Effect of a Gap Nonlinearity on Recursive Parameter Identification Algorithms

摘要

This work examines the performance of the Recursive Least Squares (RLS) Algorithm, and derivative, the Recursive Lattice Least Squares (RLLS) Algorithm in matching a linear model to a simple nonlinear system. The response of a single degree of freedom mass-spring-dashpot system to continuous forcing is simulated, and estimates for the modal parameters are obtained. Nonlinearity is introduced by allowing the restoring spring to slide without friction in a gap of specified width. Such a nonlinearity is of interest since it is a simple model of the effect of loose joints in a deployable spacecraft restoring spring to slide without friction in a gap of specified width. Such a nonlinearity is of interest since it is a simple model of the effect of loose joints in a deployable spacecraft structure. The effect of the gap size on the algorithms is studied, and the noise level and sampling rate are adjusted to see how they influence these results. The RLS Algorithm is found to be the most reliable. Finally, the RLS Algorithm is extended to include an estimate for the gap width, and good results are obtained for the noise-free case. Additional conclusions are made on how to minimize the effect of such a gap nonlinearity on the identification process. Keywords: Parameter identification; Modal analysis; Theses; Nonlinear systems; Space structures. (jhd)