When nothing deforms
The theory of deformation keeps returning to one question: has the neighbourhood of a point really deformed, or has it only moved? The answers are scattered over Part 2. This page collects them; like an “equivalence theorem”, it grows as the course goes on.
Theoremrigid motion of a neighbourhood
For an admissible deformation gradient at a point (), the following are equivalent:
- every fibre keeps its length: for every unit ;
- ;
- is a rotation: , with , ;
- the right stretch tensor is the identity: (and then );
- every principal stretch is 1;
- the Green–Lagrange strain vanishes: ;
- every fibre keeps its length and every right angle stays right: and for all orthogonal .
Proof
- 1 ⇔ 2: Proposition 2.2.5.
- 2 ⇔ 3: Key observation 2.6.1; makes the orthogonal a rotation.
- 2 ⇔ 4: (Proposition 2.4.3) and is the positive square root of ; .
- 4 ⇔ 5: the eigenvalues of are the principal stretches (Corollary 2.4.4), and a symmetric matrix with all eigenvalues 1 is .
- 2 ⇔ 6: (Definition 2.6.2).
- 2 ⇒ 7: Proposition 2.2.12; 7 ⇒ 1 trivially.
Warningwhat the linear theory can and cannot say
For small displacement gradients the same role is played by the infinitesimal strain: means the displacement is, to first order, a translation plus an infinitesimal rotation (Definition 2.7.6). But is not equivalent to the statements above for a finite rotation: a turn has and (Exercise 2.7.2).
Stress tensors at a glance
The three stress tensors of Part 3, for comparison:
| Cauchy | first Piola–Kirchhoff | second Piola–Kirchhoff | |
|---|---|---|---|
| force | current | current | pulled back by |
| per unit area | current | reference | reference |
| symmetric | yes | no | yes |
| from | |||
| stress vector | |||
| balance | |||
| small strain |