Bilinear System Characteristics from Nonlinear Time Series Analysis

Detection of changes in the resonant frequencies and mode shapes of a system is a fundamental problem in dynamics. This paper describes a time series method of detecting and quantifying changes in these parameters for a ten degree-of-freedom bilinear system excited by narrow band random noise. The method partitions the state space and computes mode frequencies and mode shapes for each region. Different regions of the space may exhibit different mode shapes, allowing diagnosis of stiffness changes at structural discontinuities. The method is useful for detecting changes in the properties of joints in mechanical systems or for detection of damage as the properties of a structure change during use