3 FailureCriteria ConstitutiveModels
3 FailureCriteria ConstitutiveModels
3 FailureCriteria ConstitutiveModels
S cross-section is horizontal
but has small deviations up
and down so that it passes
only by particle contacts and
does not cut any particle.
hw
u = γw . hw
Pore water pressure
N.F.
hw
• Pore water pressure: can be defined as the pressure exerted by the pore
water
𝝈′ = 𝝈 − 𝒖
❖ Effective stress controls the mechanical
behaviour of the soil!
1. AT REST STRESS STATE
Definition of the initial stress state -
Example:
γ = 20 kN/m3
γw = 10 kN/m3
𝜎𝑣0 = K0 =0.5
K0 – coefficient of
𝑢0 = earth pressure at rest
′
𝜎𝑣0 =
′ Z = 10 m
𝜎ℎ0 =
𝜎ℎ0 =
1. AT REST STRESS STATE
Definition of the initial stress state -
Example:
Δu Tresca envelope
Δσ = 𝐷 . Δε
Constitutive matrix
Elastic behaviour: all strains are recovered when the load is removed. If the relation is linear,
the material is linearly elastic. If not, the material is nonlinearly elastic.
Linear Nonlinear
elastic elastic
Elastic perfectly plastic behaviour: up to a given value (yield stress), the material exhibits elastic
behaviour. After that, the material presents purely plastic deformations under a constant stress equal
to the yield stress. The elastic range can be linear or nonlinear.
Linear Nonlinear
elastic elastic
3. CONSTITUTIVE LAWS
Stress-strain behaviour
Definition: soils and rocks, as other materials, have a finite strength. There is, therefore, a
limit that divides the stress states compatible with the resistance of soils and rocks from
those that they cannot resist. This boundary is called the FAILURE SURFACE and its
equation corresponds to the so-called FAILURE CRITERION:
𝑞
Impossible stress 𝐹=𝑓 𝜎 −𝑘 =0
states (F>0)
Stress Stress variable
Failure → Yield surface that depends
state
(F=0) on hardening
parameter
𝑓 𝜃
Possible
stress states
(F<0)
Elastic strains Plastic strains
𝑝
4. FAILURE CRITERION
Stress state
As mentioned before, the definition of the stress state at a point:
𝜎𝑥
𝜎𝑦
𝜎𝑥 𝜏𝑥𝑦 𝜏𝑥𝑧 𝜏𝑥𝑦 = 𝜏𝑦𝑥 𝜎𝑧
𝜏𝑦𝑧 = 𝜏𝑧𝑦 𝜎 = 𝜏
𝜎 = 𝜏𝑦𝑥 𝜎𝑦 𝜏𝑦𝑧 𝑥𝑦
𝜏𝑧𝑥 𝜏𝑧𝑦 𝜎𝑧 𝜏𝑥𝑧 = 𝜏𝑧𝑥 𝜏𝑦𝑧
𝜏𝑥𝑧
Normal and
Equilibrium of
shear stress
moments components
4. FAILURE CRITERION
Stress state
𝜎𝑥
𝜎𝑦
The magnitude of the 6 stress components necessary 𝜎𝑧
𝜎 = 𝜏
to define the stress state depends on the selected 𝑥𝑦
reference system 𝜏𝑦𝑧
𝜏𝑥𝑧
The alternative is to define other parameters STRESS
independent of the reference system INVARIANTS
Some options:
1. Principal stresses and their directions
2. First invariant of the stress tensor (I1), second invariant of the deviatoric
tensor (J2) and the angle of the deviatoric plane (θ)
3. Mean stress (p), deviatoric stress (q) and angle of the deviatoric plane (θ)
4. FAILURE CRITERION
1. Principal stresses and their directions
2. First invariant of the stress tensor (I1), second invariant of the deviatoric
tensor (J2) and the angle of the deviatoric plane (θ)
3. Mean stress (p), deviatoric stress (q) and angle of the deviatoric plane (θ)
For isotropic materials (materials that have the same properties in all directions)
𝜎1 0 0
only 3 parameters are required (𝜎1 , 𝜎2 𝑎𝑛𝑑 𝜎3 ) since
𝜎 = 0 𝜎2 0
the response of the material is the same in all
directions
0 0 𝜎3
The magnitude of the principal stresses can be calculated solving the characteristic
equation: −𝜎 3 + 𝐼 𝜎 2 − 𝐼 𝜎 + 𝐼 = 0 (𝐼 , 𝐼 and 𝐼 : invariants of the stress tensor)
1 2 3 1 2 3
𝐼1 = 𝜎𝑥 + 𝜎𝑦 + 𝜎𝑧
𝜎𝑥 𝜏𝑥𝑦 𝜎𝑥 𝜏𝑥𝑧 𝜎𝑦 𝜏𝑦𝑧 2 2 2
𝐼2 = 𝜏 𝜎𝑦 + 𝜏𝑥𝑧 𝜎𝑧 + 𝜏𝑦𝑧 𝜎𝑧 = 𝜎𝑥 𝜎𝑦 + 𝜎𝑥 𝜎𝑧 + 𝜎𝑦 𝜎𝑧 − 𝜏𝑥𝑦 − 𝜏𝑥𝑧 − 𝜏𝑦𝑧
𝑥𝑦
𝜎𝑥 𝜏𝑥𝑦 𝜏𝑥𝑧
𝐼3 = 𝜏𝑥𝑦 𝜎𝑦 𝜏𝑦𝑧 = 𝜎𝑥 𝜎𝑦 𝜎𝑧 + 2𝜏𝑥𝑦 𝜏𝑥𝑧 𝜏𝑦𝑧 − 𝜎𝑥 𝜏𝑦𝑧
2 − 𝜎 𝜏2 − 𝜎 𝜏2
𝑦 𝑥𝑧 𝑧 𝑥𝑦
𝜏𝑥𝑧 𝜏𝑦𝑧 𝜎𝑧
4. FAILURE CRITERION
1. Principal stresses and their directions
2. First invariant of the stress tensor (I1), second invariant of the
deviatoric tensor (J2) and the angle of the deviatoric plane (θ)
3. Mean stress (p), deviatoric stress (q) and angle of the deviatoric plane (θ)
𝜎2 Hydrostatic axis
𝜎1 = 𝜎2 = 𝜎3
P
3
𝐼
3 1
𝜎1
4. FAILURE CRITERION
1. Principal stresses and their directions
2. First invariant of the stress tensor (I1), second invariant of the
deviatoric tensor (J2) and the angle of the deviatoric plane (θ)
3. Mean stress (p), deviatoric stress (q) and angle of the deviatoric plane (θ)
1 2 2
Related to the
𝐽2 = 𝜎𝑥 − 𝜎𝑦 + 𝜎𝑦 − 𝜎𝑧 + 𝜎𝑧 − 𝜎𝑥 2 2
+ 𝜏𝑥𝑦 + 2
𝜏𝑦𝑧 + 2
𝜏𝑥𝑧 radial distance
6 between the
𝜎2 isotropic axis and
𝜃 angle of the deviatoric plane that defines the the stress point
position of the stress point
𝜃 = 30°
𝜃 = 0°
𝜃
2𝐽2
P
𝜎1
𝜎3 𝜃 = −30°
Deviatoric plane
4. FAILURE CRITERION
1. Principal stresses and their directions
2. First invariant of the stress tensor (I1), second invariant of the deviatoric
tensor (J2) and the angle of the deviatoric plane (θ)
3. Mean stress (p), deviatoric stress (q) and angle of the deviatoric plane
(θ)
1 1 𝐼1
𝑝 = (𝜎1 + 𝜎2 + 𝜎3 ) = 𝜎𝑥 + 𝜎𝑦 + 𝜎𝑧 =
3 3 3
1 2 2 2
𝑞= 𝜎 − 𝜎2 + 𝜎2 − 𝜎3 + 𝜎3 − 𝜎1 =
2 1
1 2 2 2 2 2 2
𝜎 − 𝜎𝑦 + 𝜎𝑦 − 𝜎𝑧 + 𝜎𝑧 − 𝜎𝑥 + 6 𝜏𝑥𝑦 + 𝜏𝑦𝑧 + 𝜏𝑧𝑥 = 3𝐽2
2 𝑥
1
In the triaxial plane, p= 3 (𝜎1 +2. 𝜎3 )
there are two equal 𝜎2 = 𝜎3
principal stresses q= 𝜎1 − 𝜎3
4. FAILURE CRITERION
1. Principal stresses and their directions
2. First invariant of the stress tensor (I1), second invariant of the deviatoric
tensor (J2) and the angle of the deviatoric plane (θ)
3. Mean stress (p), deviatoric stress (q) and angle of the deviatoric plane
(θ)
p is related to the distance measured in the isotropic axis between origin and the deviatoric
plane that contants the stress point
q is related to the radial distance between the isotropic axis and the stress point
𝜃 is the angle of the deviatoric plane that defines the position of the stress point
𝜎2
𝜎2 Hydrostatic axis
𝜎1 = 𝜎2 = 𝜎3
𝜃 = 30°
P
𝜃 = 0°
3p 𝜃
2
P
q
𝜎1 3 𝜎1
𝜎3 𝜃 = −30°
4. FAILURE CRITERION
Different failure criterion:
➢ Tresca
➢ Von Mises
➢ Mohr-Coulomb
➢ Drucker-Prager
➢ Gudheus
➢ Hoek-Brown
4. FAILURE CRITERION
Different failure criterion:
➢ Tresca
➢ Von Mises
➢ Mohr-Coulomb
➢ Drucker-Prager
➢ Gudheus
➢ Hoek-Brown
4. FAILURE CRITERION
Tresca failure criterion:
• The Tresca model is independent of I1, so it can only simulate the behaviour of materials whose
strength is independent of the mean stress without permanent volumetric strains
e.g., saturated clays in undrained conditions
• is expressed in terms of total stresses and apply to undrained soil behaviour conditions.
• the failure at a given point occurs whenever the maximum shear stress reaches a given limit
value
• The failure criterion relates the undrained strength, cu, to the radius of the Mohr's circle at failure.
𝐹= 𝐽2 . cos 𝜃 − 𝑐𝑢 = 0
1
𝐹= 𝑞. cos 𝜃 − 𝑐𝑢 = 0
3
4. FAILURE CRITERION
Mohr-Coulomb failure criterion:
• The Mohr-Coulomb failure criterion is widely used in soils and rocks and establishes that the
shear stress that causes failure at a given point varies linearly with the normal stress in the
failure plane (𝑓 – failure): ′
𝜏𝑓 = 𝑐′ + 𝜎𝑓 tan ∅′
• In this model, failure occurs for stress states where the Mohr circle is tangent to the strength envelope
• This strength envelope is typically defined in terms of two parameters: effective cohesion , c’,
and the angle of shearing resistance (effective internal friction angle), 𝝓’.
• The Mohr-Coulomb failure criterion is widely used in soils and rocks and establishes that the
shear stress that causes failure at a given point varies linearly with the normal stress in the
failure plane (𝑓 – failure): ′
𝜏𝑓 = 𝑐′ + 𝜎𝑓 tan ∅′
In general,
▪ Granular soils
▪ Normally consolidated plastic soils c’ = 0
▪ Overconsolidated plastic soils
▪ Residual soils (due to their cemented structure) c’ ≠ 0
▪ Artificially cemented materials
4. FAILURE CRITERION
Mohr-Coulomb failure criterion:
https://soilmodels.com/soilanim/
4. FAILURE CRITERION
Mohr-Coulomb failure criterion:
https://soilmodels.com/soilanim/
5. CONSTITUTIVE MODELS
➢ Linear elastic model
➢ Linear elastic perfectly plastic model
➢ Hyperbolic models
➢ Cambridge models (Cam Clay, modified Cam Clay, p-q-θ)
➢ Soft soil and soft soil creep
➢ Hoek-Brown
The choice of a suitable constitutive model that represents the behavior of soil should take into
account two key factors:
• The model to be adopted should be able to simulate the soil behavior in a satisfactory way;
• The model should be relatively simple, so as not to become prohibitive given the number of
parameters it involves or the difficulty of obtaining them through the usual laboratory and in-situ
tests.
• The linear elastic model is based on Hooke’s law, establishing a linear relationship between stresses
and strains.
• Although, in general, this constitutive model is not suitable for the modelling of soil behaviour, it
is often used in the modelling of high-stiffness elements introduced into the soil, such as
structural elements or intact rock masses.
5. CONSTITUTIVE MODELS
➢ Linear elastic perfectly plastic model
• The linear elastic perfectly plastic model (also known as Mohr-Coulomb model) can be used as a
first approximation of soil behaviour. It is used to model soil behaviour in general.
• The linear elastic part of the linear elastic perfectly plastic model is based on Hooke’s law of
isotropic elasticity
• The perfectly plastic part is based on the Mohr-Coulomb failure criterion (or Tresca).
• Perfectly plastic model is a constitutive model with a fixed yield surface, i.e. a yield surface that
is fully defined by model parameters and not affected by (plastic) straining.
5. CONSTITUTIVE MODELS
➢ Linear elastic perfectly plastic model
Advantages:
• A well-known model that offers first approximation of soil behaviour
• Requires parameters which are generally familiar to most geotechnical engineers and which
can be obtained from basic tests on soil samples
• Soil failure is generally well captured (at least for drained conditions)
Main limitations:
• Failure and yield stresses are coincident;
• Limited ability to model deformation behaviour before failure, especially when stress
level changes significantly or multiple different stress paths are followed;
• A constant Young's modulus is considered. The definition of the value to adopt must be
made with some prudence:
▪ For materials with a large linear elastic range it is realistic to use: 𝐸0
▪ For loading of soils one generally uses: 𝐸50
▪ Considering unloading problems, as in the case of tunneling and
excavations, one needs an unload-reload modulus: 𝐸𝑢𝑟
5. CONSTITUTIVE MODELS
➢ Hyperbolic models
More
realistic !
▪ Elasto-plastic model whose yield surface
is not fixed but expands generating
plastic strains
Examples:
- Hardening Soil Model (PLAXIS ®)
- Hardening Soil Model Small Strains (PLAXIS ®)
5. CONSTITUTIVE MODELS
➢ Hyperbolic models
Parameters:
𝑟𝑒𝑓
𝐸50 Secant stiffness obtained from drained triaxial tests
𝑟𝑒𝑓
𝐸𝑜𝑒𝑑 Secant stiffness obtained from oedometric drained tests
𝑟𝑒𝑓
𝐸𝑢𝑟 Stiffness in unloading – reloading
𝑐 ′ , 𝜙 ′ , ψ or 𝑐𝑢 Strength parameters
5. CONSTITUTIVE MODELS
➢ Hyperbolic models
Limitations:
❖ First Cam Clay model and then other related models: Modified Cam Clay and p-q-θ
❖ Models based on critical state soil mechanics
❖ It was developed for clays where the volumetric strain is more importante and shear strain
❖ Elasto-plastic models with strain hardening
5. CONSTITUTIVE MODELS
➢ Soft soil and Soft soil creep
❖ Models conveniente to simulate the behaviour of soft soils which are very compressible
❖ Soft soil creep : useful for soft soils where secondary consolidation is very important
❖ Available in PLAXIS ®
𝑸𝒖𝒍𝒕
• The bearing capacity is given by 𝒒𝒖𝒍𝒕 =
𝑩𝑳
• In compact granular soils and in clayey soils loaded in undrained conditions, the final
branch is short and practically vertical, failure occurring suddenly and with heave of
the foundation surrounding soil
• In loose soils and in clayey soils loaded in drained conditions, the final branch is
quite inclined and at an angle with the vertical, with the occurrence of significant
settlement without general soil failure, due to the fact that the soil strength keeps
increasing with settlement, failure being thereby progressively postponed
• In wedge I the angle y is p/4+f/2, while in wedge III the angle y’ is p/4-f/2
Displacement vectors
• In an effective stress analysis, the general expression for load bearing capacity is
written as
𝟏
𝒒𝒖𝒍𝒕 = 𝒄′𝑵𝒄 + 𝒒′𝑵𝒒 + 𝑩𝜸𝑵𝜸
𝟐
• In this expression:
• 𝒄′ is the cohesion in terms of effective stress
• 𝑵𝒄 , 𝑵𝒒 and 𝑵𝜸 are functions of 𝝓′
• 𝒒′ represents 𝜎𝑣′ at the level of the footing base
• 𝜸 should be replaced by 𝜸’ in case the water level is at or above the base of the
footing,
while in other cases an intermediate value between 𝜸 and 𝜸’ should be
estimated using the Meyerhof plot