Structural system identification using Least Mean Square (LMS) adaptive technique

Abstract

This paper presents an approach to identify structural parameters using the Least Mean Square (LMS) adaptive transversal filter. This method features easy computer simulation and fast data processing. This approach is effective even in the presence of input noise when it is applied in communications; however, this advantage is not obvious when it is used in structural systems, since the signal bandwidht it deals with in communications is much wider than the bandwidth in structure systems. With a low-pass filter added to a data acquisition system, this approach functions effectively for both linear and nonlinear structural systems. Numerical examples and experimental tests have been included to demonstrate the technique for both linear and nonlinear systems.