This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Section 1.11 Solid Mechanics Part III Kelly 96 1.11 The Eigenvalue Problem and Polar Decomposition 1.11.1 Eigenvalues, Eigenvectors and Invariants of a Tensor Consider a second-order tensor A . Suppose that one can find a scalar and a (non-zero) normalised, i.e. unit, vector n such that n n A = (1.11.1) In other words, A transforms the vector n into a vector parallel to itself, Fig. 1.11.1. If this transformation is possible, the scalars are called the eigenvalues (or principal values ) of the tensor, and the vectors are called the eigenvectors (or principal directions or principal axes ) of the tensor. It will be seen that there are three vectors n (to each of which corresponds some scalar ) for which the above holds. Figure 1.11.1: the action of a tensor A on a unit vector Equation 1.11.1 can be solved for the eigenvalues and eigenvectors by rewriting it as ( ) = n I A (1.11.2) or, in terms of a Cartesian coordinate system, ( ) ( ) ( ) = = = i i j ij r r i j ij r r q p pq k k j i ij n n A n n A n n A e e e e e e e e e In full, [ ] [ ] [ ] ) ( ) ( ) ( 3 3 33 2 32 1 31 2 3 23 2 22 1 21 1 3 13 2 12 1 11 = + + = + + = + + e e e n A n A n A n A n A n A n A n A n A (1.11.3) Dividing out the base vectors, this is a set of three homogeneous equations in three unknowns (if one treats as known). From basic linear algebra, this system has a solution (apart from = i n ) if and only if the determinant of the coefficient matrix is zero, i.e. if A n n Section 1.11 Solid Mechanics Part III Kelly 97 det ) det( 33 32 31 23 22 21 13 12 11 = = A A A A A A A A A I A (1.11.4) Evaluating the determinant, one has the following cubic characteristic equation of A , III II I 2 3 = + A A A Tensor Characteristic Equation (1.11.5) where ( ) [ ] A A A A A A A det III ) tr( ) (tr II tr I 3 2 1 2 2 2 1 2 1 = = = = = = k j i ijk ij ji jj ii ii A A A A A A A A (1.11.6) It can be seen that there are three roots 3 2 1 , , , to the characteristic equation. Solving for , one finds that 3 2 1 1 3 3 2 2 1 3 2 1 III II I = + + = + + = A A A (1.1.7) The eigenvalues (principal values) i must be independent of any coordinate system and, from Eqn. 1.11.5, it follows that the functions A A A III , II , I are also independent of any coordinate system. They are called the principal scalar invariants (or simply invariants ) of the tensor. Once the eigenvalues are found, the eigenvectors (principal directions) can be found by solving ) ( ) ( ) ( 3 33 2 32 1 31 3 23 2 22 1 21 3 13 2 12 1 11 = + + = + + = + + n A n A n A n A n A n A n A n A n A (1.11.8) for the three components of the principal direction vector 3 2 1 , , n n n , in addition to the...
View Full Document
- Fall '11