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 secondorder tensor A . Suppose that one can find a scalar and a (nonzero) 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
 Staff

Click to edit the document details