An improved time domain spectral matching method using particle swarm optimization.

Seismic analyses of structures typically require the spectrum of the scaled seed accelerogram to be consistent with a specified target spectrum. Peak ground acceleration (PGA) and Arias intensity levels of scaled and matched ground motions are difficult to maintain after spectral matching with existing time-domain methods. An improved method is thus proposed to address these issues by introducing Arias Intensity Fitting (AIF) and PGA correction algorithms. In addition, adaptive relaxation parameters obtained from particle swarm optimization (PSO) are used to improve convergence efficiency. The effectiveness of the proposed method is validated by a comparison of the spectrally matched records with the scaled time series characteristics. Moreover, the proposed method is compared with two existing time-domain spectral matching methods. Results show that the proposed method performs better in generating the acceleration time series compatible with the target spectrum while preserving the nonstationary characteristics of the scaled seed accelerogram. Considerable improvements in the aspects of PGA and Arias intensity consistency between scaled and matched accelerograms are achieved without reducing the spectral matching accuracy. The convergence efficiency of the proposed method is improved with the implementation of adaptive relaxation parameters. • An improved time domain spectral matching method is proposed. • PGA correction and AIF methods are proposed to appropriate intensity measures. • The intensity measures contain Arias intensity and peak ground acceleration. • Adaptive relaxation parameter is applied to improve convergence efficiency. [ABSTRACT FROM AUTHOR]