Simulating nonstationary non-Gaussian vector process based on continuous wavelet transform

In the present study, we propose a new iterative algorithm based on discretized continuous wavelet transform (CWT) to simulate nonstationary non-Gaussian vector process for prescribed marginal probability distribution functions and the time-scale power spectral density function matrix. The algorithm is designed for a CWT pair within the wavelet analysis framework. It can be used with analytical wavelets satisfying the admissibility condition and can cope with time dependent and time-independent coherence. The proposed algorithm is applied to generate nonstationary downburst winds and seismic ground motions at multiple sites within the wavelet analysis framework. The numerical validation analysis confirms that the proposed algorithm can lead to the simulated records matching the prescribed marginal probability distribution, marginal time-scale power spectral density function and coherence function.

Automated summary

This study introduces a new iterative algorithm based on discretized continuous wavelet transform (CWT) for simulating nonstationary non-Gaussian vector processes. The algorithm is designed for a CWT pair within the wavelet analysis framework, allowing for handling time-dependent and time-independent coherence. Applied to generate nonstationary winds and ground motions, the algorithm matches prescribed probability distributions and coherence functions.