The explicit inverse of the stiffness matrix (Q1205009)

This paper has for the first time presented the explicit inverse of the stiffness matrix of a linear elastic structure. It is based on the property that the inverse of a nonsingular matrix is equal to the adjoint of the matrix divided by the determinant of the matrix. The method for writing the inverse employs the congruent product form for the system stiffness matrix of the structure. The matrix is expressed as the product of the statics matrix, the unassembled element stiffness matrix in deformation coordinates, and the kinematic matrix. For the general case of redundant structure with \(M\) elements and \(N\) nodal degrees of freedom, the statics and kinematics matrices are rectangular matrices of order \(N\times M\) and \(M\times N\) respectively. It has indicated that by a proper transformation the unassembled element stiffness matrix is a diagonal matrix of order \(M\times M\). This results from modeling the structure as an assemblage of unimodal elements. Based on results obtained independently, it was shown that the determinant of the stiffness matrix is equal to the sum of the determinants of the stiffness matrices of all the statically determinate substructures which can be derived form the original structure. The expression for the adjoint of the stiffness matrix was obtained from the theorem on the product of compound (rectangular) matrices. Having obtained the inverse of stiffness matrix, and the explicit expression of the nodal displacements, the element internal forces were computed by premultiplying the displacements vector by the natural stiffness and kinematic matrices.