Model-Order Reduction by Simultaneous Realization of Eigenvalues and Mode Shapes (SREM)

In this paper, a unique technique "cost function" has been presented to simultaneously realize eigenvalues and mode shape vectors to attain a reduced model. Differential evolution algorithm has been utilized in order to numerically optimize the non-linear cost function instead of the least squares solution of the characteristic equation of the system. The modal matrix is reduced by effective independence distribution vector (EIDV) method to remove the slave degrees of freedom and retain the master ones which have the most contribution in the system response. EIDV retains those degrees of freedom (DOFs) in such a way as to reserve the system information content, as much as possible. This procedure has been verified with some examples and good results have been obtained. It is shown that the algorithm has several advantages, e. g., the coupling between selected modes of full-order model will be attained to guarantee the stability of closed-loop system. [DOI: 10.1115/1.4004573]