Simulation of Stochastic Fluctuating Wind Field Using the Wave Superposition Method with Random Frequencies

Abstract

The wave superposition method is one of the effective methods to simulate the stochastic fluctuating wind field. The simulated wind speed samples with single-indexing frequencies are periodic, with period of T ₀ = 2π · N/ω _u, in which N → ∞ and ω _u being the upper cutoff frequency. While using the double-indexing frequencies, theoretically the samples’ period could increase to nT ₀, where n is the number of multivariate. However, in fact, the period is not increased. It still keeps the value of T ₀. Two examples using the double-indexing frequencies indicate this problem. In this paper, the reason of this problem is studied in detail. And an improved method to generate wind speed samples with non-periodic characteristics is proposed. Formulation is given. In this formulation, the problem existing is that of the enormous memory consumption and time consuming. Therefore, characteristics for the curves of the factorized power spectral density functions changing with frequencies are further studied. And the cubic uniform B-spline interpolation method and the cut-off part of cosine terms method are utilized to improve the simulation efficiency. An example using proposed methods indicates that the simulation is successful. The simulated stochastic fluctuating wind field has features of non-periodicity. And the algorithm is effective and accurate.