Abstract
An approach for the use of the Weighted Least Squares method for time domain modal parameter identification is introduced. The approach enables standard identification methods to be employed in the usual manner, except that sections of the time records with high signal-noise ratios are given a greater emphasis in the identification process. It is shown how it is possible to apply the method to a wide range of applications, including the analysis of impulse response, input-output and time varying data. A number of different weighting schemes are explored, based upon exponential weighting and the envelope of the measured data. A series of simulated tests demonstrate how the performance of conventional time domain algorithms can be improved significantly with little extra computational effort though the use of a Weighted Least Squares scheme.