Probe Into Dynamic Condensation Technique For Computing Several Low Order Modal Pairs

Reducing the order of a large structural finite element model is important in improving the efficiency of finite element analysis in the field of structural dynamics and vibration. This has been broadly applied to many fields such as structural damage detection, parameter identification, test-analysis model correlation, modal testing, dynamic substructure technique and so on.In 1965,static reduction was proposed by Guyan and Irons(reduction method).After that,a lot of work have been done in this area and many efficient static and dynamic condensation methods have been proposed., The dynamic condensation method, started in early 1990's, is a widely-adopted effective method for order reduction.dynamic condensation technique has been improved and consummated in the following areas.1. A study was presents on the convergence property of computing the few lowest eigenpairs in large finite element models using dynamic condensation method. Using the power series expansion of the matrix. And two numerical examples have demonstrated that influence the calculating precision if the master and slave DOFs are not separated appropriately.2 This paper studies the convergence property of the dynamic condensation method on computing the few lowest eigenpairs in large finite element models. And a sufficient condition of the method is derived by the theoretical part solving matrix equation.3.The key characteristics and effectiveness of the dynamic condensation method are further illustrated by comparing with the subspace iteration method.