Computer simulation of non-Gaussian multiple wind pressure time series

The paper presents a technique for generating artificial multiple wind pressure time series with strong cross correlations in fluctuating features. The technique, for given multiple target time series, simulates an identical shape of sample auto-and cross-spectra of target signals in records of calculated artificial signals and has great flexibility in controlling non-Gaussian distributional properties while preserving the auto-and cross-spectra. The technique utilizes the unique properties of the Fourier phases in Fourier representation of time series in connection to non-Gaussian fluctuating feature generation and the phase shift relations in connection to simulating cross correlations of multiple target signals.