Practical Reduction of Structural Eigenproblems

Abstract

An evaluation of publications on reduction of structural eigenvalue problems has revealed the following: (1) Most of the published techniques are essentially equivalent to Guyan reduction; (2) published techniques to decrease the error in Guyan reduction require matrix inversions or an iterative process, or both. This paper presents a modified Guyan reduction method which decreases errors in the calculation of the eigenvectors of the system and which does not require matrix inversions whatsoever. It also shows that a simple iterative process could be used to improve further the eigenvalue solution of the system.