A new simulation algorithm of multivariate short-term stochastic wind velocity field based on inverse fast Fourier transform

• The presented formula is theoretically proved and firstly adopted to simulate wind.
• The discrete frequency is started from 1 for ergodicity, zero frequency is excluded.
• The ergodicity of the cross-correlation function is carefully investigated.
• The period of the any simulated process is exactly equal to the length of the sample.
• The computing efficiency is increased four times, specially more than one thousand times.

A novel algorithm to generate samples of stationary multivariate stochastic processes is presented according to a Fourier–Stieltjes integral and based on a strict mathematical deduction. The algorithm is very efficient as it takes full advantage of inverse fast Fourier transform technique, and is especially applicable to the simulation of short-term wind velocity field with many simulation points. By creation of a conjugate-symmetric series, the summing operation of the decomposed lagged coherence matrix is conducted in the frequency-domain, which decreased the executions of IFFT. Furthermore, the ergodic property of the algorithm is carefully investigated, and computational efficiency is compared with that of the existing method.