New Parameter-Identification Method Based on QR Decomposition for Nonlinear Time-Varying Systems

Abstract

A new identification method based on QR decomposition for nonlinear time-varying systems is proposed in this paper, which is applied in a time-varying multiple-degree-of-freedom (MDOF) dynamic system to locate and estimate nonlinearities without a priori knowledge of them. This new procedure provides a continuous-time model with which a MDOF system can be partitioned into a number of different subsystems. After an orthogonalizing algorithm and an error reduction ratio (ERR) process, all information about the masses and linear and nonlinear connections in the subsystems can be detected and estimated with a high degree of accuracy. In this procedure, the time expressions of the time-varying parameters are also given. Because of its simplicity and efficiency, this new method will have wide application in practical engineering. The reliability of the identification method is demonstrated by a numerical example.