Influence of physical parameters on system eigenvalues and its use in physical model reduction

The identification of subsystems and/or components that determine a given eigenvalue of the overall system is a challenging and important topic. This paper proposes a set of theorems and definitions that lead to an efficient procedure for this purpose. In the procedure, the bond graph representation of dynamic systems is utilized since they lead to better understanding of the system structure. The procedures include the calculation of eigenvectors. After this calculation an "effect" matrix is produced that indicates the relative importance of physical parameters in a selected eigenvalue. In addition to the relative importance, an efficient physical model reduction procedure can be constructed. Two examples are given to illustrate the approach.