Application of Dual-Tree Complex Wavelet Packet Transform for Generating Synthetic Multivariate Nonstationary Non-Gaussian Thunderstorm Wind Records

Abstract

The available thunderstorm wind records with subsecond sampling intervals is scarce for a given site; stochastic models that can be used to sample multivariate nonstationary non-Gaussian thunderstorm winds at multiple points or tricomponent thunderstorm winds at a point are lacking. We propose the use of the dual-tree complex wavelet packet transform (DT-CWPT) within the framework of the iterative power and amplitude correction (IPAC) algorithm to generate multivariate nonstationary non-Gaussian thunderstorm wind records. This is a data-driven or seed-record-based approach, and the use of the IPAC algorithm ensures the matching of the marginal cumulative probability distribution function. The DT-CWPT is used to gain computational efficiency because it is a redundant transform with a low redundancy factor, and it provides phase information. The statistics of the time-frequency power spectral density of the sampled records and the seed record were compared to show the adequacy and effectiveness of the proposed approach. The results also show that the use of the DT-CWPT instead of the (discretized) continuous wavelet transform and S-transform significantly reduces the computational time.