 , G23

z
MCS  X

Constitutive equation and coordinate system conversion

The 3D elastic constitutive equation for the local coordinate axis xyz is equation (4.1.10).
ANALYSIS REFERENCE Chapter 4. Materials

1 v12 v12 
E   0 0 0 
E1 E1
 1 
 1 v23 
  0 0 0 
  x   E2 E2    x 
     
 y   1
0 0 0   y 

  z    E2   z  
    
(4.1.5)
   1
x y 
symmetric 0 0  x y  
 y z    G12   y z  
     
 z x  
   1
0   z x  
 G23 
 
 1 
 G12 

Generally, the MCS XYZ and local coordinate system xyz are not the same. Because equation (4.1.10) is the
constitutive equation corresponding to the local coordinate system, the stiffness matrix of the local coordinate
system needs to be converted to the MCS stiffness matrix using the normalized direction vector of the local
coordinate system in the MCS. The following equations (4.1.11) and (4.1.12) are the coordinate system
conversion equations for stress and strain respectively:

σ xyz  R σ XYZ (4.1.11)

ε xyz  R ε XYZ (4.1.12)

 nX2 nY2 nZ2 2nX nY 2nY nZ 2nZ nX 

 2 
 sX sY2 sZ2 2s X sY 2sY sZ 2 sZ s X 
 t2 tY2 tZ2 2t X tY 2tY tZ 2tZ t X 
R   X  (4.1.13)
 nX s X nY sY nZ sZ nX sY  nY s X nY sZ  nZ sY nZ s X  n X sZ 
s t sY tY sZ t Z s X tY  sY t X sY tZ  sZ tY sZ t X  s X t Z 
 X X 
 t X nX tY nY tZ nZ t X nY  tY n X tY nZ  t Z nY t Z n X  t X nZ 

 nX2 nY2 nZ2 nX nY nY nZ nZ nX 

 2 
 sX sY2 sZ2 s X sY sY sZ sZ s X 
 t2 tY2 tZ2 t X tY tY tZ tZ t X 
R   X  (4.1.14)
 2n X s X 2nY sY 2nZ sZ nX sY  nY s X nY sZ  nZ sY nZ s X  nX sZ 
 2s t 2sY tY 2 sZ t Z s X tY  sY t X sY tZ  sZ tY sZ t X  s X t Z 
 X X 
 2t X nX 2tY nY 2tZ nZ t X nY  tY nX tY nZ  tZ nY tZ nX  t X nZ 

n   nX , nY , nZ  , s   sX , sY , sZ  , t   t X , tY , tZ  are the normalized direction vectors of the x , y , z  axis

respectively in the MCS XYZ . For 2D problems, nZ  sZ  t X  tY  0 , tZ  1 .

Section 1. Elastic Material Properties | 125

126 | Section 1. Elastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

On the other hand, the following equation is established between the two conversion matrices R , R .

R  RT R  RT (4.1.15)

Using this, the stress strain relationship equation in the local coordinate system can be expressed as the MCS
stress and strain as follows:

σ xyz  D xyzε xyz 

  T
 σ xyz  R σ XYZ   R σ XYZ  D xyz R ε XYZ (4.1.16)
 ε R ε 
 xyz  XYZ 

Rearranging the equation above gives the constitutive equation of the MCS XYZ , as shown in equation (4.1.17).

1
σ XYZ   RT Dx1yz R  ε XYZ (4.1.17)

Defining interface direction

The axial direction vectors n , s , t of the local coordinate system xyz defined in the MCS XYZ are defined
by the dip angle 1 and dip direction 2 . Figure 4.1.3 displays the definition of the two angles.

Figure 4.1.3 MCS according to dip Z

angle and dip direction Z

declination t
N n

2
3 X 1 s
1
s
s
s

1 is the angle between the MCS Z axis and the sliding plane corresponding to the y  z plane of the local
coordinate system with reference to the Y axis, and 2 is the angle of rotation for the y ' axis of the sliding
plane in the Z axis direction, with reference to the N axis of the X  Y plane. Here, 1 needs to be in the
[0 ,180 ] domain, and 2 needs to be in the [0 ,360 ] domain.
 ANALYSIS REFERENCE Chapter 4. Materials

Generally, reference axes N and X of the sliding plane and horizontal plane are not the same. Hence the
auxiliary angle 3 , which subtracts the declination corresponding to the angle between the reference axes of
the two planes from 2 , is used when composing the actual transformation matrix.

3  2  declination (4.1.18)

From the definition above, the vectors n , s , t that form the equations (4.1.13) and (4.1.14) for 3D element
transformation matrices can be obtained. Here, n is the normal vector of the inclined plane and s and t are
vectors on the inclined plane that are perpendicular to it.

3D:

 sin 1  cos  3   cos 1  cos  3   sin  3 

n    sin 1  sin  3  , s    cos 1  sin  3  , t  cos  3  (4.1.19)
 cos 1    sin 1   0 

Meanwhile, because the MCS axes of 2D elements are different from that of 3D elements in FEA NX, the
definition of the axial direction vector and dip angle of the local coordinate system changes. The dip angle 1
is defined as the angle between the MCS Y axis and x axis of the local coordinate system. However, because
the rotation angle of the vertical axis Y is not considered, the dip direction 2 and declination is not used.
The figure below displays the definition of the axis direction vector and dip angle of the 2D local coordinate
system xyz .

Figure 4.1.4 Definition of dip angle Y

and dip direction on 2D MCS n

1

s
Z

Section 1. Elastic Material Properties | 127

128 | Section 1. Elastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

From the definition above, the vectors n , s , t that form the equations (4.1.13) and (4.1.14) for 2D element
transformation matrices can be obtained:

2D:

 sin 1   cos 1  0

n  cos 1  , s    sin 1  , t   0  (4.1.20)
 0   0   1

Interface affiliated elements (interface, shell interface, pile elements) are models used to simulate interface
1.4 behavior. The linear stiffness matrix used for these elements are applied such that the elements are separated
Elastic Material of or do not penetrate each other.
Interface Affiliated
The linear stiffness matrix for interface, pile elements are expressed as equation (4.1.21) and the added
Elements rotational DOF form for shell interface elements is expressed as equation (4.1.22).

 kn 0 0
D   0 kt 0  (4.1.21)
 0 0 kt 
 kn 0 0 0 
0 kt 0 0 
D  (4.1.22)
0 0 kt 0 
 3 
0 0 0 knt / 12 

kn : Normal stiffness
kt : Tangential stiffness
t : Thickness of shell interface element

Precautions need to be taken on the units for the stiffness above. For example, when the SI system of units is
used, the units for stiffness are N / m 3 , not the units for the modulus of elasticity N / m 2 .
Equation (4.1.23) is recommended for calculating the stiffness, where the modulii of elasticity around the target
element is divided by the characteristic length. Here, the characteristic length ( lch ) is recommended for the
thickness of line interface, shell interface, and pile elements. The use of the square root of the element area ( A )
is recommended for the plane interface.

E G
kn   , kt   kn or  (4.1.23)
lch lch
 , : Scale factor
 ANALYSIS REFERENCE Chapter 4. Materials

E ,G : Modulus of elasticity, Shear modulus

The scale factor in the equation above needs to be selected empirically depending on the analysis. If the scale
factor is too large, numerical problems can occur. If it is too small, accurate result values for the relative
displacement of the interface element cannot be obtained. A value of '0.1 ~ 10' is recommended.

The Coulomb friction model is provided for as the nonlinear material model for interface, shell interface
1.5 elements, and the details are explained in Chapter 2. For pile elements, the multiple curve input or a value to
Nonlinear Elastic simulate perfectly plastic behavior is used as the nonlinear material model.
Behavior of FEA NX supports the following nonlinear elastic behavioral models for truss or embedded truss elements:

Truss/Embedded Truss ► Compression only behavior

Elements ► Gap behavior
► Tension only behavior
► Hook behavior
► User defined nonlinear elastic behavior

Figure 4.1.5 Various nonlinear

elastic behavioral models
 

 
allowable
tension stress gap strain

(a) Compression only behavioral mode (b) Gap behavioral model

 

hook strain
allowable
compression
stress




(c) Tension only behavioral model (d) Hook behavioral model

Section 1. Elastic Material Properties | 129

130 | Section 1. Elastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Gaps and hooks have inputs with a length unit. Internally, the gap strain and hook strain are calculated using the
element length.

FEA NX supports the following nonlinear elastic behavioral models for elastic link elements:
1.6
Nonlinear Elastic ► General behavior
Behavior of Elastic Link ► Tension only behavior
► Compression only behavior
Elements
► User defined nonlinear elastic behavior

Tension only, compression only and user defined nonlinear elastic behavior are defined in a similar way as the
nonlinear elastic behavior of truss and embedded truss in section 1.4. However, whilst the behavior of truss and
embedded truss elements are represented using the stress strain relationship, elastic link elements do not have
sectional properties and hence, their behavior is defined using the force displacement relationship shown in
Figure 4.1.6, not the stress strain concept. Because tension only and compression only behaviors do not require
separate inputs for allowable strength and allowable displacement, a user defined function needs to be used for
application.

Figure 4.1.6 Tension only and P P

compression only behavior of compression
elastic link elements

d

d
tension

(a) Tension only (b) Compression only

In Figure 4.1.6,  d represents the relative displacement between connected nodes and P represents the
internal member forces.

Jardine1 suggested the use of material models that define nonlinear behavior to consider the nonlinear behavior
1.7 that occurs in small strain states for clay.
Jardine
The Jardine model is a nonlinear elastic model that can simulate nonlinear behavior at small strain states, and
the Tresca model is used for plastic analysis when the stress of the material is larger than the input shear
strength. Here, the behavior is completely plastic and the stiffening behavior is not considered.

1 Jardine, R. J., Symes, M. J. and Burland, J. B. "The measurement of soil stiffness in the triaxial apparatus," Geotechnique
34, No. 3, 323-340, 1984.
 ANALYSIS REFERENCE Chapter 4. Materials

Nonlinear elastic behavior

The Jardine model derives the nonlinear formula based on the relationship between the secant modulus of
elasticity and axial strain measured from the undrained triaxial compression test. The undrained triaxial
compression test applies an incremental load in the axial direction of a cylindrical sample, and the stress along
the circumferential side is maintained.

The secant modulus of elasticity ( Eu ) can be directly calculated from the measured value from the triaxial
compression test.

 a   a;0
Eu  (4.1.23)
a

a : Axial strain
a : Axial stress
 a;0 : Initial(when  a  0 ) axial stress

Because the Jardine model assumes the nonlinear relationship between the secant modulus of elasticity and
axial strain, the modulus of elasticity in the elastic region can be defined using the following equation:

    
Eu  G   F  G  cos  log a   (4.1.25)
  A  

A : Strain at maximum stiffness

F : Maximum stiffness value
G : Average stiffness value

Figure 4.1.7 Jardine parameters Eu

F test data

best fit
G

2G  F

A  min B  max C
a

Section 1. Elastic Material Properties | 131

132 | Section 1. Elastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

A , F , G can be directly computed from the experimental stiffness-strain curve, and the B , C
parameters needed for computing  ,  in the following equation are assumed to be the strain at the average
stiffness and minimum stiffness. In this case, the average stiffness becomes equation (4.1.25) where  / 2 input
is for the cosine term, and the minimum stiffness is where  is input.

log 2
 (4.1.26)
 log(C / A) 
log  
 log( B / A) 
 /2
 
(4.1.27)
 log  B / A  

F , G do not need to be the same as the experimental maximum, average stiffness. F is the maximum
value of the trend line that best fits the data. When the strain is outside the range of the maximum strain (  max )
and minimum strain (  min ), the tangent modulus of elasticity is assumed as a constant value.

The general value of  min represents the minimum strain of the experimental data values. A value that is stable
in the plastic region needs to be selected for  max . If  max the value is too large, a negative (-) elastic tangent
stiffness is computed, which can cause numerical instability. Hence, the  max value is generally defined to be
smaller than C .

The equivalent stiffness of the Jardine model is as follows:

2
 eq  1   2    2   3    3  1  
2 2 2
(4.1.28)
3
1  2 3 : Major strain at the elastic state
, ,
Here,  eq can be calculated as 3 a at the stress state of the undrained triaxial test ( 1   a ,
 2   3  1/ 2  a ).
The tangent modulus of elasticity Eut can be expressed in the same way as equation (4.1.25), and expressing
this as a relationship equation with  eq is as follows:

Eu  f1  eq   G   F  G  cos  I   (4.1.29)

Eut  f 2  eq   G   F  G  cos  I   
 F  G  I  1
sin  I   (4.1.30)
2.303

And,
ANALYSIS REFERENCE Chapter 4. Materials

  
I  log  eq  (4.1.31)
 3A 

The equivalent linear strain considering the boundary range  max and  min is as follows.

 eq min   min 3 ,  eq max   max 3 (4.1.32)

The tangent modulus of elasticity is assumed to be a constant value outside the boundary range, and the general
form of Eu can be expressed as follows:



 f1   eq min   eq   eq min

Eu   f1   eq   eq min   eq   eq max (4.1.33)

 eq max
 2  eq max 
f 

 f1  eq max   f 2  eq max 
 eq
  eq   eq max


The computed average tangent stiffness from equation (4.1.33) is used when calculating the increment strain
through recursive calculations. When calculating the actual stiffness, the secant stiffness equation (4.1.24) is
used for accurate computation.

The relationship between the tangent and secant modulus of elasticity in the triaxial test is as follows:

d a d  Eu a 
Eut   (4.1.34)
d a d a

The axial incremental stress  a due to the axial incremental strain  a , found from the relationship between
the given axial strain and secant modulus of elasticity, can be expressed as follows:

 Eu  a0   a   Eu0 a0
 a0  a
Eut d a   Eu a 
 a0  a
 a   0
(4.1.35)
 a0 a

To express the relationship between incremental stress and incremental strain linearly, the average modulus of
elasticity Eut is used, and can be expressed as follows.

Eu eq  Eu0 eq
0

E ut  (4.1.36)
 eq   eq0

Section 1. Elastic Material Properties | 133

134 | Section 1. Elastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

 eq : Renewed equivalent strain

Eu : Secant stiffness, computed from equation (4.1.33)

The 3D material stiffness matrix can be expressed as follows, using the average modulus of elasticity:

1    0 0 0 
  1    0 0 0 
 
   1  0 0 0 
 
 0 1  2
D
E ut 0 0 0 0  (4.1.37)
1  1  2   2 

 0 1  2
0 0 0 0 
 2 
 1  2 
 0 0 0 0 0 
 2 

The D-min model is a sectioned linear model applied to general rocks (hard rock, soft rock etc.), proposed by
1.8 Japan Central Research Institute of Electric Power Industry (CRIEPI), Hayashi, Hibino. Sectioned linear models
D-min have different stiffnesses for each construction step, but are normalized such that the stiffness has a fixed value
within a construction step.

It is assumed that the modulus of elasticity decreases and the Poisson’s ratio increases as the Mohr circle
approaches the failure envelope. Hence, the relative distance between the Mohr circle and failure envelope
determines the modulus of elasticity and Poisson’s ratio of each section. The material property values of this
model are constant for each load step and so, repeated analysis is not required for each load step.

The failure envelope equation can be expressed as follows:

a
  
  1 (4.1.38)
R  t
 : Hydrostatic stress
 : Shear stress
a : Mohr circle coefficient
t : Tensile strength
R : Shear strength

The relationship between the failure envelope and Mohr circle is expressed in Figure 4.1.8
 ANALYSIS REFERENCE Chapter 4. Materials

Figure 4.1.8 Failure envelope and

Mohr circle of D-min model  1  a  2 
  
  1
R   t
R
d min D1
d min
R 
 2
t  1
2

t m 3

 compression
D2

Failure is determined using the buffer index ( R ), as shown in Figure 4.1.8. If the buffer index is larger than '1',
it is in the elastic region and if the buffer index is less than '0', failure is assumed.

R  k  R' 0.0  R  1.0 (4.1.39)

Here, the modified buffer index is:

d min
R'  (4.1.40)
  3
t  1
2
d min : Minimum distance between failure envelope and Mohr circle

k : Buffer index from user input variable

The factor of safety ( Fs ) is as follows:

min  D1 , D2 
Fs  (4.1.41)
 1   3 
 
 2 

Here, the buffer index is computed at the integral point, and this can be used to compute the modulus of
elasticity and Poisson's ratio of the next step:

Section 1. Elastic Material Properties | 135

136 | Section 1. Elastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

E  R m ( Ei  Ecr )  Ecr
(4.1.42)
  R n ( i   cr )  cr
Ei : Initial modulus of elasticity
Ecr : Critical modulus of elasticity
m : Nonlinear material coefficient
i : Initial Poisson's ratio
 cr : Critical Poisson's ratio
n : Nonlinear material coefficient

The Mohr circle coefficient ( a ) and buffer index ( k ) increases with the increase in initial modulus of elasticity
( Ei ). The relationship between the Mohr circle coefficient ( a ) of various rocks, buffer index ( k ) and initial
modulus of elasticity ( Ei ) is shown below in table 4.1.2. The data is based on the triaxial compression test and
was put together by the Japan Road Traffic Information Center in 1986.

Table 4.1.2 Parameters based on Initial Elasticity modulus

initial Elasticity modulus (JARTIC, Buffer index ( k ) Mohr circle coefficient ( a )
1986) ( Ei , kgf / cm2 )
100  Ei  1,000 2.0 1.0
1,000  Ei  10,000 4.0 2.0
10,000  Ei  100,000 6.0 3.0
100,000  Ei 10.0 4.0

Ground stress-strain behavior becomes nonlinear as it approaches the failure criterion, and this can be simulated
1.9 by modifying the foundation modulus. The function proposed by Duncan and Chang2 is used to calculate the
Hyperbolic Model foundation modulus. The stress-strain curve of the function is a hyperbola and the foundation modulus is a
(Duncan-Chang) function of confining stress and shear stress. It is very useful because nonlinear material models only need
material properties that can be easily obtained from the triaxial compression test or literature,

The Duncan and Change nonlinear stress-strain curve displays a hyperbolic form between the axial strain space
generated by shear stress 1   3 . Three foundation modulus are needed depending on the stress state and
stress path; the initial modulus Ei , tangent modulus Et , and unloading-reloading modulus Eur . (Refer to
Figure 4.1.9)

2 Duncan, J. M., and Chang, C.-Y. “Nonlinear Analysis of Stress and Strains in Soils,” J. Soil Mech. Found. Div., ASCE
96, 5 (1970), 1629-1653.
 ANALYSIS REFERENCE Chapter 4. Materials

Figure 4.1.9 Nonlinear stress-

strain behavior
Et
1
B
A

1   3
Eur

1
Ei C
1
O strain

The modluii and coefficient of the nonlinear elastic model can be obtained from the graph with a vertical axis of
the ratio between modulus of elasticity and atmospheric pressure ( E pa ) or ratio between bulk modulus and
atmospheric pressure ( Bm pa ), and a horizontal axis of the ratio between maximum confining pressure and
atmospheric pressure (  3 pa ) in log scale, as shown in Figure 4.1.10. The initial loading coefficient (K) can be
obtained when the vertical axis is E pa at the point where  3 pa  1 and slope at this point can be used to
calculate the coefficient n for the initial stiffness. The bulk modulus index m can be found from the slope when
the vertical axis is Bm pa .

Figure 4.1.10 Determination of

nonlinear ground material
properties
103

1
E/Pa or Bm/Pa

n or m
102
K modulus

101

100 101 102

3/Pa

The bulk modulus Bm is defined by equation (4.1.43).

Bm 
 1   2   3  3
(4.1.43)
 v

Section 1. Elastic Material Properties | 137

138 | Section 1. Elastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

 : Amount of principal stress change,

 v : Amount of principal stress change,

Initial modulus
When the ground experiences a ‘0(zero)’shear stress (when 1   3  0 ), the stress-strain behavior is calculated
using the initial modulus Ei . This initial tangent modulus is controlled by the confining stress  3 and can be
calculated using equation (4.1.44).

n
  (4.1.44)
Ei  K L pa  3 
 pa 
Ei : Initial tangent modulus, a function of confining stress
KL : Loading coefficient
pa : Atmospheric pressure
3 : Confining stress
n : Index for defining the effects of confining stress on initial modulus

If the index n is 1.0, the initial modulus Ei is directly proportional to the confining stress. If it is ‘0(zero)’, Ei
is unrelated to the confining stress.

If the confining stress is in the tensile state, the initial modulus can be ‘0 (zero)’ or ‘-(negative)’. To prevent this,
FEA NX sets a lower bound for the confining stress. The set value is 0.01 pa .

Tangent modulus
The ground is known to follow the load path when it experiences a larger shear stress than it has experienced
before. The constitutive behavior is dominated by the tangent modulus Et when the load path is followed. This
tangent modulus can be defined as a function of the ground material properties, triaxial deviatoric stress
1   3 and confining stress  3 of a Duncan-Chang model.

 R    3 1  sin   
2

Et  1  f 1  Ei (4.1.45)
 2c cos   2 3 sin  

Et : Tangent modulus
Ei : Initial tangent modulus
 : Internal friction angle of the ground
ANALYSIS REFERENCE Chapter 4. Materials

c : Cohesion of the ground

Rf
: Ratio between maximum shear stress and asymptote of the hyperbola (generally a value
of 0.75 ~ 1)

Here, the minimum value of Et can be restricted. The basic minimum tangent modulus is pa . If this value is
too small, it can cause convergence problems.

Unloading-reloading modulus
Nonlinear models use the unloading-reloading modulus Eur when the ground is unloading from a large shear
stress state. This coefficient is calculated in a similar manner to the initial modulus, except that the unloading-
reloading coefficient number K ur is used instead of KL .

n
 
Eur  K ur pa  3  (4.1.46)
 pa 

Unlike the tangent modulus, this modulus is not affected by the shear stress. If the unloading-reloading
coefficient number K ur is not defined, it is defined to be the same as the loading coefficient number KL .

Friction Angle
The friction angle can be defined by incremental form according to the confining pressure.

 3 
  0   lg   (4.1.47)
 pa 

Poisson’s ratio
The Poisson's ratio of nonlinear models are either set as a constant unrelated to the stress state or calculated
from the bulk modulus of the soil depending on the confining stress. For the latter case, the bulk modulus can
be found using equation (4.1.47).
m
 
Bm  K b pa  3  (4.1.48)
 pa 
Bm : Bulk modulus
Kb : Bulk modulus number
m : Bulk modulus index

The relationship between the Poisson’s ratio and bulk modulus can be defined from the elastic theory, as shown
in equation (4.1.48).

Section 1. Elastic Material Properties | 139

140 | Section 1. Elastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

E
Bm  (4.1.49)
3 1  2 

If the Poisson’s ratio in the equation above is ‘0(zero)’, Bm  E 3 and if the Poisson’s ratio is 0.49, Bm  17 E ,
the calculated Poisson’s ratio is limited to '0 ~ 0.49'.
And the Poisson’s ratio in Duncan-Chang model can be defined by the experimental constant not the bulk
modulus.

 
G  F lg  3 
t   pa  (4.1.50)
2
 
 
 D  1   3  
1  
 3   
m
 Rf
 K pa  p  1      1   3   
  a    1 3f  

Failure region
The failure condition of a nonlinear elastic model cannot be defined. However, to show that the shear region is
large for this material, the failure region is defined as the region that satisfies the following condition:

1   3 1   3
 sin   R f c cos  (4.1.51)
2 2

The failure ratio R f in the Duncan-Chang equation is used as shown in equation (4.1.50):

1   3  f  R f  1   3 ult (4.1.52)

The ultimate strength  1   3 ult term represents the asymptote which the hyperbolic stress-strain curve
approaches at high strains. Also, 1   3  f is the deviatoric strain at failure.
 ANALYSIS REFERENCE Chapter 4. Materials

Section 2 Plastic Material Properties

FEA NX includes various plastic material models to simulate actual ground and structural phenomena. This
section briefly introduces the plastic theory used and the properties of each material model. The table below
lists the available plastic materials for each element.

Table 4.2.1 Available plastic

materials for each element type
Element type

Axisymmetric Solid
Plane Stress

Plane strain
Failure condition

Interface

Geogrid
Beam
Truss

Shell

Solid
von Mises ∨ ∨ ∨ ∨ ∨ ∨

Tresca ∨ ∨ ∨

Mohr-Coulomb ∨ ∨ ∨

Drucker-Prager ∨ ∨ ∨

Strain-Softening ∨ ∨ ∨

Modified Cam Clay ∨ ∨ ∨

Jointed Rock ∨ ∨ ∨

Modified Mohr
∨ ∨ ∨
Coulomb

Hoek Brown ∨ ∨ ∨

Generalized
∨ ∨ ∨
Hoek Brown

Modified UBCSAND ∨ ∨ ∨

Sekiguchi-Ohta
∨ ∨ ∨
(Inviscid)

Soft Soil ∨ ∨ ∨

Hardening Soil with

∨ ∨ ∨
Small strain stiffness

Section 2. Plastic Material Properties | 141

142 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Generalized
∨ ∨ ∨
SCLAY1S

CWFS ∨ ∨ ∨

Inverse Rankine ∨

GeoGrid ∨

Coulomb Friction ∨

Janssen ∨

Principal stress invariance

2.1 Principal stress invariance is a convenient method of expressing the yield function. The stress induced at an
Failure Criterion and arbitrary point within the material can be expressed using the following equation, which uses the direction
Invariance vector n j that defines the principal stress direction:

 ij  ij  n j  0 (4.2.1)
 ij : Kronecker delta

n j  0 in the equation (4.2.1) above, and the necessary and sufficient condition for equation (4.2.1) is as follows:

 ij  ij  0 (4.2.2)

The matrix equation (4.2.2) can be expressed as a cubic equation for principal stress, as shown below:

 3  I1 2  I 2  I 3  0 (4.2.3)

Here,
I1   x   y   z   ii

I 2   x y   y z   z x    xy2   yz2   zx2    I1   ij ji 

1 2
(4.2.4)
2
 x  xy  xz
1 1 1
I 3   yx  y  yz   ij jk ki  I1 ij ji  I13
3 2 6
 zx  zy  z

I1 , I 2 , I3 can be expressed using the principal stresses 1 ,  2 ,  3 as follows.

ANALYSIS REFERENCE Chapter 4. Materials

I1  1   2   3
I 2  1 2   2 3   31 (4.2.5)
I 3  1 2 3

Deviatoric stress invariance

The stress tensor  ij can be divided into the hydrostatic pressure and invariant stress components, as shown
below:

 ij  sij   m ij (4.2.6)
Here,  m   x   y   z  / 3  I1 / 3 and represents the average stress. Also, sij   ij   m ij is the deviatoric
stress and represents the pure shear state.

The deviatoric stress invariance can be expressed as shown below:

sij  s ij  0 (4.2.7)

Equation (4.2.7) can be expressed as follows:

s 3  J1s 2  J 2 s  J 3  0 (4.2.8)

Here,
J1  sii  sx  s y  sz  0
1
J2  sij s ji
2
  x   y    y   z    z   x     xy2   yz2   zx2
1 2 2 2

6  (4.2.9)
sx  xy  xz
1
J 3  sij s jk ski   yx s y  yz
3
 zx  zy sz

J1 , J 2 , J3 can be expressed using the deviatoric principal stresses s1 , s2 , s3 as follows:

J1  s1  s2  s3  0

J2 
2
 s1  s2  s3   16 1   2 2   2   3 2   3  1 2 
1 2 2 2
(4.2.10)

J 3   s13  s23  s33   s1s2 s3

1
3

Section 2. Plastic Material Properties | 143

144 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

I1 , I 2 , I 3 , J1 , J 2 , J3 are all scalar invariants, which have properties independent of the coordinate axes.
To conveniently express the yield function geometrically, I1 , J 2 , J3 invariants are often used.

Geometric meaning of the three stress invariants

Figure 4.2.1 Stress state definition 1

in principal stress space
P  1 ,  2 ,  3 

0 r
1   2   3

 N

O
3

2

Vector OP can be defined when point P 1,  2 ,  3  is expressed as an arbitrary stress state in the principal
stress space, as shown in Figure 4.2.1. Vector OP can be divided into vector ON , which follows the hydrostatic
pressure axis; and vector NP , which exists in the deviatoric plane perpendicular to the hydrostatic pressure axis.
Their size is as follows:

1
ON    I1
3 (4.2.11)
NP  r  2 J 2

Vector NP needs to be rotated by 0 in the 1 axis to define point P on the deviatoric plane. Here, 0 is
called the similarity angle and its equation is as follows:

1 3 3 J 
 0  cos 1  3/2 
3
 (4.2.12)
3  2 J2 

Here, 0 has the following range:

 ANALYSIS REFERENCE Chapter 4. Materials


0  0  (4.2.13)
3

For numerical analysis, it is more convenient to use Lode's angle  rather than 0 and it can be defined using
the following equation:

1  3 3 J 
  sin 1   3
 (4.2.14)
3  2 J 23/2 


Here,    0  and has the following range:
6

 
   (4.2.15)
6 6

It is often more convenient to express the principal stress as an invariant stress when defining the yield function
of the material, and it can be rearranged using Lode's angle to give the following equation:

  2 
sin    3   
 1     1
  2 J2   I1  
 2   sin     1 (4.2.16)
  3   3 1
 3 sin    4    
  
  3  

Plastic materials display permanent deformation on structures even after the external load is removed, unlike
2.2 elastic materials. To express such behavioral properties, strain is formulated following additive decomposition,
Formulation of Plastic which divides strain into elastic and plastic components, as shown below:
Behavior
ε  ε el  ε pl (4.2.17)
ε : Total strain
ε el : Elastic strain
ε pl : Plastic strain

Because Hook's law defines the relationship between deformation and stress in the elastic region, applying this
to equation (4.2.17) and rearranging gives the following equation for stress:

Section 2. Plastic Material Properties | 145

146 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

σ  Dεel  D(ε  ε pl ) (4.2.18)

σ : Stress vector
D : Material stiffness matrix

The failure criterion defines the plasticity criteria and can be defined differently depending on the material
properties such as steel or concrete. The material failure criterion can be modeled in function form using various
experiments on the material. Generally, this function has variables that represent stress and hardening, and can
be expressed as follows:

f (σ,  )  0 (4.2.19)
f : Yield function
 : Hardening parameter

If the yield function f is equal to or smaller than ‘0’(zero), plastic flow does not occur and if f is larger than
'0', plastic flow occurs.

Plastic flow rule

Material failure induces plastic flow, and this plastic flow causes stress redistribution to maintain the equilibrium
state of the material. The plastic flow calculation is done in nonlinear form and the increment form is generally
used for formulation. The general values used for calculating the plastic flow in plasticity analysis for materials
are the incremental stress direction and plastic strain increment direction. The incremental stress direction is as
follows:

f i
ni  (4.2.20)
σ
n : Gradient vector representing the stress increment direction perpendicular to the failure surface
i : Number of yield functions

The plastic strain increment can be divided into the size and directional components using Koiter’s law as follows:

n
gi n
 p   i  i mi (4.2.21)
i 1 σ i1

Here, gi is the plastic potential function, which can be expressed as gi σ,   using stress and hardening
variable  , generally obtained from material tests. i is the plastic multiplier, and it needs to satisfy the
following Kuhn-Tucker condition:

f  0 , i  0 , i f  0 (4.2.22)
ANALYSIS REFERENCE Chapter 4. Materials

From the conditions above, plastic flow does not occur when the yield function f is smaller than 0 and i is
always 0. When plastic flow occurs ( i is larger than 0), the yield function is always 0. m is the vector that
defines the plastic strain increment in equation (4.2.21). Here, the method of defining the plastic strain
increment by f / σ , which uses the yield function f and not the plastic potential function g , is called the
associated flow rule and the method which uses the plastic potential function to define the plastic strain
increment direction by g / σ is called the non-associated flow rule. Using the non-associated flow rule on a
material model can suppress the excessive cubical expansion phenomena due to the discord between the stress
direction and strain direction. However, the amount of calculation increases because the stiffness matrix is
asymmetric and an asymmetric solver needs to be used.

The hardening variable  used for strain hardening can be defined using the dimensionless equivalent plastic
strain as shown below:


3
ε  Q εp
2 p T
(4.2.23)

Here,
 xxp  1 0 0 0 0 
0
 p 0
 yy   1 0 0 0 
0
  p 
 zz  0 0 1 0 0 
0
ε  p, Q
p
 (4.2.24)

 xy  0 0 0 1/ 2 0 0 
 p  0 0 0 0 1/ 2 0 
 yzp   
 zx  0 0 0 0 0 1 / 2

Stress Return Method

► Implicit backward Euler method

Section 2. Plastic Material Properties | 147

148 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Figure 4.2.2 Implicit backward

Euler method
B
σ e

X C

yield criterion

The Implicit backward Euler method can be expressed using the following equation:

σC  σ B  DmC (4.2.25)

Because the unknown C values exist on both sides of equation (4.2.25), the concept of residual vectors r is
introduced to find the value using repeated analysis:

r  σc  σB  DmC  (4.2.26)

The residual vector r converges to 0 when the final stress state lies on the failure surface. The new residual
vector rnew for recursive calculations using the 1st order Taylor expansion can be defined using the following
equation:

m
rnew  rold  σ   Dm   D σ (4.2.27)
σ

Because the residual vector is rnew  0 for the converged final stress, substituting this into equation (4.2.27) and
rearranging for σ gives equation (4.2.28).

1
 m 
σ    I   D


σ 
r
old  
  Dm  R 1 rold   Dm  (4.2.28)

Also, using the 1st order Taylor expansion on the yield function gives the following equation.
 ANALYSIS REFERENCE Chapter 4. Materials

f f
f new  f old  σ   f old  nT σ  h  0 (4.2.29)
σ 

Substituting equation (4.2.29) into equation (4.2.28) and rearranging for  gives the following equation:

fold  nT R 1rold
 (4.2.30)
h  nT R 1Dm

► Cutting Plane Method

Figure 4.2.3 Cutting Plane Method
B
σ e

X D

yield criterion

The cutting plane method can be defined as follows with reference to Figure 4.2.3:

σ C  σ X  σ e   Dm (4.2.31)

Defining the stress return direction above at point B in the perpendicular direction modifies equation (4.2.31) as
follows:

σC  σ B  DmB (4.2.32)

Also, using the 1st order Taylor expansion on the incremental function gives the following equation.

f f
fC  f B  σ    f B   n B Dm   h  0 (4.2.33)
σ 

Hence, the plastic multiplier increment   is as follows   :

fB
  (4.2.34)
n B Dm B  h

Constitutive equation

Section 2. Plastic Material Properties | 149

150 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

The plastic constitutive equation can be composed as follows. The small stress increment is determined by the
elastic part of the strain increment vector.

σ  D  ε  ε p   Dε  Dm (4.2.35)

Because the current stress always needs to be positioned on the failure surface, the consistency condition
 f  0 needs to be satisfied. Rearranging equation (4.2.35) for the small strain increment gives the following
equation (4.2.36) for the small stress increment:

 DmnT D 
σ  D ε  D ε
ep
(4.2.36)
 h  nT Dm 

The Dep in equation (4.2.36) is called the continuum tangent stiffness matrix,

When using the consistent tangent stiffness matrix for the Newton-Raphson recursive formula, it converges
faster than when equation (4.2.36) is used because of the 2nd order convergence property. This 2nd order
convergence property can be obtained from the following process. First, differentiating equation (4.2.25) gives
the following equation:

m m 
σ  Dε   Dm   D σ   D  (4.2.37)
σ  

Here,  is the change in   .

Equation (4.2.37) can be rearranged as follows:

Aσ  Dε   Dm (4.2.38)

m m 
Here, A  I   D , m  m   D
σ  

If H  A1D , equation (4.2.38) can be arranged as follows.


σ  H ε  m  (4.2.39)

If equation (4.2.39) is rearranged for the total strain term using the consistency condition, the following equation
can be obtained:
 ANALYSIS REFERENCE Chapter 4. Materials

 HmnT H 
σ  H  ε  C ε
ep
(4.2.40)
 h  nT Hm 

The Dep in equation (4.2.36) is the continuum tangent stiffness matrix, and the C ep in equation (4.2.40) is the
consistent tangent stiffness matrix.

The von Mises failure condition assumes that failure occurs when the 2nd order invariant of deviatoric stress
2.3 reaches a certain value. This condition is often used to simulate plastic behavior of metallic materials. The
J2

von-Mises perfect plastic failure condition that does not consider hardening can be expressed using the following equation:

3
f (σ)  3J 2   y  σ dev : σ dev   y  0 (4.2.41)
2
σ dev : Deviatoric stress
y : Failure stress

Because only the deviatoric stress is used to express the failure condition, it is appropriate in expressing the
ductile materials where failure occurs regardless of hydrostatic pressure. The radius of the von Mises failure
surface in 3D stress state is 2 / 3 y , and surface is expressed as a cylinder parallel to the hydrostatic axis.

Figure 4.2.4 von Mises failure

surface in principal stress
coordinate system
2

1   2   3

1
3

Section 2. Plastic Material Properties | 151

152 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Figure 4.2.5 von Mises failure

surface shape in the meridian

deviatoric axis

plane for   
6

rc 
 
6

hydrostatic axis

rt 

6

The associated plastic flow is assumed for the von Mises failure condition. The plastic strain variation is as
follows:

f 3
dε pl  d   d σ dev (4.2.42)
σ 2σ dev : σ dev

Hardening factor
FEA NX supports isotropic, kinematic and combined hardening model in von-Mises yield function. In the case of
isotropic hardening, the central axis of initial yield surface isn’t change since it supposes that the initial yield
surface expands uniformly.

3
f (σ, q)  3 J 2   y (q)  σ dev : σ dev   y (q)  0
2 (4.2.43)

Figure 4.2.5-1 Change of the yield 2

surface of isotropic hardening
model
Isotropic hardening
Initial yield surface

· 1
 ANALYSIS REFERENCE Chapter 4. Materials

 
The hardening factor of isotropic hardening model consists of effective plastic strain such as q  e p   .
The yield stress due to hardening is given by the function of effective plastic strain,  y (e p ) and directly uses
the hardening function, hy (e p ) .
The combined hardening model supposes that expansion and movement of the yield surface occurs
simultaneously by plastic deformation. In the combined hardening model, the yield surface is defined by yield
stress and back stress as follows:

3
f (σ, q)  Σ dev : Σ dev   y (q)  0 (4.2.44)
2
Σdev : σ dev  α

α : back stress

Figure 4.2.5-2 Change of the yield 2

surface of combined hardening
model
Combined hardening

Initial yield surface

· · 1

Kinematic hardening

The hardening factors of combined hardening model are effective plastic strain and back stress.

e p 
 
q  (4.2.45)
 α
  

The yield stress is calculated from hardening function using the combined variable c as follows:

Section 2. Plastic Material Properties | 153

154 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

 y  c hy (0)  (1  c )hy (e p ) (4.2.46)

In case of the combined variable, c  0 , it is isotropic hardening and kinematic hardening when c  1 . The
plastic strain of combined hardening and the change rate of back stress which follows hardening rule of Ziegler
can be expressed using the following equation:

3
dε pl  d  Σ dev (4.2.47)
2 Σ dev : Σ dev
dhy
dα  c dε pl (4.2.48)
de p

Hardening curve
The hardening curve is a material property which expresses plastic property of material. It is generally obtained
from test and uniaxial tension/compression test or pure shear test is widely used. The hardening curve in FEA
NX consists of inputting true stress-plastic strain curve and the conversion process from test result is as follows:

If you know load-displacement curve, true strain and true stress can be calculated using the following equation.

 L0  d   L  Pe
  log    log  L  ,   A
(4.2.49)
 L0   0  0

L0 , L : Length of before/after deformation

A0 : Area of before deformation

If you know engineering stress-strain, it can be calculated as follows:

  log 1   E  ,    E e
(4.2.50)
 E , E : Engineering strain/stress

Since the plastic strain begins to occur from the moment that the material yield, it can be calculated ad follows:


e p     el   
E (4.2.51)
E : Elastic modulus
ANALYSIS REFERENCE Chapter 4. Materials

The Tresca criterion was originally developed to be used on failure conditions of metallic materials. In
2.4 geotechnical analysis it is often used to simulate the ground material behavior for undrained conditions. The
Tresca failure condition for this criterion can be expressed using the uniaxial compression strength, as shown below.

 3  1   y (4.2.52)
y : uniaxial compression strength

Equation (4.2.52) can be expressed using the stress invariant term J2 and 0 , as shown in equation (4.2.53).
(0  0  60 )

1   2 
1   3  J 2 cos 0  cos 0      y (4.2.53)
3   3 

Rearranging this equation:

 1 
f  J 2 , 0   2 J 2 sin 0      y  0 (4.2.54)
 3 

  
Or, it can be expressed using the terms I1 , J 2 ,  as follows.      
 6 6

2   2   4 
f  J 2 , 0   J 2 sin       sin         y
3   3   3  (4.2.55)
 2 J 2 cos    y  0

The effects of hydrostatic pressure on the failure plane are not considered for this criterion and so, it is unrelated
to I1 . The Tresca failure criterion is a hexagonal column parallel to the hydrostatic axis in the principal stress
space, as shown in Figure 4.2.6, and is expressed as a regular hexagon in the deviatoric plane, as shown in Figure
4.2.7(a).

According to the experimental results, the shear strength of the saturated soil is unrelated to I1 for undrained
loading. The Tresca model can obtain appropriate results under these conditions:

Section 2. Plastic Material Properties | 155

156 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Figure 4.2.6 Tresca failure surface

shape in principal stress space
2

1   2   3

1
3

Figure 4.2.7 Tresca failure surface

1
shape in  plane and meridian

deviatoric axis
plane

rc 
 
 rc 6
r
rt

hydrostatic axis

rt 

2 3 6


(a) Failure surface shape in  plane (b) Failure surface shape in the meridian plane for   
6

If the von Mises and Tresca criteria are congruent for rc (0  0 ) and rt (0  60 ) , the von Mises surface
becomes a circle that circumscribes the Tresca hexagon (Figure 4.2.7(a)) in the deviatoric plane. In this case, the
expected maximum difference in failure stress occurs along (0  30 ) , and the failing shear stress ratio
between the von Mises and Tresca criteria is 2 / 3  1.15 . If the two criteria are conformed for simple shear,
 ANALYSIS REFERENCE Chapter 4. Materials

the von Mises circle inscribes the Tresca hexagon, and the maximum error between the two criteria occur along
( 0  0 ) and (0  60 ) .

Figure 4.2.8 von Mises and Tresca

1 1
failure surface shape

Von mises Von mises

Tresca Tresca

 
rc
r r
rt

2 3 2 3

(a) Relation with Tresca-Circumscription in  plane (b) Relation with Tresca-Inscription in  plane

The Mohr-Coulomb model is used to simulate most terrain and it displays sufficiently reliable results for general
2.5 nonlinear analysis of the ground.
Mohr-Coulomb

Figure 4.2.9 Yield function of

Mohr-Coulomb model 

Mohr-Coulomb
  : constant 
Real soil

n

FEA NX can simulate changes in the modulus of elasticity and cohesion with height for a Mohr-Coulomb model
using equation (4.1.4). If the amount of cohesion change with height is ‘0(zero)’, a constant value is used. If the
amount of change is not ‘0(zero)’, the cohesion can be calculated with respect to a reference height using
equation (4.2.56).

c  cref   yref  y  cinc y  y 

ref
(4.2.56)
c  cref y  y 
ref

Section 2. Plastic Material Properties | 157

158 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

cref : Input cohesion value

cinc : Cohesion increment with respect to depth
yref : Depth at which cref is measured

The y in equation (4.2.56) represents the integral position of the element. If the integral position is located
higher than yref , the cohesion can be smaller than '0'. To prevent this, the cohesion value is not decreased any
further and the cref value is used.

Yield function of Mohr-Coulomb model

According to Mohr(1900), failure can be expressed using the following equation:

  c   n tan  (4.2.57)
c : Cohesion

 : Internal friction angle

Here, the limit shear stress  of an arbitrary plane is only related to the normal stress  n of the same plane.
Equation (4.2.57) shows that material failure occurs at the stress state where the largest Mohr circle comes
across the Coulomb friction failure envelope. It also shows that the intermediate principal stress
 2 (1   2   3 ) does not have an effect on the failure condition.

Hence, the yield function of the Mohr-Coulomb failure plane is as follows:

f     n tan   c  0 (4.2.58)

The failure criterion of equation (4.2.58) is called the Mohr-Coulomb criterion and it is the most widely used
method for ground materials due to its simplicity and accuracy.

Expressing the Mohr-Coulomb criterion using principal stress terms (1   2   3 ) , equation (4.2.58) can be
rearranged into the following equation:

1   3 1   3
 sin   c cos 
2 2
1  sin  1  sin 
1 3 1 (4.2.59)
2c cos  2c cos 
1 3
 1
ft f c
 ANALYSIS REFERENCE Chapter 4. Materials

2c cos 
fc  : Uniaxial compressive strength when maximum principal stress is 0
1  sin 
2c cos 
f t  : Uniaxial tensile strength when minimum principal stress is 0
1  sin 

Equation (4.2.59) provides convenience when defining material properties because it uses the uniaxial
compressive and tensile strengths.

Figure 4.2.10 Geometric diagram


of principal stresses
yield envelope

c cos  


1   3
2
c

3 1 
1   3
 sin 
2

1   3

2

Equation (4.2.58) can be expressed using terms I1 , J 2 and  , which are often used in numerical analysis.

1  1 
f  I1 , J 2 ,     I1 sin   J 2  cos   sin  sin    c cos   0 (4.2.60)
3  3 

Assuming associated flow for the plastic potential function gives the following equation:

1  1 
g  I1 , J 2 ,     I1 sin  J 2  cos   sin  sin   c cos  0 (4.2.61)
3  3 

The Mohr-Coulomb criterion is an irregular hexagonal pyramid with a straight meridian in the principal stress
space, as shown in Figure 4.2.11, and the deviatoric shape in the  plane (1   2   3  0) is an irregular
hexagon. To draw the irregular hexagon, the lengths rt 0 and rc 0 are required and can be expressed as follows:

Section 2. Plastic Material Properties | 159

160 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

2 6c cos 
rt 0  (4.2.62)
3  sin 
2 6c cos 
rc 0  (4.2.63)
3  sin 

The rt 0 / rc 0 from equations (4.2.62) and (4.2.63) is as follows:

rt 0 3  sin 
 (4.2.64)
rc 0 3  sin 

Because the deviatoric sections of the Mohr-Coulomb failure surface are all geometrically similar, the ratio
rt / rc is always constant for an arbitrary deviatoric section.

rt rt 0 3  sin 
  (4.2.65)
rc rc 0 3  sin 

If the tensile strength is input, the tensile principal stress of the Mohr-Coulomb cannot surpass the input tensile
strength. FEA NX applies a complex of the Mohr-Coulomb failure function and the tensile Rankine failure
function to consider Mohr-Coulomb failure with allowed tensile strength.

In the Mohr-Coulomb model tensile strength can be considered based on two types: Pressure and Rankine.

- In the first “pressure type” method, the average of the principal stresses can not exceed the tensile strength:
1   2   3
 t
3
- For Rankine type the maximum principal stress should not exceed the tensile strength.

1   t

For more information on the Rankine model, see Section 2.14.

 ANALYSIS REFERENCE Chapter 4. Materials

Figure 4.2.11 Mohr-Coulomb

failure surface shape in principal
stress space  1

1   2   3

 2
 3

Section 2. Plastic Material Properties | 161

162 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Figure 4.2.12 Mohr-Coulomb

1

deviatoric axis
failure surface shape in  plane
and meridian plane


 
6 2 6 c cos 
 rt 0 
rt 0 3  sin 
r
rc 0 3c cot 
hydrostatic axis

2 6 c cos 
rc 0 
3  sin 
2 3 

6


(a) Failure surface shape in  plane (b) Failure surface shape in the meridian plane for   
6

As shown in Figure 4.2.12(b), tan  corresponding to the slope of the straight failure surface of the Mohr-
Coulomb failure criterion does not change with the confining pressure (or hydrostatic pressure). Hence, the
criterion is accurate when the confining stress is within a limited range, but it does not agree with actual physical
phenomena when the confining stress is larg enough to cause compressive failure. However, this criterion gives
highly accurate results within the confining stress ranges of the field and it is easy to use. Hence, it is the most
widely used failure model.

The Drucker-Prager model3 was developed to solve the numerical problems that occur on the corners of the
2.6 yield shape of the Mohr-Coulomb model. This model is an expansion of the von Mises model and because the
Drucker-Prager function is defined such that the deviatoric stress can increase or decrease depending on hydrostatic pressure,
it is also called the Extended von Mises criterion.

Yield function of Drucker-Prager model

Thye Drucker-Prager failure criterion ( f ) and plastic potential function( g ) can be expressed using the stress
invariant terms I1 and J2 as follows:

f  I1 , J 2   3J 2   I1   c  0
(4.2.66)
g  I1 , J 2   3J 2   I1  0
Here, ,  ,  are as follows:

3 Drucker, D. C. and Prager, W. “Soil mechanics and plastic analysis for limit design,” Quarterly of Applied Mathematics,
vol. 10, no. 2, 1952, pp. 157–165.
 ANALYSIS REFERENCE Chapter 4. Materials

2sin  6cos  2sin

  
3  sin  , 3  sin  , 3  sin (4.2.58)

The Drucker-Prager failure surface can be expressed in the principal stress space, as shown in Figure 4.2.13. This
failure surface has a conical shape with the hydrostatic axis ( 1   2   3 ) as its center. The Drucker-Prager
failure surface can be thought of as a Mohr-Coulomb failure surface with no edges, or it can be thought of as the
expanded form of the von Mises failure surface for materials that depend on hydrostatic pressure, such as soil.
If it is assumed to circumscribe the outer boundary of the Mohr-Coulomb failure surface, then  and  can be
expressed as follows.

tan  3
 ,  (4.2.68)
9  12 tan   9  12 tan 2  
1/2 1/2
2

Figure 4.2.13 Drucker-Prager

failure surface shape in principal
stress space  1

1   2   3

 2
 3

Section 2. Plastic Material Properties | 163

164 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Figure 4.2.14 Drucker-Prager 1

failure surface shape in  plane

deviatoric axis
and meridian plane

 
6
6
1 r0  2  c
 r0 c
r 3
hydrostatic axis

1 r0  2  c
6


6
2 3

(a) Failure surface shape in  plane (b) Failure surface shape in meridian plane

FEA NX provides the strain softening model with the stress-strain curve shown in Figure 4.2.15. This stress-strain
2.7 curve is composed of 3 linear sections. The linear sections are the elastic section to peak shear strength, the
Strain Softening strain softening section from peak to residual shear strength, and the constant residual shear strength section.

Figure 4.2.15 Strain softening

composition relationship
shear strength

peak, cu

elastic strain residual strain

softening

Failure criterion
The strain softening model of FEA NX is an elastic-soft plastic model that uses the von Mises model. The
softening behavior is isotropic softening behavior and is formulated based on the strain softening theory. The
yield function of the strain softening model can be expressed using the shear stress and shear strength terms,
as shown in equation (4.2.69).

f  3J 2  3Cu   (4.2.69)
ANALYSIS REFERENCE Chapter 4. Materials

Here, the shear strength Cu can be expressed using the softening coefficient  as shown in Figure 4.2.15
using equation (4.2.70):

 Cu when   0

Cu  Cu  R when 0<   res (4.2.70)
 C when    res
 res
Cu : Maximum cohesive shear strength
Cres : Residual cohesive shear strength
 : Softening coefficient
 res : Softening coefficient at intersection of residual strength line and softening line
R : Softening rate

The softening coefficient  is a control variable that controls the plastic softening behavior and is calculated
from the principal plastic strain. The principal plastic strain of the von Mises model is as follows :

2 1   2   3 
1  
ε  m  
p
 1  2 2   3  (4.2.71)
2 e  
 1   2  2 3 

Here,

 xx   2 1 1 0 0 0
   1 2 1
 yy   0 0 0 
   1 1 2 0 0 0
1 T  zz 
e  σ Pσ , σ , P  (4.2.72)
2  xy  0 0 0 6 0 0
  0 0 0 0 6 0
 yz   

 zx 
  0 0 0 0 0 6 

The softening coefficient  can be calculated from the following equation and the  - Cu relationship is
shown in Figure 4.2.16.


3
ε  Q εp
2 p T
(4.2.73)

Section 2. Plastic Material Properties | 165

166 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Figure 4.2.16 Definition of

shear strength
softening behavior

Cu
R

Cres

 res 

Simulating clay like materials as elastic-plastic hardening materials is a widely used concept (Britto and Gunn4).
2.8 The Modified Cam-clay model provided in FEA NX is also based on the elastic-plastic hardening theory.
Modified Cam-Clay
Formulation of the Modified Cam-clay model in FEA NX uses all effective stresses and is materialized using
nonlinear elastic and the implicit backward Euler method (R.I. Borja5). Nonlinear elastic behavior represents the
increase in bulk modulus when pressure is applied to the material. Also, the associated flow rule is used and the
failure surface can increase or decrease depending on hardening/softening behavior.
Figure 4.2.17(a) displays the relationship between the ground volume change and hydrostatic pressure using the
normal consolidation line and over-consolidation line, or swelling line. If the stress increases and surpasses the
hydrostatic pressure, the volume change follows the over-consolidation line. If the increase in hydrostatic
pressure is enough, the volume change passes through the intersection of the normal and over-consolidated
lines and follows the normal consolidation line.

Rotating Figure 4.2.17(a) in the counter clockwise direction by 90°shows similarities with the elastic-plastic
hardening stress-strain curve in Figure 4.2.17(b). In other words, the overconsolidation line corresponds to the
initial linear elastic section and the normal consolidation line corresponds to the hardening plastic stress-strain
relationship.

4 Britto, A. M., Gunn, M. J. Critical state soil mechanics via finite elements, Ellishorwood Limited, 1987.
5 Borja, R. I., “Cam-clay plasticity, Part II: Implicit integration of constitutive equation based on a nonlinear elastic stress
predictor,” Computer Methods in Applied Mechanics and Engineering, Vol. 88, Issue 2, 1991, pp. 225-240.
 ANALYSIS REFERENCE Chapter 4. Materials

Figure 4.2.17 Similarity between

volume-hydrostatic pressure and
stress-strain relationships normal consolidation

volume change

stress
hardening plastic

overconsolidation linear elastic

pressure strain
(a) Volume-hydrostatic pressure relationship (b) Stress-strain relationship

Figure 4.2.18 displays the pressure, volume and critical state line relationship. M is defined as the slope of the
critical state line in Figure 4.2.18(a) projected onto the p  q plane, as shown in Figure 4.2.18(b).

Figure 4.2.18 Critical state line

V q

isotropic normal critical state line

consolidation line

overconsolidation line
 M
critical state line
k

ln P P
ln(1)
(a) Pressure and specific volume relationship (b) Critical state line

Table 4.2.2 Modified Cam-clay Symbol Significance

material properties
 Slope of over-consolidation line
 Slope of normally consolidation line
M Slope of critical state line

The material properties of the ground are generally obtained from 1D consolidation experiments. The
compression index Cc and recompression index Cs are generally obtained from the void ratio, e and
log10  p  graph. The compression index and recompression index have the following equation using the slope
of normal consolidation line  and slope of over-consolidation line  :

Section 2. Plastic Material Properties | 167

168 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Cc Cs
 ,  (4.2.74)
2.303 2.303

The slope of the critical state line M can be estimated from the relationship with the effective shear resistance
angle (shear resistance angle from drained tests).

6sin 
M
3  sin  (4.2.75)
 : Internal friction angle, calculated from triaxial compression test

 can be calculated using the following equation, after the specific volume N of the normal consolidation
line at p  1.0 is found from Figure 4.2.18(a).

  N       ln 2 (4.2.76)

The yield function of Modified Cam-clay is as follows, and it displays an elliptic shape:

q2
f (σ)   p( p  pc )  0
M2
1
p '    'kk  ij
3 (4.2.77)
3
q s:s
2
1
sij   'ij   'kk  ij
3
pc : Pre-consolidation pressure
M : Slope of critical state line
 'ij : Effective stress
 ANALYSIS REFERENCE Chapter 4. Materials

Figure 4.2.19 Yield function of q

Modified Cam-clay model
Critical state line

Yield function

P
PC

When the ground reaches the critical state, the following relationship is satisfied:

q  Mp (4.2.78)

The size of the Modified Cam-clay model failure surface is determined by pc . In other words, increasing pc
increases the failure surface and can simulate hardening behavior, and reducing pc can simulate softening
behavior. The hardening/softening equation for Modified Cam-clay models can be obtained from the following
process:

Firstly, the volumetric strain change and its relationship with the specific volume change are defined by the
following equation:

dV dv
d v   (4.2.79)
V 1 e
e : Void ratio
v : Specific volume
v : Volumetric strain

Also, the additive decomposition of strain is assumed for the Modified Cam-clay model as shown below:

d v  d ve  d vp (4.2.80)

The following equation can be obtained from equations (4.2.79) and (4.2.80).

Section 2. Plastic Material Properties | 169

170 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

dv  1  e  d  v  1  e  (d  ve  d  vp )  dv e  dv p
(4.2.81)
dv p  dv  dv e

The v and pc relationship can be rearranged using Figure 4.2.20.

Figure 4.2.20 Pressure and

specific volume relationship
ln p
Normal consolidation line

 ln  pc  

 

v
v p ve
v

dpc '
dve   d ln  pc '  
pc '
(4.2.82)
dpc
d   d ln  pc   
pc

Rearranging equations (4.2.81) and (4.2.82) are as follows:

dpc
d p  d  d e  (   )
pc
(4.2.83)
dpc 1 e
 d p
pc   v

Integrating equation (4.2.83) gives the following hardening/softening equation:

1 e
pc  pc 0 exp(  vp ) (4.2.84)
 

Also, Modified Cam-clay material models have the following nonlinear elastic properties:
ANALYSIS REFERENCE Chapter 4. Materials

1 e 3(1  2v)
K p', G  K (4.2.85)
 2(1  v)
K : Bulk modulus
G : Shear modulus
e : Void ratio

However, the effective pressure is unknown when calculating the initial stress state and so, the given linear
elastic modulus is used.

To use the Modified Cam-clay model, the initial void ratio, in-situ stress and initial pre-consolidation pressure
pc is required. FEA NX uses a direct input value or a value automatically calculated from in-situ stresses and
over-consolidation ratio (OCR) for the pre-consolidation pressure pc . The user needs to input the initial void
ratio.

The over-consolidation ratio (OCR) is defined using equation (4.2.86). The p 'max is the maximum effective
hydrostatic pressure on the material, and p 'v is the initial effective hydrostatic pressure. Generally, the
maximum effective normal stress experienced by the ground is determined from oedometer tests.

p 'max
OCR  (4.2.86)
p'

For a clear explanation, it is assumed that the shear stress does not exist and the gravitational direction is the y
axis. Then, the in-situ stress becomes equation (4.2.87).

σ0   X 0  Y 0  Z 0 0 0
T
0 (4.2.87)

To calculate pc , first use the OCR and equations (4.2.88) and (4.2.89) to calculate pmax and qmax .

σmax  OCR X 0 0 0 0
T
OCR Y 0 OCR Z 0 (4.2.88-a)
σmax  OCR Y 0 K0 0 0 0
T
OCR Y 0 OCR Y 0 K0 (4.2.88-b)


pmax
1

  x max   y max   z max
3

 x max   y max    y max   z max 
1 2 2
  z max   x max 
2

qmax 
2 (4.2.89)

Section 2. Plastic Material Properties | 171

172 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

The K0 condition is also applied to equation (4.2.88-b). K0 can be estimated from the internal friction angle
using the following equation:


K0  h  1  sin   (4.2.90)
 v
 h : Horizontal direction effective stress

FEA NX uses equation (4.2.77) to calculate pc . If the user inputs the pc value directly, FEA NX tests whether the
input value and in-situ stress state satisfy equation (4.2.77) and adjusts the value when it is not satisfactory.

General ground surface strata have brittle fracture surfaces, and these are called 'joints' for rock models. The
2.9 material models that reflect these properties are called jointed rock models.
Jointed Rock
Jointed rock models are transversely isotropic perfectly plastic material models. The material can have
transversely isotropic properties depending on the rock layer characteristics in the elastic region. In other words,
rock layers have isotropic properties in the layer direction, but have anisotropic properties in the normal
direction to the layer. The perfectly plastic behavior is based on the Coulomb friction function in the major joint
direction. Hence, perfect plasticity occurs in the major joint direction when maximum shear stress is reached.
The major joint direction can be defined in a maximum of 3 directions and the first major joint direction is equal
to the transversely isotropic material direction.

Orthotropic elastic material stiffness

The elastic material behavior of jointed rock models are already explained for transversely isotropic elastic
materials above.

Plastic behavior in 3 directions

The yield function in the major joint direction i can be defined using equation (4.2.82):

fi   s2   t2   n tan i  ci (4.2.91)
 ANALYSIS REFERENCE Chapter 4. Materials

Figure 4.2.21 Yield criterion for τ

individual planes fi  τ   n tan i  ci

tan i
ci

n

To examine the plasticity condition for each failure surface, stress transformation to the local coordinate system
( n , s , t ) is required.

σ nst  TiT σ XYZ (4.2.92)

σ nst : Stress in local coordinate system
σ XYZ : Stress in material coordinate system
Ti : Transformation matrix in the i active plane

The general 3D transformation matrix that considers the dip angle and dip direction is as follows:

 nx2 n y2 nz2 2 nx n y 2n y nz 2nz nx 

 
Ti   nx sx
T
ny s y nz s z nx s y  n y s x nz s y  n y s z nz s x  nx s z  (4.2.93)
 nx t x nyt y nz t z n y t x  nx t y n y t z  nz t y nz t x  nxt z 


Here, the 3D ni , s i , t i is as follows:

 sin 1  cos 3   cos 1  cos 3   sin 3 

ni   sin 1  sin 3  , si   cos 1  sin 3  , t i  cos 3 
   
(4.2.94)
 cos 1    sin 1   0 

Expressing the local coordinate system by rotation in the GCS is as follows:

Section 2. Plastic Material Properties | 173

174 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Figure 4.2.22 MCS from dip angle Z

and dip direction Z

declination t
N n

2
3 X 1 s
1
s
s
s

The Modified Mohr-Coulomb model is the expanded version of the Mohr-Coulomb model, and is a specialized
2.10 model for silt or sand. Modified Mohr-Coulomb models are complex material models which combine nonlinear
Modified Mohr-Coulomb elastic and plastic models.

Nonlinear elastic
The Modified Mohr-Coulomb model defines the elastic region as nonlinear elastic and the power-law is used to
obtain the elastic volumetric stress-strain relationship. In other words, the tangent compression modulus is
expressed as a water supply form of the current hydrostatic pressure, as shown below:

1 m
 p 
K t  K ref 
 p 
(4.2.95)
 ref 
K ref : Reference coefficient of compressibility
pref : Reference pressure

m : Rational number, '0.5' used for sand 0  m  1

The following equation can be expressed by considering the tensile pressure ( pt ):

1 m
 p  pt 
K t  K ref   (4.2.96)
 p
 ref 
 ANALYSIS REFERENCE Chapter 4. Materials

Here, the tensile pressure is a numerical invention used to consider the tensile stress when an in-situ pressure of
'0' is assumed. However, actual soil analysis nearly always considers non-zero in-situ stresses.
The equation (4.2.96) above is derived as the volumetric stress-strain relationship, as shown in the equation
(4.2.97):

m1
 p  pt 
dp  K ref d Ve
 p 
(4.2.97)
 ref 

Integrating equation (4.2.97) and rearranging gives equation (4.2.98):

 p  p  
1

p   pt  m1
 mpref K ref Ve  F  Ve 
m m
0 t (4.2.98)

Yield function
The failure surface of the Modified Mohr-Coulomb model is a decoupled double hardening model, where shear
failure and compressive failure do not affect each other. This coupled failure surface has the following equation
in the p  q space:

q 6sin 
f1    p  p   0
R1   3  sin 
2 (4.2.99)
 q 
f 2   p  p       pc  0
2 2

 R2   
f1 : Shear yield function

f2 : Compressive yield function

Figure 4.2.23 Shape of yield q

function in p-q plane

pc

6sin 
3  sin 
p
p 0 pc  p

Section 2. Plastic Material Properties | 175

176 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Figure 4.2.24 Shape of yield  1

function in deviatoric plane


 Drucker-Parger
 0 6 ( 1  0)

 
6 fit for   15
 fit for   35

 3  2

The functions R1   , R2   of equation (4.2.99) model the strength difference between triaxial compression
and triaxial tension as a function of  . The Modified Mohr-Coulomb model can be expressed in the same way
as a Mohr-Coulomb model using functions R1   , R2   in the deviatoric plane.
A relationship like the one found in equation (4.2.100) is derived to fit the triaxial tensile Mohr-Coulomb model.

n
 1  1 sin 3 
R1      (4.2.100)
 1  1 

Here, n = -0.229 .

1 is coupled with the friction angle, as shown in equation (4.2.101).

1

 3  sin   n
  1
3  sin  
1   1 (4.2.101)

 3  sin   n
  1
 3  sin  

Here, 1  0.7925 .

The maximum value of 1 is the friction angle (  ) 46.55°. Also, the shape of the compression cap can be
modified using R2   . R2   is the same as equation (4.2.102) below.

2
 1   2 sin 3 
R2      (4.2.102)
 1  2 
ANALYSIS REFERENCE Chapter 4. Materials

Here, n = -0.229 and the cap is circular when the basic value ‘0’ is used for 2 .

Flow rule
The plastic potential function in the p  q plane can be expressed as follows. For the Modified Mohr-Coulomb
model, it is applied to 2 faces that consider shear and compression.

6sin
q1  q   p  p 
3  sin (4.2.103)
q2   p  p    q 2  pc2
2

Here, the dilatancy angle  can be expressed using the friction angle  , as shown in equation (4.2.104).

sin   sin cv

sin  (4.2.104)
1  sin  sin cv

sin cv : Friction angle when volume is constant

Hardening behavior
Two types of shear and compression hardening behavior are applied in the Modified Mohr-Coulomb model.
The shear hardening behavior is determined by the friction angle and can be expressed as the following
equation. If the shear hardening occurs, the dilatancy angle is recalculated by the Row’s rule6 (4.2.104).

sin   sin    (4.2.105)

2
 γp  γp (4.2.106)
3

 : Equivalent deviatoric plastic strain

γp : Deviatoric plastic strain

The compression hardening behavior is expressed by the pre-consolidation stress as the following equation.

1
  P m m m
Pc  pref   c 0    vp  (4.2.107)
  pref   
 

Pc : Pre-consolidation stress

6 ROWE, P. W. The stress-dilatancy relation for static equilibrium of an assembly of particles in contact.
Proc. Roy. Soc. London A269 (1962), 500-527.

Section 2. Plastic Material Properties | 177

178 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Pc 0 : Pre-overburden pressure

 : Compression cap hardening parameter

Hoek and Brown 7 suggested the use of the equivalent continuum concept to define the stress reduction
2.11 phenomena due to failure of jointed rocks. Hoek and Brown first suggested a failure function to separate intact
Hoek-Brown rock and broken rock. After the failure is defined, the stress reduction phenomena were simulated by decreasing
the particular coefficient values that define the failure function. This method suggested by Hoek and Brown
defines the uniaxial compressive strength, which cannot be considered in the existing Mohr-Coulomb method.
This allows for the accurate and simple representation of rock behavior.

Yield function
The failure criterion suggested by Hoek and Brown is as follows. The intermediate principal stress (  2 ) term is
ignored in this failure criterion.

 1   3  m c 3  s c2
(4.2.108)
1   2   3
c : Uniaxial compressive strength
m,s : Empirical coefficient for defining rock failure

The yield function ( f ) can be expressed using the stress invariant as follows:

 sin   I
f  4 J 2 cos2   m c 2    cos    m c 1  s c2  0 (4.2.109)
 3  3
I1 : First order invariant
J2 : Second order invariant
c : Uniaxial compressive strength of the rock (  / 6     / 6 )

In the principal stress space, the Hoek-Brown model has a diverging hexagonal pyramid shape along the
hydrostatic axis and its deviatoric plane shape is expressed as an angular hexagonal shape made up of 6 curved
surfaces. This hexagonal shape has an edge where the curved surfaces meet, and this creates difficulties. To
solve this problem, FEA NX processes these edges as curved surfaces using the modified Hoek-Brown criterion
suggested by Wan8.

7 Hoek, E. and Brown, E. T., “Empirical strength criterion for rock masses,” Journal of Geotechnical and
Geoenvironmental Engineering, Vol. 106, Issue GT9, 1980.
8 Wan, R. “Stress return solution algorithm for generalized Hoek-Brown plasticity model,” Proceedings of the 8th
 ANALYSIS REFERENCE Chapter 4. Materials

f  q 2 g 2 ( )   c*qg ( )  3 c* p  s c2  0 (4.2.110)

Here,  c*  m c / 3 , q  3J 2 , p  I1 / 3 and the g ( ) used to define the deviatoric plane shape is as follows:

4(1  e2 ) cos 2 ( / 6   )  (1  2e) 2

g ( )  (4.2.111)
2(1  e2 ) cos 2 ( / 6   )  (2e  1) D

Here, D 4(1  e2 ) cos2 ( / 6   )  5e2  4e

Figure 4.2.25 displays the shape of the Hoek-Brown model in the stress space.

Figure 4.2.25 Hoek-Brown failure

surface  1

1   2   3

 2
 3

The Coulomb friction model assumes that the frictional force is proportional to the size of the value obtained by
2.12 multiplying the coefficient of friction and the tangent direction force. FEA NX defines the yield function of the
Coulomb Friction model using the equation below:

f  tt2  tn tan  ( )  c( )  0 (4.2.112)

tt : Lateral direction force
tn : Normal direction force
 : Internal friction angle

International Conference of the Association for Computer Methods and Advances in Geomechanics, Morgantown, USA,
1994, pp. 719-724.

Section 2. Plastic Material Properties | 179

180 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

c : Cohesion
 : Size of plastic relative displacement

In FEA NX, the internal friction angle and cohesion can be set as a functional value that depends on the plastic
relative displacement.

The equation above can be expressed as Figure 4.2.26. FEA NX supports additional tensile strength inputs to
express brittle behavior in the tensile direction.

Figure 4.2.26 Coulomb friction tt

function
c / tan 

tensile strength

c

tn

The plastic relative displacement u p can be defined using the plastic multiplier, which represents size and
plastic direction components as shown below:
g
u p  
t
(4.2.113)
g  t t2  tn tan
 : Dilatency angle
 : Plastic multiplier

Here, the plastic multiplier can be calculated from the map regression method.
The Janssen model, which is applied to the rotational DOF of shell interface elements, simulates the nonlinear
2.13 elastic relationship between the moment and rotational displacement. FEA NX provides for the Coulomb
Janssen friction model and Janssen model for shell interface elements. The Coulomb friction model is used to define the
normal and lateral direction forces.
 ANALYSIS REFERENCE Chapter 4. Materials


 0 & tx  t y  tz  0 if u x  0



 K nb3 2t x
my    y if  y  , t x  K x bu x (4.2.114)
 12 K nb 2
  
  y t x b 1  8 tx
 if  y 
2t x
  2  9   2
 K nb2
  y K nb 
 y
t x / ux : Normal interface traction / Normal relative displacement
t y , tz : Tangential interface traction
m y ,  y : Axial moment / rotation angle
b : Thickness of shell interface element
Kn : Tangential stiffness

Only perfectly plastic behavior is supported for Coulomb friction models used on shell interface elements.

Geogrids are reinforcing structures used to strengthen the ground and have only the tensile only structural
2.14 behavioral properties. When selecting the geogrid element in FEA NX, the inverse Rankine model applied on a
Rankine/Inverse Rankine truss elements is used for 2D models and the inverse Rankine model applied on a plane stress element is used
for 3D models. Here, the allowable compressive strength is '0(zero)'. The inverse Rankine model only needs to
be computed using the opposite sign from the Rankine model defined below.

The Rankine material failure assumes that failure occurs when the maximum principal stress (  max ) reaches the
tensile strength and the yield function is as follows:

f   max   t    0 (4.2.115)

Expressing equation (4.2.115) using the invariants I1, J 2 ,  are as follows.

2 I
f  J 2 A    1 (4.2.116)
3 3

Here, 𝐴(𝜃) is

Section 2. Plastic Material Properties | 181

182 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

 3 1
 cos   sin  for  1
 2 2
A    sin  for  2 (4.2.117)

  3 cos   1 sin  for  3
 2 2

Figure 4.2.27 displays the 3D shape of the inverse Rankine model in the stress space. The shape is a right triangle
in the deviatoric plane (  plane) as shown in Figure 4.2.28, and it can be defined as a linear function about the
hydrostatic axis in the meridian plane.

Figure 4.2.27 Rankine failure

surface shape in principal stress  1
space

 2

 3

Figure 4.2.28 Rankine failure

surface shape in deviatoric plane
deviatoric axis

(  plane) and meridian plane 1


 
6
3
1 rt  f t
2
2
3 f t
r  rt
hydrostatic axis
rc

1 rc  6 f t
2

2 3 
6


(a) Failure surface shape in deviatoric plane (  plane) (b) Failure surface shape in meridian plane for   
6
ANALYSIS REFERENCE Chapter 4. Materials

The Modified UBCSAND model is developed to simulate liquefaction phenomenon using plastic theory based
2.15 on effective stress. It is extended to enable implicit nonlinear analysis for 3D stress state based on the
Modified UBCSAND constitutive model9, 10 developed to simulate liquefaction phenomenon with explicit method for 2D stress state.
Basically, in the elastic domain, it represents a nonlinear elastic behavior that elastic modulus changes with
respect to pressure. A plastic behavior is determined by three yield functions of shear, compression and pressure
off. In particular, the shear yield function is able to consider the effect of material densification for cyclic loading.
The Modified UBCSAND model is implemented as the implicit backward Euler method to maximize
convergence and efficiency and uses a consistent tangent stiffness matrix

Nonlinear elasticity
In elastic zone, it represents a nonlinear elastic properties that elastic modulus changes with respect to effective
pressure ( p ' ).

ne
 p ' pt 
G e  K Ge pref   (4.2.118)
 p
 ref 

K Ge : Elastic shear modulus number

pref : Reference pressure

ne : Elastic shear modulus exponent

pt : Allowable tension pressure

Here, the allowable tension pressure is calculated automatically based on the cohesion and maximum friction
angle. Assuming that Poisson’s ratio doesn’t change according to the pressure and isotropic properties are
maintained, the bulk modulus is determined as follows:

2 1    e
Ke  G (4.2.119)
3(1  2 )

Shear yield function

The Modified UBCSAND model represents a plastic shear behavior using Mohr-Coulomb yield function.

9 Beaty, M. and Byrne, PM., “An effective stress model for predicting liquefaction behaviour of sand,” In Geotechnical
earthquake engineering and soil dynamics III, Americal Society of Civil Engineers, Geotechnical Special Publication 75(1),
1998, pp. 766-777.
10 Puebla, H., Byrne, PM., and Phillips, R., “Analysis of CANLEX liquefaction embankments: protype and centrifuge
models,” Canadian Geotechnical Journal, 34, 1997, pp 641-657.

Section 2. Plastic Material Properties | 183

184 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

1
f s  Rmc 3 J 2  I1 tan m  c  0 (4.2.120)
3

m : Mobilized friction angle

Here, Rmc which expresses the shape of  plane is as follows:

1   1  
Rmc  sin      cos     tan m
3 cos m  3 3  3
(4.2.121)
3 3 J3

2 J 23/2

Shear flow rule

The flow rule utilizes the following plastic potential based on the non-associated plastic flow rule11. Therefore,
the non-associated matrix operation is performed in case of using the Modified UBCSAND model.

I1
gs   c tan m   3Rmw 2 J 2  tan m
2

3
(4.2.122)
4 1  e 2  cos 2    2e  1
2
 
Rmw  Rmc  , m 
2 1  e 2  cos    2e  1 4 1  e 2  cos 2   5e 2  4e 3 

m : Mobilized dilatancy angle

The size of dilatancy angle changes with the similar form to the stress-dilatancy angle theory12 for the variation
of mobilized friction angle

sin m  sin m  sin cv (4.2.123)

cv : Constant volume friction angle

In other words, the plastic deformation describes shrinkage when mobilized friction angle is smaller than
constant volume friction angle whereas it describes swelling when mobilized friction angle is larger than
constant volume friction angle

11 Menetrey, P. and Willam, KJ, “Triaxial failure criterion for concrete and its generalization,” ACI Structural Journal, 92:3,
1995, pp. 11-18.
12 Rowe, P.W., “Stress-dilatancy relation for static equilibrium of an assembly of particles in contact,” Proceedings of the
Royal Society of London, Mathematical and Physical Scieces, Series A, 269, 1962. pp-500-557.
 ANALYSIS REFERENCE Chapter 4. Materials

Figure 4.2.29 Swelling/Shrinkage

according to the direction of
plastic strain

Shear Stress
Constant volume
cv
Dilative

Contractive

Mean Stress

Hardening shear behavior

The hardening rule for the variation of maximum plastic shear strain is represented by the variation of stress
ratio.

  sin m
(4.2.124)
f s m ,    0

The hardening phenomenon is determined respectively for two shear yield functions. The primary yield surface
is used when the present stress ratio is the maximum stress ratio of material. On the other hand, the secondary
yield surface is activated when the present stress ratio is smaller than the maximum value of material. At this
time, if the stress ration exceeds the previous maximum value, the primary yield surface is activated again.
The secondary yield surface is introduced to simulate the increase of plastic stiffness for cyclic loading.
Therefore, the plastic strain by secondary yield surface is smaller than by the primary yield surface. This
phenomenon is called the densification of soil. On the other hand, if the load is unloading, the material maintains
elastic state and the stress ratio decreases.
The hardening rule of primary yield surface is expressed with the following equation:

np 1 2
Gp  p'    sin   
 sin m   s  KGp  1    R f   s
 p 
m
p'  ref    sin  p   (4.2.125)
 s  1p   3p

K Gp : Plastic shear modulus number

pref : Reference pressure

np : Plastic shear modulus exponent

Section 2. Plastic Material Properties | 185

186 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

p : Peak friction angle

Rf : Failure ratio

 1P ,  3P : Min./Max. plastic strain of principal axis

Figure 4.2.30 Plastic shear

hardening behavior Gp / p'

Stress Ratio

 sin m

 S

Maximum Plastic Shear Strain

Considering the densification of soil due to cyclic loading, the hardening rule of secondary yield surface is as
follows:

np 1 2
 p '    sin m  
 sin m  KGp,2  1 R  s
 p    sin   f 
 ref    p   (4.2.126)
p n 1 
KG ,2  KG  4 
p
 Fdens
 2 

K Gp,2 : Cyclic plastic shear modulus number

n : Number of half cycles

Fdens : Soil densification fitting factor

As the mobilized friction angle for cyclic loading closes to the maximum mobilized friction angle, the plastic
shear modulus decreases and finally it closes to the perfect plastic state. In this case, the ground is determined
with liquefaction. In case of reaching liquefaction, some residual hardening stiffness can be given through Fpost .

KGp, post  KGp Fpost (4.2.127)

Fpost : Post liquefaction fitting factor
 ANALYSIS REFERENCE Chapter 4. Materials

Cap yield function, flow rule and hardening behavior

The cap yield function is same with that of Modified Mohr-Coulomb model (4.2.99). It uses the associated plastic
flow rule and the hardening model can be expressed with the following equation in the form what the size of
compression limit increases for the plastic volumetric strain.

mp
 p' 
pc  K Bp pref   vp
 p 
(4.2.128)
 ref 
K Bp : Plastic bulk modulus number

mp : Plastic shear modulus exponent

Pressure cut-off yield function and flow rule

The pressure cut-off yield function can be considered additionally to add the condition what the effective
pressure is always larger than specified value ( f pr  pcut  p ' ). It uses the associated plastic flow rule and
the pressure cut-off yield function does not have hardening behavior.

The Sekiguchi-Ohta model is widely used in Japan and still improved since it is developed by Sekiguchi and
2.16 Ohta13. There are Inviscid and Viscid type. The Inviscid type is plastic model without time dependency. Although
Sekiguchi-Ohta (Inviscid) it shares several characteristic with Cam-Clay14 model, there is a difference that the irreversible diliatancy15 is
strictly described considering the K0 stress state of normally consolidation. However, it causes a numerical
problem because the plastic flow value is only undetermined under the preconsolidation stress state. In FEA NX,
it resolves numerical problem in the preconsolidation stress using the algorithms 16 which calculates the
specificity vertex using the crossing of the two yield function.

Nonlinear elastic
Similarly to the Modified Cam-Clay model of FEA NX, it shows the nonlinear elastic characteristic what the
elastic modulus changes with the effective stress ( p ) in the elastic range.

1  e0
31  2 
K  ,G  K (4.2.129)
2 1   
p


13 Sekiguchi, H. and Ohta, H., "Induced anisotropy and time dependency in clays", 9th ICSMFE, Tokyo, Constitutive
equations of Soils, 1977, 229-238
14 Roscoe, K. H., Schofield, A. N. and Thurairajah, A., "Yielding of Clays in States Wetter than Critical", Geotech., 1963,
Vol. 13, No. 3, pp. 211-240
15 Ohta, H., Sekiguchi, H., "Constitutive equations consdiering anisotropy and stress reorientation in clay", Proceedings of
the 3rd International Conference on Numerical in Geomechanics., 1979, pp. 475-484
16 Pipatpongsa, T., Iizula, A., Kobayashi, I., Ohta, H., "Fem formulation for analysis of soil constitutive model with a corner
on the yield surface", Journal of Structural Engineering, Vol. 48, pp. 185-194

Section 2. Plastic Material Properties | 187

188 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

K : Bulk modulus

e0 : Initial void ratio

 : Slope of overconsolidation line

 : Poisson's ratio

G : Shear modulus

Yield function
The yield function of Sekiguchi-Ohta (Inviscid) model is as follows:

 p
f  MD ln    D  0 (4.2.130)
 pc 

M : Slope of critical state line

D : Dilatancy modulus

pc : Preconsolidation pressure

 : Generalized relative stress ratio

The generalized relative stress ratio (  ) which is the value for the degree of volume expansion is expressed as
follows:

3  sij scij  sij scij 

      (4.2.131)
2 p pc  p pc 

sij : Stress deviator tensor

scij : Preconsolidation stress deviator tensor of K0 state

In the above equation, it can be found that  is affected by the preconsolidation pressure and each component
of stress deviator tensor. Through this, it describes strictly the dilatancy effect than Cam-Clay model which
considers only the preconsolidation pressure and the present stress deviator tensor.

The dilatancy modulus ( D ) has the relation with slope of critical state line ( M ), slope of normally
consolidation line (  ), slope of overconsolidation line (  ) and initial void ratio as follows:
 ANALYSIS REFERENCE Chapter 4. Materials

MD       (4.2.132)

Here,    1  e0  ,     1  e0  . The dilatancy modulus ( D ) is calculated internally in FEA NX.

Figure 4.2.31 displays the yield function in principal stress space. As described before, it can be found that the
yield function of Sekiguchi-Ohta model has the specificity which causes a numerical problem in the
preconsolidation stress.

Figure 4.2.31 Yield function in

principal stress space

 1

 3

 2

Hardening behavior
The preconsolidation stress ( pc ) is used as a parameter of isotropic hardening and defined as follows:

 p   p 
pc  pc 0 exp  v v0  (4.2.133)
   
pc 0 : Initial preconsolidation pressure

 vp : Plastic volumetric strain

 vp0 : Initial plastic volumetric strain

Section 2. Plastic Material Properties | 189

190 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

The Hoek-Brown failure criterion for rock masses is widely accepted and has been applied. While, in general, it
2.17 has been found to be satisfactory, there are some uncertainties and inaccuracies that have made the criterion
Generalized Hoek-Brown inconvenient to apply and to incorporate into numerical models. In particular, the difficulty of finding an
acceptable equivalent friction angle and cohesive strength for a given rock mass has been a problem since the
publication of the criterion in 1980. The Generalized Hoek-Brown model resolves all these issues and sets out a
recommended sequence of calculations for applying the criterion. In order to link the empirical criterion to
geological observations by means of one of the available rock mass classification schemes, the Rock Mass Rating
is used17.

Yield function
The non-linear Generalized Hoek-Brown criterion for rock masses defines material strength in terms of major
and minor principal stresses as:

a
 m 
f HB   1   3    ci   b  1  s 
  ci  (4.2.134)
1   2   3
 ci : Uniaxial compressive strength

mb , s, a : Parameter for defining rock mass failure

Here, mb , s , a can be expressed with the parameters related to the geological strength index(GSI) and the
disturbance factor(D).

 GSI  100  ,  GSI  100  , a  e

1 1  GSI /15
 e 20/3  (4.2.135)
mb  mi exp   s  exp  
 28  14 D   9  3D  2 6

GSI : Geological strength index

mi : Intact rock material property

D : Disturbance factor (D=0 for undisturbed rock masses, D=1 for very disturbed rock masses)

Flow rule
If the flow rule is used same as the yield function of Generalized Hoek-Brown, the corner from hexagon should
be handled. However, this difficulty is removed by using the flow rule of conical shape such as Drucker-Prager
model.

17 Hoek E., C. Carranza-Torres, and B. Corkum. 2002. Hoek-Brown criterion – 2002 edition. In Proceedings of the 5th North
American Rock Mechanics Symposium and the 17th Tunnelling Association of Canada: NARMS-TAC 2002, Toronto,
Canada, eds. R.E. Hammah et al, Vol. 1, pp. 267-273.
 ANALYSIS REFERENCE Chapter 4. Materials

I1 (4.2.136)
g HB  S  tan
3

Here, S  3Rmw2 J 2 , 1  sin / 3 and  is dilatancy angle.

Rmw 
2 cos

Figure 4.2.32 shows the shape in stress space of the Generalized Hoek-Brown model. The tensile stress is the
following equation.

s ci (4.2.137)
t 
mb

Figure 4.2.32 Hoek-Brown failure

 3 1
surface in the principal stress
space

 1
t 3
2

The Soft Soil model is suitable for simulation of normally consolidated or near normally consolidated clay soils.
2.18 Although this model is replaced by the advanced constitute model that simulates the hardening behavior better
Soft Soil or the Soft Soil Creep model that simulates the secondary consolidation, the Soft Soil model is better capable
to model the compression behavior of very soft soils. The Soft Soil model uses the yield surface of Modified
Mohr Coulomb model to resolve the convergence problem due to the discontinuity of yield function. The main
features of this model are the stress-dependent nonlinear elastic behavior, the hardening behavior through pre-
consolidation stress and the failure by shear stress.

Nonlinear elastic
The Soft Soil model has the nonlinear elastic characteristic which has the logarithmically relationship between
volumetric strain and mean effective pressure. This is the same stress-dependent stiffness with Modified Cam-
Clay.
The bulk modulus K and the shear modulus G have the following relationship for the effective stress p .

Section 2. Plastic Material Properties | 191

192 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

K
1  e0  p 31  2 
, G K (4.2.138)
 2 1   

e0 : Initial void ratio

 : Slope of over-consolidation line

 : Poisson's ratio

If the tensile strength p is considered from the above equation, the bulk modulus can be expressed as follows:

1  e0 
K  p  p  (4.2.139)


By using the cohesion C and the friction angle  , the tensile strength p can be calculated as follows:

C
p  (4.2.140)
tan 

Yield function and flow rule

The Soft Soil model uses the yield function of Modified Mohr Coulomb model. The yield function and flow rule
of Soft Soil model are same with the equation (4.2.99) and (4.2.103) respectively. For more details, please refer
to the ‘2.10 Modified Mohr-Coulomb’.

Hardening behavior
Even though the Soft Soil model has each yield function of shear and compression, the hardening behavior
occurs for the compression yield function. Same as the Modified Cam-Clay model, the compression hardening
behavior is defined by the pre-consolidation stress which is the function of plastic strain.

 p   p 
pc  pc 0 exp  v v0  (4.2.141)
   
pc 0 : Initial pre-consolidation stress

 v
p
: Plastic volumetric strain

 vp0 : Initial plastic volumetric strain

 ANALYSIS REFERENCE Chapter 4. Materials

The Hardening Soil with small strain stiffness model is a modification of the Modified Mohr-Coulomb model
2.19 that considers the increased stiffness of soils at small strains. This behavior is described in this model using an
Hardening Soil with small additional strain-history parameter and two additional material parameters.
strain stiffness As the sign convention for stresses and strains is displayed that compression is negative and tensile is positive,
it assume that 1   2   3 for the principal stresses and 1   2  3 for the principal strains. For example,
the relationship of 1   2   3 is established in case of triaxial tests.

Nonlinear elastic
In the Modified Mohr-Coulomb model, the following characteristics are used to the stress dependent value.

m m m
ref   1  px    1  px  ref   3  px 
E50  E50  ref  , Eur  Eur
ref
 ref  , Eoed  Eoed  ref 
p p  p p  p p  (4.2.142)
 x   x  x
px  c cot 

c : Cohesion
ref
p : Reference stress for stiffnesses

 : Friction angle

m : Power law for stress dependent stiffness

ref
E50 : Reference secant stiffness in standard drained triaxial test
ref
Eoed : Reference tangent stiffness for primary oedometer loading
ref
Eur : Reference unloading / reloading stiffness

As the stiffness modulus can be changed according to the stress, this model shows the nonlinear elastic
characteristic what the elastic modulus changes.

Yield function, plastic potential function and flow rule

T. Schanz, P.A. Vermeer and P.G. Bonnier18 have developed the Hardening Soil model based on the hyperbolic
relationship between deviatoric stress and vertical strain in the triaxial test, and suggested the shear &
compressive yield function as the following equation.

18 Schanz T., Vermeer P.A., Bonnier P.G. (1999). The hardening-soil model: Formulation and verification. Beyond 2000 in
Computational Geotechnics, Balkema, Rotterdam. pp. 281-290.

Section 2. Plastic Material Properties | 193

194 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

qa q 2q qf
fs   p 0 , qa : and  p :  1p   2p   3p
E50 qa  q Eur Rf
q2 1  1
fc    p  px    pc  px  , q :  1  1  2 3
2 2
(4.2.143)
2    
3  sin 

3  sin 
q : Deviatoric stress

qa , q f : Asymptotic & ultimate deviatoric stress

E50 : Confining stress dependent stiffness modulus for primary loading

Eur : Young's modulus for unloading and reloading

p : Hardening parameter (plastic shear strain)

Rf : Failure ratio q f / q a

f s , fc : Shear & compressive yield function

pc : Pre-consolidation stress

If  v p  0 , the above shear yield function can be considered the strain relationship as the following equation.

qa q 2q
0  1  1e  1 p ,  1  , 1e  , 1 p   p (4.2.144)
2 E50 qa  q Eur

The yield function of Modified Mohr-Coulomb model consists of the shear, compressive and tensile yield
function as the following equation.

2qa  1   3  2  1   3  2 E50
f13   p 0 , Ei :
Ei qa   1   3  Eur 2  Rf
1 2sin 
qa : qf , q f :  1  c cot   (4.2.145)
Rf 1  sin 
q2
  p  c cot     pc  c cot    0
2 2
fc 
2
f t   p  pt

f13 : Shear yield function ( 1   2   3 )

fc : Compressive yield function

ANALYSIS REFERENCE Chapter 4. Materials

ft : Tensile yield function

pc : Pre-consolidation stress

 : Cap parameter

pt : Tensile strength

The maximum shear stress q f is calculated by the Mohr-Coulomb criteria, and the Mohr-Coulomb model is
used in case of q f  1   3  .
 is an auxiliary model parameter which control the value of compressive yield function in p-q space. It is
decided by considering the stress ratio in the normally consolidated state ( K 0nc ) and the friction angle.
The plastic potential function uses the Mohr-Coulomb criteria for shear, and the yield function for compressive
and tensile as the following equation.

g13 
1   3   1   3  sin
m
2 2
2
q
  p  c cot  
2
gc  (4.2.146)
2
gt   p
g13 : Shear plastic potential function (in case of 1   2   3 )

gc : Compressive plastic potential function

gt : Tensile plastic potential function

m : Mobilized dilatancy angle

The mobilized dilatancy angle  m can be obtained from the following equation, and it is limited to satisfy the
condition 0   m   considering physical behavior.

sin m  sin cv 1   3 sin   sin

sin m  ,sin m  ,sin cs  (4.2.147)
1  sin m sin cv   1   3   2c cot  1  sin  sin

m : Mobilized friction angle

cs : Critical state friction angle

 : Dilatancy angle

Section 2. Plastic Material Properties | 195

196 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

The mobilized friction angle and the critical state friction angle are consistent with the Rowe theory as described
by T. Schanz, P.A. Vermeer and P.G. Bonnier.

Hardening behavior
The Modified Mohr-Coulomb model shows hardening behavior while increasing the effective plastic strain and
it reaches the perfect plastic state in case of q f  1   3  as mentioned in the previous yield function.
In the process of compressive hardening, the pre-consolidation stress pc is defined as below:

1
 n 1 m 1 m
pc  px 

pcn1  pcref 
 px   ref 
  pc  px 
1 m
 ref
pc  px
H  vp 

 px
  (4.2.148)
ref
K sref Kcref ref
Eur Eoed
H  , K sref  , Kcref 
K sref  Kcref 3 1  2v  3 1  2v 

pcref : Reference pre-consolidation stress

 vp : Plastic volumetric strain

Hardening Soil with small strain stiffness

The Hardening Soil with small strain stiffness model is implemented by using the Modified Mohr-Coulomb
model and Small strain overlay19 model, and needed two additional parameters as below:

G0 ref : Initial or very small-strain shear modulus

 0.7 : Shear strain at which the shear modulus is about 70% of the initial small-strain shear modulus

The strain range in which soils can be considered truly elastic is very small. With increasing strain range, soil
stiffness decrease nonlinearly as the following graph.

19 Benz, T. "Small strain stiffness of soil and its numerical consequences", PhD thesis, University Stuttgart, 2007.
 ANALYSIS REFERENCE Chapter 4. Materials

Figure 4.2.33 Characteristic

stiffness-strain behavior of soil
with the ranges for typical
geotechnical structures and
different tests

To reflect the above characteristics, the Hardening Soil with small strain stiffness model uses the modified
Hardin & Drnevich relationship20 as the following equation.

Gs 1
 , a  0.385 (4.2.149)
G0 
1 a
 0.7
Gs : Shear modulus

G0 : Initial shear modulus

 : Shear strain

 0.7 : Shear strain at which the shear modulus is about 70% of the small-strain shear modulus

Once the direction of loading is reversed, the stiffness regains a maximum recoverable value which is in the
order of the initial soil stiffness. Then, while loading in the reversed direction is continued, the stiffness
decreases again.

To reflect the above characteristics, the Hardening Soil with small strain stiffness model writes the history of
strain in the internal model, and follows Masing’s rule as follows:

20 B.O. Hardin, V.P. Drnevich, "Shear modulus and damping in soils: Design equations and curves", Journal of the Soil
Mechanics and Foundations Division, 98(SM7):667-692, 1972.

Section 2. Plastic Material Properties | 197

198 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

► The shear modulus in unloading is equal to the initial tangent modulus for the initial loading curve.
► The shape of the unloading and reloading curves is equal to the initial loading curve, but twice its size.
Masing’s rule can be fulfilled by using twice of the initial loading  0.7 for the reloading  0.7 .

The initial shear modulus G0 is calculated by the following equation.

m
   c cot  
G0  G0ref  ref1
 p  c cot   (4.2.150)
 

And the hysteresis shear strain (  H ) is defined as the following equation.

eH
H  3 , e  e : e (4.2.151)
e

H : Strain hysteresis tensor (for more details, refer to Benz19)

e : Incremental shear strain

In the numerical analysis, the following incremental equation is used with the tangential stiffness of modified
Hardin & Drnevich relationship.

 
 n 1 
 H Hn
G n 1 
G0
  , a  0.385 (4.2.152)
n 1
H H
n   n 1 n 
 1  a H 1 a H 
  0.7  0.7 

Soft clay in its natural state has a significant anisotropy in the interior soil particle fabric by deposition and
2.20 consolidation21. Also as the strain is continuously generated, the degree of anisotropy is changed whereby the
Generalized SCLAY1S interior soil particles are rearranged and contact between the particles is changed. The Generalized SCLAY1S
model of FEA NX is rooted in the SCLAY122 model which considered the change due to the initial stress induced
anisopropy of the soft clay and anisotropy of rotational hardening.

21 S.J. Wheeler, M. Cudny, H.P. Neher, C. Wlitafsky, "Some developments in constitutive modelling of soft clays",
Proceedings of the International Workshop on Geotechnics of Soft Soils-Theory and Practice, Noordwijkerhoud, the
Netherlands, 2003, pp. 17-19.
22 S.J. Wheller, A. Naatanen, M. Karstunen, M. Lojander "An anisotropic elastoplastic model for soft clays", Canadian
Geotechnical Journal, 40.2., 2003., pp. 403-418.
ANALYSIS REFERENCE Chapter 4. Materials

On the other hand, the structure of the soil particle is composed of two parts 23: bonding as well as fabric. The
fabric is composed of spatial arrangement of particles and inter-particle contact, and the bonding is weakened
gradually as the plastic straining is caused by the forces acting between the particles. The phenomenon that the
bonding is weakened gradually by the plastic straining is called destructuration, and SCLAY1S model considers
the destructuration phenomenon of the SCLAY1 model additionally.
The initial SCLAY1(S) model was the model assumed the triaxial stress state, and later it is improved by the
model considering the general stress state. The Generalized SCLAY1S model is that the shape of yield function
is complicated and needs more variables to represent the hardening behavior. However, It has a advantage that
it can simulate the behavior of the general stress state strictly as well as the triaxial stress state.
In the SCLAY1S model, ignoring the initial anisotropy and bonding, and in case of assuming a related material
constant to 0, it can be found that the Modified Cam Caly model and the yield function are matched exactly.

Nonlinear elastic
The Generalized SCLAY1S model of FEA NX does the stress-dependent non-linear elastic behavior like Modified
Cam Clay, Sekiguchi-Ohta, Soft-Soil models. This being so, the detailed description and formulas will be
omitted. (Refer to the equation 4.2.139)

Yield function and plastic potential function

As the Generalized SCLAY1S model follows the associated flow rule, it is equal to the yield function and plastic
potential function.
The yield function of the SCLAY1 model simplified of the triaxial stress state is represented about the effective
stress, and the signs of the stress is that compression is (+) and tensile is (-).


f   q   p   M 2   2
2
  pc  p p  0
1 3
p   'kk  ij , q  s:s (4.2.153)
3 2
1
sij   'ij   'kk  ij
3
p : Pressure
q : Shear stress

pc : Preconsolidation pressure

M : Slope of critical state line

 : Degree of anisotropy

23 J.B. Burland, "On the compressibility and shear strength of natural clays", Geotechnique, 40.3, 1990, pp. 329-378.

Section 2. Plastic Material Properties | 199

200 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Substituting 0 to the degree of anisotropy  and summarizing the equation in the equation (4.2.153), it can be
found that the yield function of the Modified Cam Clay is equal to the equation (4.2.77). In other words, the
SCLAY1 model is the generalized model considering the degree of anisotropy from the Modified Cam Clay
model.
Figure 4.2.34 Yield function of q
triaxial stress state

M
1

1

pc p

The generalized yield function in the general stress state is represented as follows.

3  3 
f   s  pα d  :  s  pα d     M 2   α d : α d    pc  p  p  0 (4.2.154)
2  2 

Here, the deviatoric fabric tensor α d of the soil particle is represented by the fabric tensor α  in the same
form of the deviatoric stress tensor.

1
 d ,ij   ij   ij ij (4.2.155)
3

The fabric tensor α  has the following properties.


 kk  ij  1 (4.2.156)

The degree of anisotropy  meaning the slope of the yield function in the simplified of the triaxial stress state
is defined by the deviatoric fabric tensor α d as follows.

3
2  αd : αd  (4.2.157)
2
ANALYSIS REFERENCE Chapter 4. Materials

And preconsolidation stress pc in the generalized SCLAY1S model considered bonding of the soil particle is
represented as follows.

pc  1  x  pci (4.2.158)

pci : Preconsolidation pressure of the intrinsic yield function

x : Degree of bonding

Here, the intrinsic yield function has the same stress in the same fabric, the void ratio, the slope and the limit
state but is the conceptional yield function24 having smaller preconsolidation stress. The detailed description
about the intrinsic yield function and the degree of bonding can be seen part of the hardening behavior.
On the other hand, the Generalized SCLAY1S model of FEA NX is under the allowable tensile pressure to handle
the convergence problem in case of occurring tensile to materials like Modified Cam Clay, Sekiguchi-Ohta
models.

3  3 
f   s  pα d  :  s  pα d     M 2   α d : α d    pc  p  p  pt   0 (4.2.159)
2  2 

Isotropic hardening law

The Generalized SCLAY1S model has three kinds of the hardening laws. Here, the isotropic hardening law is the
hardening law which the general soft clay has, and it is equal to the law of the Modified Cam Clay, Sekiguchi-
Ohta models. The following is the equation representing the general isotropic hardening law.

dpc 
1  e pc d p
v (4.2.160)
 
pc , dpc : Preconsolidation pressure and the change rate of the preconsolidation pressure

d  vp : The change rate of the volumetric plastic strain

 : The gradient of the normal consolidation line

 : The gradient of the over-consolidation line

e : The void raio

24 M. Karstunen, C. Wiltafsky, H. Krenn, F. Scharinger, H.F. Schweiger, "Modelling the behaviour of an embankment on
soft clay with different constitutive models", International journal for numerical and analytical methods in geomechanics,
30.10, 2006, pp. 953-982.

Section 2. Plastic Material Properties | 201

202 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

The isotropic hardening behavior of the Generalized SCLAY1S model is equal to the equation (4.2.160) but the
using material constants are changed and the object of hardening turns to the preconsolidation stress of the
intrinsic yield function.

 
dpci
1  e  pci d  p
v (4.2.161)
i  
i : The gradient of reconstituted soil(or intrinsic) normal consolidation line

Figure 4.2.35 Compression line of q

natural and reconstituted soil

M
1

1

pc p

The reconstituted soil without bonding and the natural soil having the initial bonding follow the intrinsic
compression line of the picture 4.2.35 and the general compression line respectively. The general compression
line generates the yielding at the bigger effective pressure than the reconstituted soil, and converges to the
compression line of reconstituted soil as the bonding decreases gradually. At this time, generally the gradient
of the post-yield compression curve  has a lot bigger value than the gradient of the reconstituted soil line i ,
but is is due to the destructuration that the bonding of natural soil reduces gradually.

Rotational hardening law

The rotational hardening law simulates the behavior changing the degree of anisotropy as the plastic strain
changes, but it causes the hardening so that the anisotropy disappears as the plastic strain increases.
The rotational hardening law in the simplified of the triaxial stress state is represented as follows.

 3  1  
d        d vp        d dp  (4.2.162)
 4   3  
 , d : Degree of anisotropy and the change rate of the degree of anisotropy
ANALYSIS REFERENCE Chapter 4. Materials

d  vp : The change rate of volumetric plastic strain

d  dp : The change rate of shear plastic strain

q
 : , Ratio of the shear stress and pressure
p
 : Coefficient of the absolute effectiveness of the rotational hardening law

 : Coefficient of the relative effectiveness of the rotational hardening law

Here,  and  are defined as follows.

x x  0  x x0
x  , x  (4.2.163)
0 x  0  x x  0

In the equation (4.2.162), it can be found that the change rate of the degree of anisotropy  is growing
together as the volumetric plastic strain and shear plastic strain grow. However, as the degree of anisotropy is
closer to    3 4 or   1 3 , the contribution of the volumetric plastic strain or the shear plastic
strain affecting to the change rate of the degree of anisotropy are reduced.
On the other hand, numerical problems happen in the dry side likewise other material models following the limit
state theory, Macaulay bracket  included in the equation (4.2.162) is to prevent the degree of anisotropy
effusing when the yielding occurs in the dry side.
The rotational hardening law of generalized model changes as follows.

 3  1  
dα d    η  α d  d vp    η  α d  d  dp  (4.2.164)
 4  3  
αd , dαd : Deviatoric fabric tensor and the change rate of the deviatoric fabric tensor
s
η : η , The raio of deviatoric stress and pressure
p

The equation (4.2.164) is similar to the equation (4.2.162), but hardening variables are changed from the scalar
values to the secondary tensor values corresponding to the respective stress components.

Destructuration law
The third hardening law, destructuration law simulates that the degree of bonding decreases gradually as the
plastic strain occurs.

Section 2. Plastic Material Properties | 203

204 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

  
dx  a 0  x  d  vp  b 0  x  d  dp  ax d  vp  bd  dp  (4.2.165)

x, dx : The degree of bonding and the change rate of the degree of bonding

a : Coefficient of the absolute effectiveness of the destructuraion law

b : Coefficient of the relative effectiveness of the destructuraion law

Similar to the rotational hardening law, the change rate of the degree of bonding also grows as the volumetric
plastic strain and the shear plastic strain grow. However it is irrelevant to the sign of the volumetric plastic strain
and only affected by its magnitude. Also, it can be found that the change rate of the degree of bonding is
reduced as the degree of bonding x is closer to x  0 , because it simulates the weakened bonding
phenomenon as the plastic strain increseas as a result.

By the development of the tunnel excavation technology, it is possible to construct structures in deep geological
2.21 environments and bedrock, and these structures under the high confining pressure can be occurred brittle
CWFS (Cohesion fracture like spalling or slabbing by the excavation of the cavity. These failure phenomena can not be predicted
Weakening and Frictional properly with perfectly elastoplastic, strain softening, brittle models applying the traditional failure criteria.
CWFS model predicted the swelling effects of bedrock and the failure behavior in deep geological environments
Strengthening) more exactly than the brittle model, therefore this model is included in FEA NX.

Shear yield function

CWFS model taking advantage of the Mohr-Coulomb yield function is that the hardening/softening behavior of
table is possible. Therefore shear plastic behavior is represented as follow equation.

1
f s  Rmc 3J 2  I1 tan     c    0 (4.2.166)
3
 : Friction angle

c : Cohesion

 : Equivalent plastic strain

Shear flow rule

The flow rule uses the following plastic potential based on the non-associated plastic flow rule. Here, It uses the
smoothing formula in order to avoid singularity occurred in the corner. For more information, refer to chapter
modified UBCSAND.
 ANALYSIS REFERENCE Chapter 4. Materials

I1
gs   c tan m   3Rmw 2 J 2  tan
2

3
4 1  e 2  cos 2    2e  1 (4.2.167)
2
 
Rmw  Rmc  , 
2 1  e 2  cos    2e  1 4 1  e 2  cos 2   5e 2  4e 3 

 : Dilatancy angle

Shear hardening behavior

In order to define the shear hardening, the relation of the plastic multiplier  and the hardening variable  are
defined as follows.

 
1
3
1  sin 2   (4.2.168)

FEA NX can define the hardening behavior about cohesion c , friction angle  and dilatancy angle  using
the table.

Geogrid is generally used as the material for reinforcing soil/ground. Geogrid is made of the polymer fabric, and
2.22 it is working with the weight of soil/ground. It is only resisted the tension, and mainly used as a sub-material of
Geogrid reinforced earth retaining wall.

Nonlinear elastic
The geogrid material in FEA NX shows the tension-only behavior. The stress-strain relationship of geogrid is
shown in the following figure.

Figure 4.2.36 Tension-only

behavior of geogrid 

tension 

The 2D geogrid shows an independent behavior each other in the axial plane.

Section 2. Plastic Material Properties | 205

206 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

 E1 
 0 0 
 xx  1  
2
  xx 
    
0   yy 
E2
 yy    0 (4.2.169)
   1  2
 xy   0   
0 G12   xy 
 

Yield function
The yield function and plastic potential function of geogrid material are same since it follows the associated flow
rule.
The plastic behavior of each direction is independently and the yield condition is as follows.

f (σ)     yield  0 (4.2.170)

The yield condition of 1-axis and 2-axis can be applied differently.

ANALYSIS REFERENCE Chapter 4. Materials

In FEA NX, the concrete smeared crack model is used to simulate the plain concrete. The concrete-smeared
2.23 crack model assumes that the monotonic load is applied to concrete under low confining pressure. The low
Concrete smeared confining pressure corresponds to the stress less than 1/5 to 1/4 of the maximum compressive strength of
crack concrete. For reinforced concrete, the embedded elements are used. In this case, effects such as bond slip from
the interface of concrete and steel are considered through tension stiffening.25
The concrete smeared crack model simulates the compression part of concrete using a typical isotropic elasto-
plastic model, and the tensile part of concrete is simulated using a smeared crack model. The smeared crack
model is a method of simulating the crack by adjusting the stress and stiffness at the integration point, without
reconfiguring the mesh. In the smeared crack model, the band-width obtained based on the size of the element
is reflected in the crack behavior to avoid mesh dependencies (Bazant and Oh 1983). The square root of an area
for a two-dimensional element and the cubic root of a volume for a three-dimensional element are the crack
width of the element. For the higher order element, the half of the square root and the cubic root are the crack
width. The tensile/compression uniaxial behavior of the concrete smeared crack model is shown in the figure
below.

Figure 4.2.37 Failure point

Uniaxial behavior of concrete in compression
smeared crack

Start of inelastic
Unload/reload
behavior
response


Cracking failure

25 H. D. Hibbit, "A simplified model for concrete at low confining pressure", Nuclear engineering and design,

104.3, 1987, 313-320

Section 2. Plastic Material Properties | 207

208 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Strain Decomposion
The following strain decomposition is used in the concrete smeared crack model.

dε  dεel  dεcpl  dεtpl (4.2.170)

dε cpl : Incremental change of compressive plastic strain

dεtpl : Incremental change of tensile plastic strain

The above equation can be integrated as follows.

ε  εel  εcpl  εtpl (4.2.171)

Compressive behavior
The concrete smeared crack model simulates compressive behavior using a typical elasto-plastic model.

Compression plastic flow

Associated flow ruls is used.

  p
2
 f
dε cpl  d c 1  c0    c (4.2.172)
 c   σ
 

c0 is the value which makes the ratio of ε11 of the monotonic biaxial test and ε11 of the monotonic uniaxial test
pl pl

to be constant.

Compression yield surface

fc  q  3a0 p  3 c  0 (4.2.173)
1
p : Effective stress (  trace  σ  )
3
3
q : Mises equivalent effective stress ( S:S , S : deviatoric stress )
2

a0 is the value calculated from the ratio of the compressive strength in the uniaxial stress state to the
compressive strength of the biaxial stress state.
 c is the yield stress under the pure shear stress, calculated from the compression curve entered by the user.
ANALYSIS REFERENCE Chapter 4. Materials

 1 a0 
c     c (4.2.174)
 3 3

Tensile behavior
Cracks occur when the stress is outside the crack detection surface, a type of yield function. If a crack occurs,
the stress is returned using an implicit backward Euler method, and the direction of the maximum principal
strain is the direction of the crack. This direction is stored to simulate the anisotropy by cracking in subsequent
analyses. Since the orthogonal fixed crack model is used in the smeared crack model, the subsequent crack shall
be in a direction orthogonal to the existing crack direction and up to three in the case of the three-dimensional
model. If the elastic strain in the direction of the crack is tensile after the crack has occurred, it is behaved as
damaged elastic and it is considered that the crack is closed when it is compressed.
Figure 4.2.38
Tensile behavior of concrete 
semared crack
Failure point

Unload/reload
response


Crack detection surface

Crack detection surface is Coulomb line.

    b  
f t  qˆ   3  b0 ut  pˆ   2  0 ut   t  0 (4.2.175)
  t   3 t 

 tu : Failure stress in uniaxial tension

p̂ , q̂ : p and q for which open-crack related components are neglected.

b0 is obtained by using the tensile strength value when the stress in one direction of the main stress is the
compressive strength in a plane stress condition.

Section 2. Plastic Material Properties | 209

210 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Tension flow rule

As with compression, associated flow ruls is used.

f t
dε tpl  d t (4.2.176)
σ

Damaged elasticity
The relationship between stress-elastic strain of concrete-smeared crack model is as follows:

σ  D : ε el (4.2.177)
D : Elastic Stiffness Matrix of Concrete Model
In uncracked concrete, D is a linear isotropic elastic matrix. If the elastic strain in the direction of the crack is
tensile after the crack occurs, the stress and stiffness are behaved as damaged elasticity by reference to the
tensile curve entered by the user.
If  is the direction of the crack, the stress and elastic strain in that direction are   and 
el
. If 
el
 0 , the
component,  , of the concrete elastic stiffness matrix resulting from Poisson ratio is not considered.

D  0 for    ,    (4.2.178)

If the maximum value of the elastic strain produced in the  direction is  

open
, and the stress at the
corresponding elastic strain is  
open
, the damaged components of the elastic stiffness matrix are as follows.

 
open

D  (4.2.179)
 
open

The shear stiffness associated with the direction of the crack is calculated by the shear function entered by the
user.
ANALYSIS REFERENCE Chapter 4. Materials

In reinforced concrete, the interaction between the reinforcement and the concrete is governed by secondary
2.24 transverse and longitudinal cracks in the vicinity of the reinforcement. This behavior can be modeled with a
Bond Slip bond-slip mechanism where the relative slip of the reinforcement and the concrete is described in a
phenomenological sense.
In FEA NX, the relationship between the normal traction and the normal relative displacement is assumed to
be linear elastic, whereas the relationship between the shear traction and the slip is assumed as a nonlinear
function.

tn  kn un
(4.2.180)
tt  f t  ut 

Differentiating (4.2.180) results in expressions for the tangential stiffness coefficients.

 D11  kn D12  0 
 
 D21  0 ft  (4.2.181)
D22 
 ut 

FEA NX offers a predefined curve, ‘polynomial function’, for the relationships between shear traction and slip,
and a user-defined multi-linear function is also available.

Polynomial function
The polynomial function describes the relationship between shear stress and slip as shown in the figure below,
and the formula is shown below.

Figure 4.2.39
Shear behavior of bond slip

Section 2. Plastic Material Properties | 211

212 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

   u   u 
2
 u  
3

 ft 0  a  b  t   c  t   d  t   if 0  ut  ut 0
ft     ut 0   ut 0   ut 0   (4.2.182)


 ft 0  a  b  c  d  if ut 0  ut
ANALYSIS REFERENCE Chapter 4. Materials

The constitutive law for discrete cracking in FEA NX is based on a total deformation theory, which expresses
2.25 the tractions as a function of the total relative displacements. The relationship between normal traction and
Discrete Cracking crack width and the relationship between shear traction and slip are assumed as nonlinear functions.

tn  f n  un 
(4.2.183)
tt  f t  ut 

In the above equation, the relationship between the normal traction and shear traction is independent to each
other, so the stiffness can be expressed as follows.

 f n 
 D11  u D12  0 
 n  (4.2.184)
 f t 
 D21  0 D22  
 ut 

In general, the normal traction is governed by a tension softening relation. For structural interface elements,
FEA NX supports the following relations:

1. Brittle cracking model

2. Linear tension softening model
3. Nonlinear tension softening model

Brittle cracking model

Brittle cracking model is characterized by the full reduction of the strength after the strength criterion has been
reached.

 ft
 f n  un  if un 
 kn
tn   (4.2.185)
 0 ft
if  un
 kn

Section 2. Plastic Material Properties | 213

214 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Linear tension softening model

In case of linear tension softening, the relation of the crack stress and displacement in the normal direction is
given by the figure below.
Figure 4.2.40
Linear tension softening tn
behavior

ft

G fI

∆u n
∆u n,ult

  un 
 f t 1   if 0  un  un ,ult
tn    un ,ult 
 (4.2.186)

 0 if un ,ult  un

un ,ult G If
: 2 ft

G If : Mode-I fracture energy

ft : Tensile strength

Unloading and reloading can be modeled according to a secant approach or an elastic approach. In the secant
approach, the relation between the traction and the relative normal displacement is linear up to the origin, after
which the initial stiffness is recovered. In the elastic approach, the initial stiffness is recovered immediately after
the relative normal displacement has become less than the current maximum relative normal displacement.
ANALYSIS REFERENCE Chapter 4. Materials

Nonlinear tension softening model

Hordijk , Cornelissen & Reinhardt proposed an expression for the softening behavior of concrete as shown in
the figure below
Figure 4.2.41
Linear tension softening tn
behavior
ft

∆u n
∆u n,ult

    
3
  1   c un   exp  c un 
    u  
1
 2 un ,ult 
 ft    n , ult 
   if 0  un  un ,ult
   
tn     u n
(1  c13 )exp(c2 )  (4.2.187)
  un ,ult 



 0 if un ,ult  un


un ,ult G If
:  5.136 ft

c1 : 3
c2 :  6.93

Unloading and reloading can be modeled according to a secant approach, an elastic approach or by application
of hysteresis.

Section 2. Plastic Material Properties | 215

216 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Shear Retention
In general, the shear traction is reduced after cracking according to the following equation.

 k u if ut  ut 0
tt   t t (4.2.188)
  kt ut if un  ut 0

 kt : Reduced shear stiffness

ut 0 : Initial shear slip

In general,  may vary between 0.1 and 0.3. If the crack surface is assumed to be smooth after Mode-I
cracking,  is defined as zero. But generally, it is assumed that the crack surface is not smooth and hence
0   1.
ANALYSIS REFERENCE Chapter 4. Materials

The Masonry material model is a material model that simulates the behavior of the Masonry wall under lateral
2.26 load using homogenization techniques. First, homogenize the brick and the head joint, and then homogenize
Masonry the homogenized material and the bed joint to obtain the physical properties of the entire homogenized
material. In the masonry material, tensile cracking is considered to be the only nonlinearity. Cracks can occur on
both bricks and on two joints. In the event of a crack, the elastic modulus of the crack is reduced by a
predetermined proportion. Then, the newly homogenized material is obtained by reflecting the reduced elastic
modulus.

The assumptions applicable to the Masonry Model are as follows.

1. Brick has linear elastic behavior and brittle failure occurs.
2. Mortar behaves linear elastically.
3. Brick and mortar are perfectly bonded.
4. Head joints are continuous.

Figure 4.2.42 shows a coordinate system of the Masonry Model for homogenizationi, where indexes b, hj, and
bj represent bricks, head joints, and bed joints, respectively.
Figure 4.2.42
Coordinate system of
masonry model

hb

tbj

x tb
thj

The stress/strain relationship of the homogenized masonry material is represented by

ε  Cσ (4.2.189)

Where C is the compliance matrix for the homogenized Masonry model and is obtained by nine independent
modulus of elasticity as shown below.

Section 2. Plastic Material Properties | 217

218 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

 1 vxy vxz 
   
 Ex Ex Ex 
 v yx 1 v yz 
  
 Ey Ey Ey 
 
  vzx vzy 1 

 E Ez Ez 
C z
 (4.2.190)
 1 
 Gxy 
 
 1 
 
 G yz 
 1 

 Gxz 

Structural matrix A is used to obtain the stress of each brick and joint from the average stress of the Masonry
model.

σbj  Abj σ
σ hj  Ahj σ (4.2.191)
σb  Abσ

Homogenization of brick and head joints

Calculate the ratio of the length of the brick and head joint.

lb t
tb  , t hj  hj (4.2.192)
lb  thj lb  thj

Calculate the coefficients needed to obtain the physical properties.


   
thj Ehj 1  vb2  tb Eb 1  vhj2
  
1  v 12

hj v 2
b

thj vhj E 1  v   t v E 1  v 
2 2

 
hj b b b b b
(4.2.193)
1  v 1  v 
2
hj
2
b

tb vb thj vhj
b  ,  hj  ,    b   hj
 b
1  v 1  vhj  
ANALYSIS REFERENCE Chapter 4. Materials

Calculate Poisson’s ratio and modulus of elasticity.


vˆ yz  , vˆzx  vˆ yx   1  vˆ yz 

E y'  Ez'     vzy'

1 t t  vˆ v   vˆ v  (4.2.194)
 hj  b  2  hj  zx  hj   2  b  zx  b 
'
Ex Ehj Eb  Eˆ   Eˆ 
 z Ehj   z Eb 
Eˆ Eˆ Eˆ
vˆxz  vˆzx x , vˆzy  vˆ yz z , vˆxy  vˆ yx x
ˆ
Ez ˆ
Ey Eˆ y

Calculate the shear modulus.

1 1 thj t
 '   b
Gxy' Gxz Ghj Gb (4.2.195)
G  thj Ghj  tbGb
'
yz

Homogenization with bed joints

Calculate the ratio of the length of the brick and bed joint.

hb t
tˆ  , t bj  bj (4.2.196)
hb  tbj hb  tbj

Calculate the coefficients needed to obtain the physical properties.

tbj ˆˆ
tE
 
1 v 2
bj 1  vˆxz vˆzx
tbj Ebj ˆˆz
tE
 
1  vbj2 1  vˆxz vˆzx
tbj vbj Ebj ˆ z vˆxz
tE
   (4.2.197)
1 v 2
bj 1  vˆxz vˆzx

 bj 
tbj vbj
, ˆ 

tˆ vˆzy  vˆzx vˆxy ,     ˆ
1  vbj 1  vˆxz vˆzx
bj

bj   bj ,  
ˆ 
tˆ vˆxy  vˆzy vˆxz ,    ˆ
1  vˆxz vˆzx
bj

Section 2. Plastic Material Properties | 219

220 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Calculate Poisson’s ratio and modulus of elasticity.

  
vzx  , vzy    , vxy   
  
   2    2
Ez  , Ex 
 
1 t tˆ v v  v vˆ 
 bj    bj  xy  bj   ˆ  zy  zy   (4.2.198)
E y Ebj E y ˆ    ˆ 
 Ez Ebj   Ez Ez 
v y v  v vˆ 
m  x  m   ˆ  xy  xy 
 ˆ 
 E x Em   Ex Ex 
Ey Ex E
v yx  vxy , vxz  vzx , v yz  vzy y
Ex Ez Ez

Calculate the shear modulus.

1 tbj tˆ
 
Gxy Gbj Gˆ xy
ˆ ˆ xz
Gxz  tbj  tG (4.2.199)
1 t tˆ
 bj 
G yz Gbj Gˆ yz

Stress of bead joint, head joint, and brick

Structural matrix A , which provides stress of individual components from the average stress of the masonry,
is as follows.

 a11 a12 a13 

 
 a21 a22 a23 
a a a33 
A   31 32  (4.2.200)
 a44 
 a55 
 
 a66 

The coefficients of the structure matrix Abj for the bed joints are as follows:
ANALYSIS REFERENCE Chapter 4. Materials

 1 1  v vbj 
a11bj  1  Cbj Ebj     Cbj Ebj vbj  zx  
 E x Ebj   E z Ebj 
   
 vxy vbj   vzy vbj 
a12bj  Cbj Ebj     Cbj Ebj vbj   
 E x Ebj   E z Ebj 
   
v vbj   1 1 
a13bj  Cbj Ebj  xz    Cbj Ebj vbj   
 E x Ebj   E z Ebj 
   
bj
a22  1.0
(4.2.201)
 1 1  v vbj 
a  Cbj Ebj vbj 
bj
   Cbj Ebj  zx  
31
 E x Ebj   E z Ebj 
   
v v  v v 
a3bj2  Cbj Ebj vbj  xy  bj   Cbj Ebj  zy  bj 
 E x Ebj   E z Ebj 
   
v vbj   1 vbj 
a3bj3  1  Cbj Ebj vbj  xz    Cbj Ebj   
 E x Ebj   E z Ebj 
   
G
bj
a44  1.0, a5bj5  1.0, a6b6j  bj
Gxz

Cbj 1
:
1  vbj2

The head joints and bricks are obtained by multiplying the common structural matrix  by the partial
structural matrix Pb , Phj .

Ahj  Phj A
(4.2.202)
Ab  Pb A

The coefficients for the common structural matrix are as follows:

Section 2. Plastic Material Properties | 221

222 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

   
ˆ ˆ  1  1   CE
aˆ11  1  CE ˆ ˆ vˆ  vzx  vˆzx 
z xz  
x
 Ex Eˆ  ˆ
 x   Ez Ez 
 ˆ   ˆ 
aˆ12  CEˆ ˆ  vxy  vzy   CE
ˆ ˆ vˆ  vzy  vzy 
z xz 
x
 Ex Eˆ  E ˆ 
 x   z Ez 
   
aˆ13  CEˆ ˆ  vxz  vˆxz   CE
ˆ ˆ vˆ  1  1 
z xz  
x
E ˆ  ˆ
 x Ex   Ez Ez 
aˆ22  1.0
(4.2.203)
   
ˆ ˆ vˆ  1  1   CE
aˆ31  CE ˆ ˆ  vzx  vˆzx 
z 
x zx
 Ex Eˆ  ˆ
 x   Ez Ez 
 ˆ   ˆ 
ˆ ˆ vˆ  vxy  vxy   CE
aˆ32  CE ˆ ˆ  vzy  vzy 
z
x zx
E ˆ  E ˆ
 x E x   z E z 

ˆ ˆ vˆ  vxz  vxz  ˆˆ  1 1 
aˆ33  1  CE x zx    CE z   
ˆ 
 Ex Ex   Ez Ez 
Gˆ
aˆ44  1.0, aˆ55  1.0, aˆ66  xz

Gxz

1
Ĉ :
1  vxz vˆzx
ˆ

The coefficients for the partial structural matrix are as follows:

ANALYSIS REFERENCE Chapter 4. Materials

p11  1.0
 vˆ v   vˆ yx v 
p21  Ci Ei vi  zx  i   Ci Ei   i 
 Eˆ E   Eˆ y Ei 
 z i   
 vˆ v   1 1 
p22  1  Ci Ei vi  zy  i   Ci Ei   
 Eˆ   Eˆ y Ei 
 z Ei   
 1 1   ˆ
v yx vi 
p23  Ci Ei vi     Ci Ei vi   
 Eˆ E   ˆ Ei 
 z i   Ey  (4.2.204)
 vˆ v   vˆ yx v 
p31  Ci Ei  zx  i   Ci Ei vi   i 
 Eˆ E   Eˆ y Ei 
 z i   
 vˆ v   1 1 
p32  Ci Ei  zy  i
 Eˆ   Ci Ei vi   
 Eˆ y Ei 
 z Ei   
 1 1   vˆ yz v 
p33  1  Ci Ei     Ci Ei vi   i 
 Eˆ   Eˆ y Ei 
 z Ei   
Gi
p44  1.0, p55  , p66  1.0
Gˆ yz

Ci 1
:
1  vi2

In the above expression, index i is a head joint (hj) or brick (b).

Section 2. Plastic Material Properties | 223

224 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Combined-Cracking-Shear-Crushing (CCSC) model in FEA NX is based on the formula presented by Lourenco26

2.27 and enhanced by Van Zijl 27 . It is based on multi-surface plasticity, comprising a Coulomb friction model
Combined combined with a tension cut-off and an elliptical compression cap. Softening acts in all three modes and is
Cracking-Shearing- preceded by hardening in the case of the cap mode.
Crushing

Figure 4.2.43
Three yield surfaces in two
dimension

The interface model is derived in terms of the generalized stress and strain vectors.

  u 
σ   ,ε n (4.2.205)
 tt   ut 

D  diag kn ks kt  (4.2.206)

 , un : Stress and relative displacement in the normal direction

t t , ut : Stress and relative displacement in the tangential directioin
kn : Stiffness in the normal direction
ks , kt : Stiffness in the tangential direction

Shear slip failure

A Coulomb friction model describes the shear-slipping in the CCSC model and adhesion and internal friction
angle assume softening behavior.

26 P. B. Lourenço, “Compulational strategies for masonry structures”, Delft University Press,

1996
27 G.P.A.G. van Zijl, “Compulational modelling of masonry creep and shrinkage”, The

Netherlands by Meinema BV, Delft, 1999

ANALYSIS REFERENCE Chapter 4. Materials

f  tt     c
c0
 
G IIf
c  c0e (4.2.207)
c c
  0  r  0  0
c0

 : Internal friction angle

0 ,  r : Initial friction ( tan  ) and residual friction
c0 : Initial adhesion

 : Plastic relative displacement

G IIf : Mode-II fracture energy

From the observations of experiments27, shear failure energy and normal stress assume the following linear
relationships.
 a  b if   0
G IIf   (4.2.208)
 b if   0

a, b : Material constants

Tension cut-off
Tensile failure and softening behavior are defined as follows, and plastic flow functions are defined using the
associated plastic flow rule.

ft     t

ft
t
(4.2.209)
G If
 t  ft e
t : Tensile plastic strain
ft : Tension force
I
G f : Mode-I fracture energy

Compression cap
The yield function of the compression cap is defined as follows, and the function of compressive plastic flow is
defined using the associative plastic flow rule.

Section 2. Plastic Material Properties | 225

226 | Section 2. Plastic Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

f c   2 Cs tt   c2
2

 c  εTp ε p (4.2.210)

 c   c  fc , p 

c : Compressive plastic strain

c : Compression force
Cs : Shear stress contribution parameter

fc : Compression failure energy

p : Plastic strain at peak

Compression force is expressed by the initial hardening behavior and softening behavior defined by
compression failure energy.

Figure 4.2.44
Hardening and softening
behavior for compression

Therefore, the compression force is defined in three areas as follows:

ANALYSIS REFERENCE Chapter 4. Materials

2 c  c2
 c1  c    i   f c   i   , i  1
fc
p  p2 3

2
  p 
  c   f c   m  f c   c
2
 , m  1
fc (4.2.211)
c
 m   p  2
 
   f   c  m 
 c3  c    r   m   r  exp  2  m c     , r  1
fc
  m   p  m r
 
7
  

Section 2. Plastic Material Properties | 227

228 | Section 3. Drained/Undrained Materials

Chapter 4. Materials
ANALYSIS REFERENCE

Section 3 Drained/Undrained Materials

The pore pressure in stress analysis can be divided into the steady state pore pressure, which does not change
3.1 with time, and the unsteady pore pressure, which changes with time or changes in load/boundary states, for
Isotropic Materials convenience.

p  psteady  punsteady (4.3.1)

Here, the state where the unsteady pore pressure is close to '0' is called the drained condition and the analysis
under this condition is called drained analysis. Generally, it is appropriate to perform drained analysis for the
following states.

► When the change in steady state pore pressure is insignificant, due to external boundary conditions or use of
sand like materials which have large coefficients of permeability
► When simulating the process after consolidation, where the excessive pore pressure has been dissipated

Pore water can display instantaneous undrained behavior, due to the use of clay like materials with small
coefficients of permeability or external conditions such as the existence of impermeable layers. In this case, un-
negligible unsteady pore pressure occurs for the change in external load state. This unsteady state pore pressure
is called the excessive pore pressure. When the pore pressure is assumed not to change with the seepage
condition time, it is determined by the permeability coefficient and the volume change of the porous ground
due to the compressibility of the pore water. FEA NX uses this process of dissipating excessive pore pressure,
caused by load state change, with time to simulate consolidation analysis.

The state where excessive pore pressure occurs due to the compressibility of the pore water is called the
undrained condition, and the analysis under this condition is called undrained analysis. The general undrained
conditions are as follows.

► When the permeability coefficient is small, or the load change is very large
► When instantaneous behavior and safety due to load change is of interest

Drained conditions do not have stiffness for change in load condition. Hence, drained materials follow the
3.2 behavior of the ground skeleton material.
Undrained Constitutive On the other hand, undrained material models consider the stiffness for the compressibility of the pore water in
Equation addition to the ground skeleton material behavior. Pore water does not have shear stiffness and has volume
change stiffness. The relationship between the excessive pore pressure change and volume change of the
ground is as follows.
ANALYSIS REFERENCE Chapter 4. Materials

K w T mech (4.3.2)
dpunsteady  dpexcess   m dε
n

Kw : Bulk modulus of water

n : Porosity

m : Normal unit vector (in 3D, m  [1,1,1, 0, 0, 0]T )

dε mech : Amount of mechanical strain change

Assuming saturated linear elastic istropic materials for convenience, the modified strain-stress relationship can
be obtained.

dσ  dσ ' mdpexcess  Cel ,u dε mech (4.3.3)

K
Cel ,u  Cel  w m mT : Undrained elasticity matrix
n
Cel : Effective elasticity matrix

The bulk modulus of water ( Kw ) is generally a very large value and so, the undrained Poisson’s ratio  u is close
to 0.5 and the porous material displays nearly incompressible behavior. When partial low order elements are
used in this case, volumetric locking occurs and the accuracy of the solution falls greatly. Hence, modeling using
high order elements is recommended for undrained analysis.

To account for undrained effects and guarantee the solution accuracy, FEA NX does not use the bulk modulus
of water directly and uses the user input undrained Poisson's ratio or Skempton factor ( B ) as a base to back-
calculate the bulk modulus of water directly. The undrained Poisson's ratio has a default value of '0.495'.

Using equation (4.3.3), the equation for calculating the bulk modulus of pore water is as follows, when the
undrained Poisson's ratio or Skempton factor is given.

Kw E '  u  '
Kf   (4.3.4)
n 1  2 '1  '1  2 u 
Kf : Bulk modulus of pore water
E ',  ' : Effective modulus of elasticity and effective Poisson's ratio

The Skempton factor ( B ) is defined as the ratio between the undrained bulk modulus and bulk modulus of pore
water. In other words,

Section 3. Drained/Undrained Materials | 229

230 | Section 3. Drained/Undrained Materials

Chapter 4. Materials
ANALYSIS REFERENCE

K f  BK u (4.3.5)
Ku : Undrained bulk modulus

Using equations (4.3.2) and (4.3.4), the relationship between the undrained Poisson's ratio and Skempton factor
can be expressed as follows.

3 ' B(1  2 ')

u  (4.3.6)
3  B(1  2 ')

Using this, the bulk modulus of pore water can be calculated. From equation (4.3.6), it can be seen that the
undrained Poisson's ratio approaches '0.5' as the Skempton factor approaches '1'.

Effective stiffness/effective strength

3.3 This is the most general case where the input stiffness parameters and strength parameters are the parameters
Undrained Material Type of the ground skeleton. Like drained analysis, FEA NX uses the input stiffness/strength parameters for undrained
analysis. The disadvantage is that the effective strength parameters in the undrained state are hard to obtain
through experimentation.

► Available material model

: Linear elastic material, Mohr-Coulomb, Drucker-Prager, Duncan-Chang, Hoek-Brown, Strain Softening,
Modified Cam-clay, Jardine, D-min, Modified Mohr-Coulomb, User-supplied, Modified UBCSAND, Sekiguchi-
Ohta

Effective stiffness/undrained strength

The undrained load path of simple material models like the Mohr-Coulomb model is known to be difficult to
express accurately. Hence, the undrained shear stiffness, which is determined by the friction angle and cohesion,
can be overestimated. If the empirical undrained shear stiffness ( su ) is known, the undrined strength can be
directly input using the cohesion when the friction angle is 0. By using the actual undrained strength, results that
satisfy the shear stiffness can be obtained. However, this case has the same disadvantage that the undrained
load path is difficult to estimate accurately.

► Available material model

: Mohr-Coulomb, Drucker-Prager, Modified Mohr-Coulomb

Undrained stiffness/undrained strength

This method directly inputs the undrained stiffness, which considers the compressibility of the pore water.
Hence, FEA NX does not calculate the excessive pore pressure and its effects are included in the calculated stress.
In other words, the calculated stress is the total stress that includes the pore pressure. The undrained stiffness
and undrained strength parameters can be directly input when known by lab testing.
ANALYSIS REFERENCE Chapter 4. Materials

► Available material model: Linear elastic material, Mohr-Coulomb, Drucker-Prager, Modified Mohr-Coulomb

Section 3. Drained/Undrained Materials | 231

232 | Section 4. Seepage Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Section 4 Seepage Material Properties

Darcy's Law is used to display the seepage phenomena within the ground.
4.1
Constitutive Equation 1
q  kh  kn g  kpw (4.4.1)
w
q : Seepage velocity
k : Permeability coefficient matrix
h : Total head
ng : Gravitational direction unit vector

Darcy's Law expresses the proportionality between the ground seepage velocity and the total head gradient.
Darcy's Law was originally derived for saturated soils, but various researches have shown that it can be applied
to unsaturated flow. Also, Darcy's Law is effective for slow viscous flow and can be applied to most groundwater
flow.

For seepage materials, FEA NX uses the permeability coefficient matrix that considers only the diagonal
component of each direction. Here, the direction is the MCS direction.

kx 0 0
 
k0 ky 0 (4.4.2)
0 k z 
 0

The seepage velocity q has velocity units, and the actual flow velocity in the soil has a value of the seepage
velocity q divided by the porosity of the soil.

q
v (4.4.3)
n

The permeability coefficient is a criterion for how much the groundwater within the soil moves in unit time and
4.2 it is dependent on the water content and void ratio change  e of the ground. The permeability coefficient has
Permeability Coefficient the largest value at the saturated state because the path size increases with larger water content. Also, because
water content is dependent on pore pressure, the permeability coefficient also changes with pore pressure. The
void ratio change is considered in consolidation analysis and fully coupled stress-seepage analysis. The void ratio
change is calculated from the initial condition of the void ratio.
 ANALYSIS REFERENCE Chapter 4. Materials

To express the change in permeability coefficient with pore pressure and void ratio change, FEA NX uses the
saturated permeability coefficient k sat , permeability ratio function kr  kr ( p) depending on pore pressure
change. The ck that defines the permeability ratio depending on the void ratio change  e . The unsaturated
permeability coefficient can be found using the following equation.

e

k  10 ck kr ( p) k sat (4.4.4)

Figure 4.4.1 Permeability ratio for

negative pore pressure
Permeability Ratio

0 Negative Pore-Water Pressure

The permeability coefficient with the pore pressure is directly input as a table in FEA NX, but widely known
permeability coefficient formulas can be used. The supported formulas in FEA NX are as follows and the h
represents the negative pore pressure head.

Gardner function

1
kr  (4.4.5)
1  a(h)n
a, n : Curve fitting parameters

Frontal function

h0 kr  1
h(r  1)
0  h  H0 kr  1 (4.4.6)
H0
h  H0 kr  r

Section 4. Seepage Material Properties | 233

234 | Section 4. Seepage Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

r : Minimum permeability ratio

H0 : Limit negative pore pressure head

Van Genuchten function

 
2

    
1  (ah)n1 1  (ah)n  m 
kr   (4.4.7)
n m/2
1   ah  
 
a, n, m : Curve fitting parameters

When water flows within the ground, a certain amount is retained and this amount is determined by the ground
4.3 properties and capillary suction. This is called the water content. Seepage analysis generally uses the volumetric
Volumetric Water water content, the ratio between the total volume and water volume.
Content
Vw
  nS (4.4.8)
V
 : Volumetric water content
Vw : Water volume
V : Total volume
n : Porosity
S : Degree of saturation

The change in volumetric water content for pore pressure is used for element calculation for seepage and
consolidation analysis, as explained in chapter 3. Differentiating equation (4.4.8) for pore pressure can express
it using the porosity and degree of saturation.

 n S
S n (4.4.9)
p p p

The first term of the right hand side represents the slope of the volumetric water content for the saturated
condition. This term is represented using the specific storage Ss , which represents the volumetric ratio of the
water inflow or outflow in the ground due to the pore pressure head change.

n Vv h Ss
S   (4.4.10)
p h p 
Vv : Void volume
 ANALYSIS REFERENCE Chapter 4. Materials

The second term of the right hand side represents the slope of the volumetric water content for the unsaturated
condition. This value uses the slope of the user input soil-water characteristic curve. The soil-water characteristic
curve represents the relationship between the volumetric water content and pore pressure for unsaturated
conditions. The general curve is shown in Figure 4.4.2.

Figure 4.4.2 Soil-water

characteristic curve

Volumetric Water Content

0 Negative Pore-Water Pressure

Like the permeability coefficient, the volumetric water content is directly input as a table. Widely known
formulas can also be used. The supported formulas in are as follows. p represents the negative pore pressure
head.

Van Genuchten fuction

s  r
  r  (4.4.11)
1  (ap) 
n m

r : Minimum volumetric water content

s : Maximum volumetric water content
a, n, m : Curve fitting parameters

The permeability coefficient and volumetric water content above were defined individually for pore pressure or
4.4 pore pressure head. However, nonlinear characteristics (permeability coefficient and volumetric water content)
Ductile Function of actual soils are affected by the pressure head change simultaneously in a coupled form.

FEA NX reflects these characteristics and uses ductile function forms (pressure head-water content, water
content-permeability ratio function or pressure head-degree of saturation, degree of saturation-permeability
ratio function) to define the characteristics of unsaturated materials. When the pressure head-degree of

Section 4. Seepage Material Properties | 235

236 | Section 4. Seepage Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

saturation and degree of saturation-permeability ratio ductile function are defined, the volumetric water
content is calculated using the porosity.
 ANALYSIS REFERENCE Chapter 4. Materials

Section 5 Viscous Material Properties

The typical behavior of visco-elastic and visco-plastic material is appeared to creep (increasing strain at constant
stress) and stress relaxation (decreasing stress at constant strain) phenomenon. Also, viscous can be changed
depending on the material temperature and strain rate. The visco-elastic property means that having both
viscosity and elasticity. Similarly, the visco-plastic property means that having both viscosity and plasticity. FEA
NX includes age independent and age dependent model for visco-elastic material, and Soft Soil Creep,
Sekiguchi-Ohta(viscid) model for visco-plastic material.

Table 4.5.1 Available viscous

materials for each element type
Element type

Axisymmetric Solid
Plane Stress

Plane strain
Material type
Interface

Geogrid
Beam
Truss

Shell

Solid
Age independent ∨ ∨ ∨ ∨ ∨ ∨

Age dependent ∨ ∨ ∨ ∨ ∨ ∨ ∨

Soft Soil Creep ∨ ∨ ∨

Sekiguchi-Ohta
∨ ∨ ∨
(Viscid)

At the macroscopic level, the creep phenomenon is best observed in the uniaxial creep test under constant load
5.1 and the relaxation test under constant strain at constant temperature. A specimen subjected to a constant
Age Independent Visco- uniaxial tension exhibits three distinct phases in the time frame: primary creep stage, secondary creep stage and
elastic Material the tertiary creep stage to rupture as shown in Figure 4.5.1. In the first stage of so-called primary creep, we
observe a decreasing strain rate. In the second stage of so-called secondary creep, the creep strain rate is
approximately constant. In the third stage of tertiary creep, the creep strain tare increases. The tertiary creep,
similar to necking in plasticity, is considered as a localized instability phenomenon, which is beyond the scope
of this creep analysis. The primary and secondary creep behavior can be used for isotropic material in FEA NX.

Section 5. Viscous Material Properties | 237

238 | Section 5. Viscous Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Figure 4.5.1 Uniaxial creep test 

Rupture
·
under constant load at constant
temperature

Primary Secondary Tertiary

If the specimen is unloaded after some creep deformation, the elastic strain is immediately recovered and a
portion of the creep strain is gradually recovered as shown in Figure 4.5.2. The recoverable portion of the creep
deformation is called primary creep and the non-recoverable portion is called secondary creep.

Figure 4.5.2 Creep strain 

relaxation upon load removal Load
removed

Elastic
recovery
Secondary
Primary not recoverable
recovery

The Kelvin-Maxwell model is employed in the formulation of the creep capability as a generalization of the age
independent visco-elastic material behavior. This model consists of one elastic spring and two viscous dampers.
The Kelvin-Voigt model, which is a spring and a damper connected in parallel, represents the primary creep and
a damper connected in series to the Kelvin-Voigt model represents the secondary creep.
 ANALYSIS REFERENCE Chapter 4. Materials

Figure 4.5.3 Kelvin-Maxwell creep

model
Primary Secondary
Creep k1 Creep k2
k p ( ) e2 or  cprimary e1 or  total
c

cs ( )

· c p ( ) · ·  (t ) or s1

The creep strain and time-dependent increase in creep strain under constant stress is given as


 total
c
 t   primary
c

cs
(4.5.1)
  k 
 cprimary  1  e p cp t

k p  

The creep strain can be calculated with two empirical laws as follows:

Empirical law 1 :  total

c
 A( ) 1  exp   R( )t    K ( )t
A( )  a b or a exp  b 
R( )  c exp  d  or c d (4.5.2)
K ( )  e sinh  f    e exp  f  
g
or
Empirical law 2 :  total
c
 a bt d  e f t
a, b, c, d , e, f : Material constants
t : Time

  
In empirical law1, k p  , cp  , cs  and k p , c p , cs are calculated by the primary and
A( ) A( ) R( ) K ( )
secondary differential equations for  total
c
in empirical law2.

The equilibrium equation of Kelvin-Maxwell creep model in the uniaxial condition is as follows:

Ce  ke  s
 cs  cs  0 0  (4.5.3)
C , k    , s  s1 0
T

 sc ( c p  c )
s  0 k p 

Section 5. Viscous Material Properties | 239

240 | Section 5. Viscous Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Substitution of the increase in creep strain using central difference method into equation (4.5.3) gives the
following equation:

2 
 t C  k  e  2Ce  s (4.5.4)
 

The stiffness of the primary, secondary creep elements and equivalent creep stiffness in Kelvin-Maxwell creep
model can be determined by

2c p 2cs kk
k1  k p  , k2  , kc  1 2 (4.5.5)
t t k1  k2

Using equation (4.5.4) and (4.5.5), the pseudo incremental strain (   ' ) which represent the stress relaxation is
as follows:

s ' c c 
 
cp c
 '   2  s  total   primary
c
  primary  (4.5.6)
kc  2
k k1 

In multi-axial creep deformation, a unique set of rheological parameters ( k p , c p , cs ) based on the effective
stress is used and the pseudo incremental strain may be expressed with the following equation:

2 2c p 2  2cs 
k1   kp  , k2   
3  t
 3  t 
(4.5.7)
4 c c 
 
cp
Δε'   s εtotal  ε cprimary  ε cprimary 
3  k2 k1 

With the total strain increment which is the summation of elastic and creep strain increment, the stress-strain
relationship gives the following equation:

σ   De  Dc   ε e  ε c  (4.5.8)

De : Material matrices for elasticity

Dc : Material tangent matrices for creep

Since the summation of elastic and creep strain increment should be equal to the exception of the pseudo
incremental strain from the total strain increment, the stress-strain relationship is as follows:
 ANALYSIS REFERENCE Chapter 4. Materials

σ  De ε  ε ' (4.5.9)

For isotropic material, the elastic-creep tangent matrix Dec may be conveniently obtained by

 2 1 1 
 K  3 kec K  kec
3
K  kec
3
0 0 0 
 
 2 1 
 K  kec K  kec 0 0 0 
3 3
 
 2
K  kec 0 0 0 
 3 
D 
ec

 1
kec 0 0 
  (4.5.10)
2
 
 1
symmetric kec 0 
 2 
 1 
 kec 
 2 
1 1 1
 
kec 2G kc

K : Bulk modulus
G : Shear modulus

The properties of material such as concrete are changed with time and non-mechanical deformation of creep
5.2 and shrinkage occurs. Also, the deformation with time varies depending on the time of the stress occurred in
Age Dependent Visco- creep deformation.
elastic Material
When a uniaxial stress exerts on a concrete specimen at the age  , creep deformation with time can be
expressed as creep compliance (total strain at the age t ), specific creep (creep function excluding elastic
deformation) and creep coefficient (ratio of creep strain to elastic strain). Various creep functions can be used
depending on the time of the specific stress applied. If the stress changes with time, the increased/decreased
stress at each time requires an independent creep function. Creep strain at a particular time is calculated
through superposition of individually calculated strains due to the stresses increased/decreased from the time
that stress starts changing. In order to use the superposition method, the histories of all the element stresses
are saved, and the creep strains are calculated from the initial steps to the present for all the stresses at every
step. Extensive data storage and calculations are thus required to use the superposition method. However, FEA
NX does not save the entire histories of stresses, rather uses the following integration method to increase the
calculation efficiency.

The total creep deformation from a particular time to a final time can be expressed as a superposition integration
of creeps due to the stresses resulting from each stage.

Section 5. Viscous Material Properties | 241

242 | Section 5. Viscous Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

t  ( )
 c (t )   C (t , ) d (4.5.11)
0 
 c (t ) : Creep strain at time t
C (t , ) : Specific creep
 : Time at which the load is applied

If we assume from the above expression that the stress at each stage is constant, the total creep strain can be
expressed as a sum of the strains at each stage.

n1
 cn    j C (tn , j ) (4.5.12)
j 1

Using the equation (4.5.12), the incremental creep strain between the times tn  tn1 can be expressed as
follows:

n1 n2
 cn   cn   cn1    j C (tn , j )    j C (tn1 , j ) (4.5.13)
j 1 j 1

If the specific creep is expressed by the degenerate kernel(Dirichlet functional summation), the incremental
creep strain can be calculated without having to save the entire stress history.

m
C (t , )   ai (t ) 1  e (t  ) i  (4.5.14)
i 1

a (t ) : Coefficients related to the initial shapes of specific creep curves at the time of loading 
 : Values related to the shapes of specific creep curves over a period of time

In FEA NX, you can use the Aging-Kelvin creep model using the five  and the Aging-Viscous creep model
which excludes spring from the Aging-Kelvin creep model.

Figure 4.5.4 Aging-Kelvin creep k1 k2 k3 k4 k5

model

· 

1 2 3 4 5
 ANALYSIS REFERENCE Chapter 4. Materials

By introducing the specific creep formula, the incremental strain can be expressed as follows:

1 5
 5
      ai (t ) 1  i     1   i  cn1
 E i 1  i1
i  (1  i ) i t (4.5.15)
i  e t / i

 cn1 : Creep strain of previous stage

E : Elastic modulus

5 5
1 1
From the above expression, it can be rearranged using    ai (t ) 1  i  and  "   1  i  cn1 to
E E i1 i 1

give the following equation:

  E (   " ) (4.5.16)

Finally, including the shrinkage strain, it can be expressed as follows:

  E(   "   sh ) (4.5.17)

The Soft Soil Creep model simulates the creep behavior expanded to three-dimensional based on one-
5.3 dimensional creep theory28,29,30.
Soft Soil Creep
Unlike the primary consolidation by the dissipation of excessive pore pressure, the secondary consolidation is a
phenomenon caused by changes in the clay structure skeleton. It has the time dependency behavior what the
compression is occurring continuously over time. Therefore, the Soft Soil Creep model is suitable for
representing the creep behavior with time dependency

28 Buisman, K., Results from long duration settlement tests., Proc. 1 st International Conference on Soil Mechanics and
Foundation Engineering, Cambridge, 1936, Vol. 1, p. 103-107.
29 Bjerrum, L., Engineering geology of Norwegian normallyconsolidated marine clays as related to settlements of buildings
(Seventh Rankine Lecture) Geotechnique, 1967, Vol. 17, p. 83-118.
30 Garlanger, J.E., The consolidation of soils exhibiting creep under constant effective stress. Geotechnique, 1972, Vol. 22,
p. 71-78.

Section 5. Viscous Material Properties | 243

244 | Section 5. Viscous Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Figure 4.5.5 Primary and log  time 

secondary consolidation curve
with time

Primary Secondary

strain

Characteristics
The Soft Soil Creep model is to simulate the time-dependent secondary consolidation (creep behavior), the
stress-dependent stiffness and the failure behavior according to the Mohr-Coulomb criterion.

1D creep model and 3D extended model

The 1D creep strain is expressed as the following equation what the total strain rate is the sum of an elastic strain
 e and a time-dependent creep strain  c . The creep strain can be considered to a time-dependent creep strain,
i.e., a visco-plastic strain.

 

   1   
   
e c
  
1  e0  1  e0    p 
(4.5.18)
  1  e0  c 
 p   p 0 exp     
     
e : Elastic strain rate
 c
: Creep strain rate
 : Swelling index
 : Compression index
e0 : Initial void ratio
 : Creep index
 : Reference time, precisely one day31

31 Vermeer, P. A., & Neher, H. P., A soft soil model that accounts for creep, Proceedings of the international symposium
'Beyond 2000 in Computational Geotechnics', 1999, Amsterdam, p.249–261.
ANALYSIS REFERENCE Chapter 4. Materials

p : Pre-consolidation stress

From the above equation of 1D creep behavior, the total strain rate of 3D extended Soft Soil Creep model can
be expressed as follows:

 

11   p eq  
p eq
ε ε ε  D σ
e c
 
1  e0 p,eqp  pp  σ (4.5.19)
 1  e0 c 
p p  p p 0 exp   v 
   
D : Elasticity matrix
p eq : Equivalent pressure
pp : Pre-consolidation pressure

From the above equation (4.5.19), the volumetric creep strain  vc can be expressed with the following equation:

 

1   peq  
 vc    (4.5.20)
1  e0    p p 

If the equation (4.5.20) is integrated over time  t for constant p eq , the volumetric creep strain increment is
given as

 
 
  t  p  
eq 
 vc   ln 1    (4.5.21)
1  e0     p p 0  
 

Equivalent pressure and yield function

Using the well-known stress invariants for isotropic stress p and deviatoric stress q , the equivalent pressure
p eq in Soft Soil Creep model can be defined as follows:

peq   p  p    q2
2
(4.5.22)

 : Cap parameter in Modified Mohr-Coulomb model

 is constant which defined by input parameter or ground material.

Section 5. Viscous Material Properties | 245

246 | Section 5. Viscous Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

The Soft Soil Creep model has the Modified Mohr-Coulomb model criterion without strain hardening to prevent
excessive deviatoric stress. Therefore, the total strain rate (4.5.19) additionally includes the plastic strain rate by
shear failure as follows:
 e  p c (4.5.23)
g
  p


(4.5.24)
6sin
g q  p  p 
3  sin
g : Plastic potential function
 : Dilatancy angle

The plastic strain rate follows the flow rule of Modified Mohr-Coulomb model and for fine grained, cohesive soils,
the dilatancy angle tends to be small, it may often be assumed that dilatancy angle is equal to zero.

Figure 4.5.6 Equivalent pressure q

shape and yield function

p eq


6sin 
3  sin 
p
p p eq

The Viscid type of Sekiguchi-Ohta model follows the nonstationary flow surface theory32 among various visco-
5.4 plastic theories for simulating creep behavior of ground. The nonstationary flow surface theory model is
Sekiguchi-Ohta (Viscid) basically based on the plastic model, but there is a difference that it contains the time dependent function.
The Viscid type follows the assumption33 that the creep is in progress with a constant stress state in initial state
ground before loading.
Like Inviscid type, it resolves a numerical problem using the specificity algorithms of Inviscid type since it
contains a stress state that the plastic flow is only undetermined in the yield function. However, except the
specificity problem, convergence problem occurs when the stress state locates dry side due to the model
characteristics. In FEA NX, it resolves convergence problem by correcting the softening behavior in the stress
state of dry side.

32 Liingaard, M., Augustesen, A., Lade, P. V., "Characterization of models for time-dependent behavior of soils",
International Journal of Geomechanics, 2004, 4.3: 157-177.
33 Sekiguchi, H. and Ohta H., "Induced anisotropy and time dependency in clays", 9th ICSMFE, Tokyo, Constitutive
equations of Soils, 1977, 229-238
ANALYSIS REFERENCE Chapter 4. Materials

Yield function
The flow function of Viscid type is defined as follows:

 t  f    vp
F   ln 1  0 exp    0
   v (4.5.25)
   
 : Coefficient of secondary compression
0 : Change speed of initial volumetric strain rate
 vvp : Visco-plastic volumetric strain

The f   of equation (4.5.25) is defined as follows:

 p
f    MD ln    D (4.5.26)
 p0 

The detailed explanation of equation (4.5.26) can be found in ‘2.16. Sekiguchi-Ohta (Inviscid)’.
However, in case of directly using the flow function of equation (4.5.25) for yield function, the problem34 that
elastic range cannot be defined occurs since the left term always has a positive value. In this case, the associated
plastic flow rule cannot be used due to the violation of Hill's principle of maximum plastic work. In order to
resolve this problem, Iizuka and Ohta35 transformed the equation (4.5.25) as follows:

f  , h   f    h   vvp , t   0 (4.5.27)

Here, the hardening function ( h ) is defined as follows:

     vvp  

h   vvp , t    ln  exp    1 (4.5.28)
 0 
v t     


However, there is a problem that the hardening function ( h ) is not defined when the initial visco-plastic
volumetric strain rate is zero (  vvp0  0 ). Therefore, the numerical problem is resolved by setting the initial value

34 Takeyama, T., Pipatpongsa, T., Iizuka, A., Mizuta, T., Ohno, S., Ohta, H., "Soil/water coupled FE Simulation of field
performance of 5 embankments placed on homogeneous clay." Proceedings of the Sri Lankan Geotechnical Society's First
International Conference on Soil & Rock Engineering, 2007
35 Iizuka, A., Ohta, H., "An interpretation of Sekiguchi and Ohta's model based on viscoplasticity theory.", Proceedings of
the 34th Japanese National Conference on Geotechnical Engineering, 1999, 595-596

Section 5. Viscous Material Properties | 247

248 | Section 5. Viscous Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

of the visco-plastic volumetric strain rate (  vvp ) which makes h   vvp , t1   0 at t1 when the stress state is
judged to the visco-plastic state violating the initial yield function with load change or overtime.

 v0t1 
 vvp0   ln   1 (4.5.29)
 

Figure (4.5.7) shows the yield function at triaxial stress state. Like the Cam-Clay material model, it is called that
the right is wet side and the left is dry side based on the critical state line. Generally, the material model which
follows critical state theory shows the hardening behavior at wet side and the softening behavior at dry side.
However, in the viscid type of Sekiguchi-Ohta model, it is known that the over-consolidation ratio is high and
the convergence problem36 occurs when the stress state locates dry side due to the nature of yield function. In
FEA NX, the convergence problem is resolved by preventing the softening behavior at dry side.

Figure 4.5.7 Simplified yield C.S.L

function and critical state line in q
triaxial stress state
Dry side Wet side

yield function

36 Takeyama, T., Ohno, S., Pipatpongsa, T., Iizuka, A., Ohta, H., "The stress update using implicit integration for the viscid
version of sekiguchi-ohta model", Technical report, 2005
 ANALYSIS REFERENCE Chapter 4. Materials

Section 6 Hysteresis Material Properties

When the crack and yield occurs by irregular cyclic load, the displacement history to the current affects the later
relationship between restoring force and displacement. The relationship between force and deformation for
uniaxial load is called the Skeleton curve. When the cyclic load is applied based on the Skeleton curve, the rule
of the relationship between force and deformation for unloading and reloading is called hysteresis model. The
properties of each hysteresis model are explained in this chapter. Table 4.6.1 lists the available hysteresis models
for each element.

Table 4.6.1 Available hysteresis

models for each element type
Element type

Axisymmetric Solid
Plane Stress

Plane strain
Elastic Link

Hysteresis model
Interface

Geogrid
Beam
Truss

Shell

Solid
Multilinear ∨ ∨ ∨

Normal Bilinear ∨ ∨ ∨

Kinematic ∨ ∨ ∨

Origin-Oriented ∨ ∨ ∨

Peak-Oriented ∨ ∨ ∨

Clough ∨ ∨ ∨

Degrading ∨ ∨ ∨

Takeda ∨ ∨ ∨

Modified Takeda ∨ ∨ ∨

Modified Ramberg
∨ ∨ ∨ ∨ ∨ ∨ ∨ ∨
Osgood
Modified Hardin-
∨ ∨ ∨ ∨ ∨ ∨ ∨ ∨
Drnevich

Section 5. Viscous Material Properties | 249

250 | Section 7. Thermal Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Response points at initial loading move about on a bilinear skeleton curve.

6.1
Normal Bilinear model P

Figure 4.6.1 Normal Bilinear

model

K1

K0

K0
D1

D1 D
K0

K 2

Hysteresis rule
► In case of Dmax  D1 , it is a linear elastic and moves on a elastic gradient straight line crossing the origin.
► In case of the deformation D first overpass D1 , or the maximum deformation point to the current, it moves
on a second gradient straight line.
► In case of unloading in the condition of D1  D , D1  D , it moves on a second gradient straight line by
unloading a elastic gradient according to the Masing rule.
 ANALYSIS REFERENCE Chapter 4. Materials

Response points at initial loading move about on a trilinear skeleton curve. The unloading stiffness is identical
6.2 to the elastic stiffness, and stiffness reduction after yielding is possible only for positive (+) and negative (-)
Kinematic model symmetry. In case of kinematic model, the interaction of axial force and biaxial bending component can be
considered by the plastic theory.

Figure 4.6.2 Kinematic model P

K2

K1

K0
K0

D2 D1 D1 D2 D

K0

K2

Uniaxial hysteresis rule

► In case of Dmax  D2 , it behaves like a bilinear.
► In case of Dmax  D2 , it moves on a third gradient straight line.
► In case of unloading, it moves on a elastic gradient according to the Masing rule.

Multi-axial hysteresis rule

It considers the interaction between axial force and biaxial bending component by the kinematic hardening rule
based on the plastic theory. The yield domain is the following equation.

Section 7. Thermal Material Properties | 251

252 | Section 7. Thermal Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE


 M  
y
 P  Pbal , y   

f ( P, M y , M z )    g y  M y , M z 
  P  P  
y
M
 y ,max   max bal , y  


 z
 M    P  Pbal , z   
 

z
  g z  M y , M z     1  0
 M z ,max   Pmax  Pbal , z  

 My 
  (4.6.1)
gy M y, M z    M y ,max 
 
 My   Mz 
    
 M y ,max   M z ,max 

 Mz 
 
gy M y, M z    M z ,max 
 
 My   Mz 
    
 M y ,max   M z ,max 
M max : Maximum bending yield strength
Pbal : Axial force at the balanced failure
Pmax : Axial yield strength
,  , : An exponent related to interaction curve

Figure 4.6.3 Movement of yield Mz

surface and stiffness change

mi 1 conjugate point
mi
oi
oi 1

My
 ANALYSIS REFERENCE Chapter 4. Materials

There are two yield surfaces corresponding to a trilinear skeleton curve and these two yield surfaces follow the
hardening rule of modified Mroz.

Response points at initial loading move about on a trilinear skeleton curve. The response point moves towards
6.3 the origin at the time of unloading. When it reaches the skeleton curve on the opposite side, it moves along the
Origin-Oriented model skeleton curve again.

P
Figure 4.6.4 Origin-Oriented
model

D2 D1 D1 D2 D

Section 7. Thermal Material Properties | 253

254 | Section 7. Thermal Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Response points at initial loading move about on a trilinear skeleton curve. The response point moves towards
6.4 the maximum displacement point on the opposite side at the time of unloading. If the first yielding has not
Peak-Oriented model occurred on the opposite side, it moves towards the first yielding point on the skeleton curve.

Figure 4.6.5 Peak-Oriented model P

D2 D1 D1 D2 D

 ANALYSIS REFERENCE Chapter 4. Materials

Response points at initial loading move about on a bilinear skeleton curve. As the deformation progresses, the
6.5 unloading stiffness gradually becomes reduced. When the loading sign changes at the time of unloading, the
Clough model response point moves towards the maximum displacement point in the region of the progressing direction. If
yielding has not occurred in the region, it moves towards the yielding point on the skeleton curve. Where
unloading reverts to loading without the change of loading signs, the response point moves along the unloading
path. And loading takes place on the skeleton curve as the loading increases.

Figure 4.6.6 Clough model P

 Dmax , Pmax 

P1

K r

D1 D1 D

K r

P1

 Dmin , Pmin 

Hysteresis rule
► In case of unloading in the condition of D1  D , it moves on the gradient of unloading stiffness, K r .


 D 
K r  K o  1   K o
 Dmax 

(4.6.2)

 D 
K  K o  1
r   K o
 Dmin 
Ko : Initial elastic stiffness

D 1 : Yield displacement in the region of the first unloading
Dmax : Maximum displacement in the region of tension

Section 7. Thermal Material Properties | 255

256 | Section 7. Thermal Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Dmin : Maximum displacement in the region of compression

 : Constant for determining unloading stiffness

► If the sign of the load changes in the unloading process, it moves toward the maximum deformation point of
the opposite side. If the opposite side is not yield, D1 is the maximum deformation point.

Response points at initial loading move about on a trilinear skeleton curve. At unloading, the coordinates of the
6.6 load-deformation move to a path along which the maximum deformation on the opposite side can be reached
Degrading model due to the change of unloading stiffness once. If yielding has not occurred on the opposite side, the first yielding
point is assumed to be the point of maximum deformation. As the maximum deformation increases, the
unloading stiffness gradually decreases.

Figure 4.6.7 Degrading model P

D max , Pmax 

P2

P1

D1 D2 D

P2

D min , Pmin 

Hysteresis rule
► In case of unloading in the condition of D2  D , it behaves like a bilinear.
 ANALYSIS REFERENCE Chapter 4. Materials

► In case of unloading in the condition of D2  D , it moves on the gradient of unloading stiffness, K r .

 P  Pmin 1 
K r  K o  max 
 Dmax  Dmin K r1  (4.6.3)
P   P2
K r1  2
D2  D2
Ko : Initial elastic stiffness
Dmax : Maximum displacement in the region of tension
Dmin : Maximum displacement in the region of compression
Pmax : Maximum force in the region of tension
Pmin : Maximum force in the region of compresion

P 2 : Yield stiffness in the region of the second unloading

D 2 : Yield displacement in the region of the second unloading

Response points at initial loading move about on a trilinear skeleton curve. The unloading stiffness is determined
6.7 by the location of the unloading point on the skeleton curve and whether or not the first yielding has occurred
Takeda model in the opposite region.

Figure 4.6.8 Takeda model P

P2

P1

P1

P2

Section 7. Thermal Material Properties | 257

258 | Section 7. Thermal Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Hysteresis rule

Figure 4.6.9 Hysteresis rule of the P

Takeda model after the first
yielding
Rule : 3
P2

P1 Rule : 2

Rule :1

P1
Rule : 4

P2

► In case of the deformation D first overpass D1 , the opposite first yielding is the maximum displacement
point on the opposite side.
► In case of unloading on the skeleton curve, the coordinates of the load-deformation moves toward the
maximum deformation point of the opposite side. (Rule 1)
► In case of reloading before it reaches to the maximum deformation point of the opposite side, the point
progresses along the same unloading curve. (Rule 2)
► In case of reaching to the skeleton curve, it moves along the skeleton curve. (Rule 3)
 ANALYSIS REFERENCE Chapter 4. Materials

Figure 4.6.10 Hysteresis rule of P

the Takeda model after the
second yielding Rule : 8
Rule : 5
P2
Rule : 7

P1

Rule : 6

D 2

D1 D2 D

P1

Rule : 9
Rule :10
P2

D
► In case of the deformation D first overpass 2 , it moves along the skeleton curve. (Rule 5)
► In case of unloading on this curve, it moves on the gradient of unloading stiffness, K r . If the opposite side is
K r P1
before experiencing the first yielding, the range of is the . (Rule 6)

0.4
Dmax
K r  K b
D2
0.4
Dmin
K r  K b
D2
(4.6.4)
 
P P
K b  2

1

D D
2 1

P2  P1
K b 
D2  D1

P1
► If the point exceeds the , it moves toward the second yielding point. (Rule 9)

Section 7. Thermal Material Properties | 259

260 | Section 7. Thermal Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Figure 4.6.11 Hysteresis rule of the P

inner loop Takeda model

P2

P1
Rule :12

Rule :13

D2 D
Rule :11
Rule :14

P2

► In case of unloading on the straight line toward the maximum deformation point of the opposite side, it enters
to the inner loop. (Rule 11)
► In case of the sign of restoring force changes in the process of unloading in the inner loop, it returns to the
previous unloading point of the opposite side. (Rule 12)
 ANALYSIS REFERENCE Chapter 4. Materials

6.8 P
Modified Takeda model

P2
Figure 4.6.12 Modified Takeda
model
P1

P1

P2

Hysteresis rule

Figure 4.6.13 Hysteresis rule of the P

modified Takeda model after the
first yielding
Rule : 3
P2

P1 Rule : 2

Rule :1

D1
D

P1

P2

Section 7. Thermal Material Properties | 261

262 | Section 7. Thermal Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

► In case of the deformation D first overpass D1 , the opposite first yielding is the maximum deformation
point of the opposite side.
► In case of unloading on the skeleton curve, the coordinates of the load-deformation moves toward the
maximum deformation point of the opposite side. (Rule 1)
► In case of reloading before it reaches to the maximum deformation point of the opposite side, the point
progresses along the same unloading curve. (Rule 2)
► In case of reaching to the skeleton curve, it moves along the skeleton curve. (Rule 3)

Figure 4.6.14 Hysteresis rule of P

the modified Takeda model after
the second yielding Rule : 7
Rule : 4
P2
Rule : 6

P1
Rule :10

Rule : 5

Rule :11 D
Rule : 9
Rule : 8

Rule :12
D min , Pmin 

D
► In case of the deformation D first overpass 2 , it moves along the skeleton curve. (Rule 4)
► In case of unloading on this curve, it moves on the gradient of unloading stiffness, K r . If the opposite side is
before experiencing the second yielding, the opposite second yielding is the maximum deformation point of the
opposite side. (Rule 5)
 ANALYSIS REFERENCE Chapter 4. Materials

 D
0.4
  D
0.4

K r  max  Kb , max  , K r  max  Kb , min 
 D1   D1 
   
(4.6.5)
P  Pmin
Kb  max
Dmax  Dmin

► At the time when the sign changes, , the coordinates of the load-deformation moves toward the maximum
deformation point of the opposite side. (Rule 8)
► In case of unloading on the straight line toward the maximum deformation point of the opposite side, it enters
to the inner loop. (Rule 9)
► At the time when the sign changes, the coordinates of the load-deformation moves toward the maximum
deformation point of the opposite side. (Rule 10)

The Ramberg-Osgood model is originally proposed for the dynamic model of metal material, but modified by
6.9 Tatsuoka.
Modified Ramberg-
Osgood model 

Figure 4.6.15 Modified Ramberg-

Osgood model

Go  1 ,  1 
Skeleton Curve

Go

Hysteresis Curve

Section 7. Thermal Material Properties | 263

264 | Section 7. Thermal Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Hysteresis rule
► In the initial loading, it moves along the following skeleton curve.


Go      

2 hmax  2  (4.6.6)
 ,   
2   hmax   r Go 
Go : Initial stiffness (Shear modulus)
r : Reference shear strain
hmax : Maximum damping constant

► The hysteresis curve is as follows:


    1     1      1  
Go    1      (4.6.7)
 2   2   2  

If only shear stress is considered, the initial confinement pressure may be considered with the following equation.

Go  Goi  mo 
' n1

(4.6.8)
 r   ri  mo 
' n2

Goi ,  ri : Normalized value at confinement pressure 1.0tf/m 2

 mo
'
: Initial confinement pressure
n1 , n2 : Influence coefficients of confinement pressure

Formulation is performed according to the additive decomposition, which divides elastic and plastic
components, as shown in the following formula.

  
  
Go Go (4.6.9)
  el   pl

In a multi-axial state, the stresses are divided into hydrostatic stress and deviatoric stress, and then it is
organized into deviatoric stress and deviatoric strain, and expressed in the following formula.
ANALYSIS REFERENCE Chapter 4. Materials

σ dev
2Goε dev  σ dev  2Go ε dev
pl
(4.6.10)
σ dev

σ dev : Deviatoric stress tensor

ε dev : Deviatoric strain tensor
ε pl
dev : Plastic strain tensor

Considering the multi-axial conditions of the plastic strain in equation (4.6.9), the equivalent deviatoric strain
 eq  2εdev : εdev and equivalent deviatoric stress  eq  0.5σdev : σdev are used to indicate the plastic strain.

ε dev  eqpl
pl

 eq   2Go  Go (4.6.11)
σ dev  eq

Using equation (4.6.10) and equation (4.6.11), the relationship between stress and strain in a multi-axial state is
defined as follows:

 eqpl
2Goε dev  σ dev  Go σ
 eq dev (4.6.12)
 σ dev   eq  σ dev

The deviatoric strain and the deviatoric stress tensor of the left and right sides of equation (4.6.12) are replaced
with the equivalent deviatoric strain and equivalent deviatoric stress, and the scalar equation is established as
follows.

Go eq   eq   eq  eq (4.6.13)

Equation (4.6.12) and (4.6.13) are used to indicate the relationship between equivalent deviatoric stress and
equivalent deviatoric strain.

 eq
σ dev  2 ε (4.6.14)
 eq dev

The differential equation for the relationship of the equivalent deviatoric strain and the equivalent deviatoric
stress is as follows.

Go
 eq   eq  Go eq (4.6.15)
1      1  eq 

Section 7. Thermal Material Properties | 265

266 | Section 7. Thermal Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

The following formula is used to define the relationship between stress and strain.
ε dev
 eq  2 ε dev
J 2e
 1 
ε dev   I  δδ  ε
 3 
      
σ dev  2 eq ε dev   eq  eq  ε dev  (4.6.16)
 eq     eq 
 eq 
p   Kδε
σ  σ dev  pδ

Using equation (4.6.15) and (4.6.16), the material stiffness matrix is calculated as follows:

  1      2 G 1  
σ   2 eq I   3K  2 eq  δδ  ε   2 eq  o   ε devε dev  ε (4.6.17)
       
 eq 3  eq    eq eq  eq eq  

This model defines the hysteresis curve by applying the Masing rule to the Hardin-Drnevich model which
6.10 suggested only the skeleton curve.
Modified Hardin-Drnevich
model 

Figure 4.6.16 Modified Hardin-

Drnevich model

Go  1 ,  1 
Skeleton Curve

Go

Hysteresis Curve
ANALYSIS REFERENCE Chapter 4. Materials

Hysteresis rule
► In the initial loading, it moves along the following skeleton curve.

Go
 (4.6.18)

1
r
Go : Initial stiffness (Shear modulus)
r : Reference shear strain

► The hysteresis curve is as follows:

Go    1  (4.6.19)
  1 
  1
1
2 r

If only shear stress is considered, the initial confinement pressure may be considered with the following equation.

Go  Goi  mo 
' n1

(4.6.20)
 r   ri  mo 
' n2

Goi ,  ri : Normalized value at confinement pressure 1.0tf/m 2

 mo
'
: Initial confinement pressure
n1 , n2 : Influence coefficients of confinement pressure

Formulation is performed according to the additive decomposition, which divides elastic and plastic
components, as shown in the following formula.

  
 
Go  r Go (4.6.21)
  el   pl

Considering the multi-axial conditions of the plastic strain in equation (4.6.20), the equivalent deviatoric strain
 eq  2εdev : εdev and equivalent deviatoric stress  eq  0.5σdev : σdev are used to indicate the plastic strain.

Section 7. Thermal Material Properties | 267

268 | Section 7. Thermal Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

 eqpl ε dev  pl
pl

 2Go  Go eq (4.6.22)
r σ dev  eq

Subsequent relationships are the same as the Modified Ramberg-Osgood model described earlier, and the
differential equation of equivalent deviatoric strain and equivalent deviatoric stress is as follows.

 r2
 eq  Go  eq  Go eq (4.6.23)
 r 
2
eq

The material stiffness matrix is calculated as follows:

  1      2 G 1  
σ   2 eq I   3K  2 eq  δδ  ε   2 eq  o   ε devε dev  ε (4.6.24)
       
 eq 3  eq    eq eq  eq eq  
ANALYSIS REFERENCE Chapter 4. Materials

Section 7 Thermal Material Properties

The material used for the conduction element in FEA NX is isotropic.
7.1 The heat flux-temperature gradient relationship of the isotropic thermally conductive material is as follows.
Constitutive equation
 fx  k 0 0   g x 
    
 y
f   0 k 0  g y  (4.7.1)
   0 0 k   g 
 fz   z

In the case of a porous material such as a ground, it is composed not only of the soil particles but also water and
7.2 gas in the pores between the soil particles. Therefore, unlike a general isotropic material, the thermal
Porous material conductivity coefficient k considering water and gas is as follows.

k  1  n ks  nSkw  n 1  S  kv (4.7.2)

ks , kw , kv : Heat conductivity of soil, water and gas

n : Porosity
S : Saturation

The heat capacity ( C ) in the porous medium can also be expressed as follows.

C  1  n sCs  nS wCw  n 1  S  vCv (4.7.3)

s ,  w , v : Mass Density of soil, water and gas

Cs , Cw , Cv : Heat capacity of soil, water and gas

Section 7. Thermal Material Properties | 269

270 | Section 7. Thermal Material Properties

Chapter 4. Materials
ANALYSIS REFERENCE

Below zero degrees, water changes phase to ice. In the case of the ground, the phase change of the groundwater
7.3 in the air gap between the soil particles occurs, not the phase change of the soil particles themselves. At this
Freezing of Porous time, additional energy must be supplied, which is called latent heat. Mentioned energy depend as well on the
change of the unfrozen water content with respect to temperature. Unfrozen water content function depicts
Material the rate at which water does not change into ice even at temperatures, generally below zero degrees.

Therefore, considering the unfrozen water content function ( fu ) and the latent heat of water ( H L ), the
thermal conductivity and thermal capacity of the ground can be expressed by the following equation.

 
k  1  n  ks  nS f u kw  1  f u  ki  n 1  S  kv (4.7.4)

 dfu 
C  1  n  sCs  nS w  Cw  H L   n 1  S  vCv (4.7.5)
 dT 

ki : Thermal Conductivity of Ice