风时程模拟的高效高精度混合法

结合谐波叠加法(Weighted Amplitude Wave Superposition,WAWS)和自回归方法(Auto Regressive,AR)的优点,提出高效高精度新型混合法。在频域上开窗,将计算定义域划分为局部定义域。推导了应用于AR方法的Levison-Durbin矩阵递推公式,提出利用ACI准则判断最佳回归阶次。利用该混合法对大跨空间结构风速时程进行模拟。计算显示在保证计算精度下,该方法风速时程的模拟效率是谐波叠加法的数十倍;计算精度与自回归方法相比亦显著提高,且易于收敛。

THE COMPOSITE APPROACH FOR WIND TIME SERIES SIMULATION

Based on weighted amplitude wave superposition method (WAWS) and auto regressive method (AR), a new composite approach for wind time series simulation was proposed. This approach can significantly reduce computation effort with no accuracy lost, where the frequency domain was divided into a group of sub-domains by creating frequency sub-windows. The wind time series simulation for a spatial structure was carried out as a numerical example. Furthermore, the matrix expression of Levison-Durbin algorithm was put forward and the optimally regression order was judged by the ACI criterion. Finally, it is shown from computation and comparison that the presented approach has much higher computation efficiency than WAWS and better precision than AR.