Alternative stress measures explained

In continuum mechanics, the most commonly used measure of stress is the Cauchy stress tensor, often called simply the stress tensor or "true stress". However, several alternative measures of stress can be defined:[1] [2] [3]

  1. The Kirchhoff stress (

\boldsymbol{\tau}

).
  1. The nominal stress (

\boldsymbol{N}

).
  1. The Piola–Kirchhoff stress tensors
    1. The first Piola–Kirchhoff stress (

\boldsymbol{P}

). This stress tensor is the transpose of the nominal stress (

\boldsymbol{P}=\boldsymbol{N}T

).
    1. The second Piola–Kirchhoff stress or PK2 stress (

\boldsymbol{S}

).
  1. The Biot stress (

\boldsymbol{T}

)

Definitions

Consider the situation shown in the following figure. The following definitions use the notations shown in the figure.

In the reference configuration

\Omega0

, the outward normal to a surface element

d\Gamma0

is

N\equivn0

and the traction acting on that surface (assuming it deforms like a generic vector belonging to the deformation) is

t0

leading to a force vector

df0

. In the deformed configuration

\Omega

, the surface element changes to

d\Gamma

with outward normal

n

and traction vector

t

leading to a force

df

. Note that this surface can either be a hypothetical cut inside the body or an actual surface. The quantity

\boldsymbol{F}

is the deformation gradient tensor,

J

is its determinant.

Cauchy stress

The Cauchy stress (or true stress) is a measure of the force acting on an element of area in the deformed configuration. This tensor is symmetric and is defined via

df=t~d\Gamma=\boldsymbol{\sigma}Tn~d\Gamma

or

t=\boldsymbol{\sigma}Tn

where

t

is the traction and

n

is the normal to the surface on which the traction acts.

Kirchhoff stress

The quantity,

\boldsymbol{\tau}=J~\boldsymbol{\sigma}

is called the Kirchhoff stress tensor, with

J

the determinant of

\boldsymbol{F}

. It is used widely in numerical algorithms in metal plasticity (where thereis no change in volume during plastic deformation). It can be called weighted Cauchy stress tensor as well.

Piola–Kirchhoff stress

See main article: Piola–Kirchhoff stress tensor.

Nominal stress/First Piola–Kirchhoff stress

The nominal stress

\boldsymbol{N}=\boldsymbol{P}T

is the transpose of the first Piola–Kirchhoff stress (PK1 stress, also called engineering stress)

\boldsymbol{P}

and is defined via

df=t~d\Gamma=

Tn
\boldsymbol{N}
0~d\Gamma

0=\boldsymbol{P}n0~d\Gamma0

or

t0=t\dfrac{d{\Gamma}}{d\Gamma0}=

Tn
\boldsymbol{N}
0

=\boldsymbol{P}n0

This stress is unsymmetric and is a two-point tensor like the deformation gradient.
The asymmetry derives from the fact that, as a tensor, it has one index attached to the reference configuration and one to the deformed configuration.[4]

Second Piola–Kirchhoff stress

df

to the reference configuration we obtain the traction acting on that surface before the deformation

df0

assuming it behaves like a generic vector belonging to the deformation. In particular we have

df0=\boldsymbol{F}-1df

or,

df0=\boldsymbol{F}-1

Tn
\boldsymbol{N}
0~d\Gamma

0 =\boldsymbol{F}-1t0~d\Gamma0

The PK2 stress (

\boldsymbol{S}

) is symmetric and is defined via the relation

df0=

Tn
\boldsymbol{S}
0~d\Gamma

0=\boldsymbol{F}-1t0~d\Gamma0

Therefore,
Tn
\boldsymbol{S}
0

=\boldsymbol{F}-1t0

Biot stress

\boldsymbol{U}

. The Biot stress is defined as the symmetric part of the tensor

\boldsymbol{P}T\boldsymbol{R}

where

\boldsymbol{R}

is the rotation tensor obtained from a polar decomposition of the deformation gradient. Therefore, the Biot stress tensor is defined as

\boldsymbol{T}=\tfrac{1}{2}(\boldsymbol{R}T\boldsymbol{P}+\boldsymbol{P}T\boldsymbol{R})~.

The Biot stress is also called the Jaumann stress.

The quantity

\boldsymbol{T}

does not have any physical interpretation. However, the unsymmetrized Biot stress has the interpretation

\boldsymbol{R}T~df=(\boldsymbol{P}T\boldsymbol{R})

Tn
0~d\Gamma

0

Relations

Relations between Cauchy stress and nominal stress

From Nanson's formula relating areas in the reference and deformed configurations:

n~d\Gamma=J~\boldsymbol{F}-Tn0~d\Gamma0

Now,

\boldsymbol{\sigma}Tn~d\Gamma=df=

Tn
\boldsymbol{N}
0~d\Gamma

0

Hence,

\boldsymbol{\sigma}T(J~\boldsymbol{F}-Tn0~d\Gamma0)=

Tn
\boldsymbol{N}
0~d\Gamma

0

or,

\boldsymbol{N}T=J~(\boldsymbol{F}-1\boldsymbol{\sigma})T=J~\boldsymbol{\sigma}T\boldsymbol{F}-T

or,

\boldsymbol{N}=J~\boldsymbol{F}-1\boldsymbol{\sigma}    and    \boldsymbol{N}T=\boldsymbol{P}=J~\boldsymbol{\sigma}T\boldsymbol{F}-T

In index notation,

NIj=

-1
J~F
Ik

~\sigmakj    and    PiJ=J~\sigmaki

-1
~F
Jk

Therefore,

J~\boldsymbol{\sigma}=\boldsymbol{F}\boldsymbol{N}=\boldsymbol{F}\boldsymbol{P}T~.

Note that

\boldsymbol{N}

and

\boldsymbol{P}

are (generally) not symmetric because

\boldsymbol{F}

is (generally) not symmetric.

Relations between nominal stress and second P–K stress

Recall that

Tn
\boldsymbol{N}
0~d\Gamma

0=df

and

df=\boldsymbol{F}df0=\boldsymbol{F}(\boldsymbol{S}Tn0~d\Gamma0)

Therefore,
Tn
\boldsymbol{N}
0

=

Tn
\boldsymbol{F}\boldsymbol{S}
0
or (using the symmetry of

\boldsymbol{S}

),

\boldsymbol{N}=\boldsymbol{S}\boldsymbol{F}T    and    \boldsymbol{P}=\boldsymbol{F}\boldsymbol{S}

In index notation,

NIj=SIK

T
~F
jK

   and    PiJ=FiK~SKJ

Alternatively, we can write

\boldsymbol{S}=\boldsymbol{N}\boldsymbol{F}-T    and    \boldsymbol{S}=\boldsymbol{F}-1\boldsymbol{P}

Relations between Cauchy stress and second P–K stress

Recall that

\boldsymbol{N}=J~\boldsymbol{F}-1\boldsymbol{\sigma}

In terms of the 2nd PK stress, we have

\boldsymbol{S}\boldsymbol{F}T=J~\boldsymbol{F}-1\boldsymbol{\sigma}

Therefore,

\boldsymbol{S}=J~\boldsymbol{F}-1\boldsymbol{\sigma}\boldsymbol{F}-T=\boldsymbol{F}-1\boldsymbol{\tau}\boldsymbol{F}-T

In index notation,

SIJ=

-1
F
Ik

~\taukl

-1
~F
Jl

Since the Cauchy stress (and hence the Kirchhoff stress) is symmetric, the 2nd PK stress is also symmetric.

Alternatively, we can write

\boldsymbol{\sigma}=J-1~\boldsymbol{F}\boldsymbol{S}\boldsymbol{F}T

or,

\boldsymbol{\tau}=\boldsymbol{F}\boldsymbol{S}\boldsymbol{F}T~.

Clearly, from definition of the push-forward and pull-back operations, we have

\boldsymbol{S}=\varphi*[\boldsymbol{\tau}]=\boldsymbol{F}-1\boldsymbol{\tau}\boldsymbol{F}-T

and

\boldsymbol{\tau}=\varphi*[\boldsymbol{S}]=\boldsymbol{F}\boldsymbol{S}\boldsymbol{F}T~.

Therefore,

\boldsymbol{S}

is the pull back of

\boldsymbol{\tau}

by

\boldsymbol{F}

and

\boldsymbol{\tau}

is the push forward of

\boldsymbol{S}

.

Summary of conversion formula

Key: J=\det\left(\boldsymbol\right),\quad\boldsymbol=\boldsymbol^\boldsymbol=\boldsymbol^,\quad\boldsymbol=\boldsymbol\boldsymbol,\quad \boldsymbol^T=\boldsymbol^, \boldsymbol=J\boldsymbol\boldsymbol^,\quad\boldsymbol=J\boldsymbol,\quad\boldsymbol=J\boldsymbol^\boldsymbol\boldsymbol^,\quad\boldsymbol=\boldsymbol^\boldsymbol,\quad\boldsymbol=\boldsymbol\boldsymbol

Conversion formulae
Equation for

\boldsymbol{\sigma}

\boldsymbol{\tau}

\boldsymbol{P}

\boldsymbol{S}

\boldsymbol{T}

\boldsymbol{M}

\boldsymbol{\sigma}=

\boldsymbol{\sigma}

J-1\boldsymbol{\tau}

J-1\boldsymbol{P}\boldsymbol{F}T

J-1\boldsymbol{F}\boldsymbol{S}\boldsymbol{F}T

J-1\boldsymbol{R}\boldsymbol{T}\boldsymbol{F}T

J-1\boldsymbol{F}-T\boldsymbol{M}\boldsymbol{F}T

(non isotropy)

\boldsymbol{\tau}=

J\boldsymbol{\sigma}

\boldsymbol{\tau}

\boldsymbol{P}\boldsymbol{F}T

\boldsymbol{F}\boldsymbol{S}\boldsymbol{F}T

\boldsymbol{R}\boldsymbol{T}\boldsymbol{F}T

\boldsymbol{F}-T\boldsymbol{M}\boldsymbol{F}T

(non isotropy)

\boldsymbol{P}=

J\boldsymbol{\sigma}\boldsymbol{F}-T

\boldsymbol{\tau}\boldsymbol{F}-T

\boldsymbol{P}

\boldsymbol{F}\boldsymbol{S}

\boldsymbol{R}\boldsymbol{T}

\boldsymbol{F}-T\boldsymbol{M}

\boldsymbol{S}=

J\boldsymbol{F}-1\boldsymbol{\sigma}\boldsymbol{F}-T

\boldsymbol{F}-1\boldsymbol{\tau}\boldsymbol{F}-T

\boldsymbol{F}-1\boldsymbol{P}

\boldsymbol{S}

\boldsymbol{U}-1\boldsymbol{T}

\boldsymbol{C}-1\boldsymbol{M}

\boldsymbol{T}=

J\boldsymbol{R}T\boldsymbol{\sigma}\boldsymbol{F}-T

\boldsymbol{R}T\boldsymbol{\tau}\boldsymbol{F}-T

\boldsymbol{R}T\boldsymbol{P}

\boldsymbol{U}\boldsymbol{S}

\boldsymbol{T}

\boldsymbol{U}-1\boldsymbol{M}

\boldsymbol{M}=

J\boldsymbol{F}T\boldsymbol{\sigma}\boldsymbol{F}-T

(non isotropy)

\boldsymbol{F}T\boldsymbol{\tau}\boldsymbol{F}-T

(non isotropy)

\boldsymbol{F}T\boldsymbol{P}

\boldsymbol{C}\boldsymbol{S}

\boldsymbol{U}\boldsymbol{T}

\boldsymbol{M}

See also

References

  1. J. Bonet and R. W. Wood, Nonlinear Continuum Mechanics for Finite Element Analysis, Cambridge University Press.
  2. R. W. Ogden, 1984, Non-linear Elastic Deformations, Dover.
  3. L. D. Landau, E. M. Lifshitz, Theory of Elasticity, third edition
  4. Book: Three-Dimensional Elasticity. 1 April 1988. Elsevier. 978-0-08-087541-5.