Sequential Prediction-Error Method for Structural Identificaion

Abstract

Time-domain methods for the identification of linear structural dynamic systems are studied. The stochastic autoregressive and moving average (ARMAX) model is used to process the measured excitation and response records contaminated by noises. The study focuses on the sequential prediction-error method incorporating several techniques for improving the parameter estimation. They are the exponential data weighting, the global data weighting, and the square-root estimation techniques. Efficient procedures of the square-root estimation are developed for the multi-input and multioutput (MIMO) case as well as the multi-input and single-output (MISO) case. Verifications of the present methods are carried out using the simulated time histories for the input excitation and output response, as well as using the experimental data on a building model. The results indicate that the square-root estimation technique is particularly effective for improving the convergence and accuracy of the sequential estimation, even with crude initial guesses.