Structural System Parameter Estimation by Two-Stage Least Squares Method

Abstract

A two-stage least-squares parameter estimation procedure using covariance function data is developed for the estimation of the natural frequency and damping parameters of randomly excited structural systems. The estimation procedure is statistically and computationally efficient and yields a measure of the statistical reliability of the structural parameter estimates. Examples of the estimation of the natural frequency and damping computed by forced vibration, correlation analysis, and two-stage least-squares methods are examined.