Simulation of stationary and nonstationary wind velocity field along a long-span bridge using a numerical truncation method

Abstract

The spectral representation method (SRM) has been extensively utilized to generate inhomogeneous stochastic wind velocity fields of long-span bridges. However, the simulation process usually needs a large number of computing resources, particularly when plenty of simulation points/samples are required. To address this issue, this study proposes a numerical truncation approach with the aid of a closed-form solution of Cholesky decomposition regarding the coherence matrix. First, the double summation of SRM is optimized as two separate inner and outer summations. Then, the closed-form solution of Cholesky decomposition is cut off to simplify the inner summation by setting an appropriate threshold. Since this inner summation dispenses with the participation of the truncated elements, the computational efforts are reduced significantly. As for the outer summation, it can be fast completed via introducing the fast Fourier transform technique. Finally, numerical examples are used to verify the effectiveness of the proposed method in terms of accuracy and efficiency. The results show that this method possesses a satisfactory accuracy and high efficiency with a given threshold. Therefore, it presents a high potential in simulating large-scale stochastic fluctuating wind velocity fields.