谐波合成法模拟随机风场的优化算法  

Optimized Algorithm of Wave Superposition Method to Simulate Stochastic Wind Field

作　　者：罗俊杰[1] 韩大建[1] 

机构地区：[1]华南理工大学建筑学院,广东广州510640

出　　处：《华南理工大学学报（自然科学版）》2007年第7期105-109,共5页Journal of South China University of Technology(Natural Science Edition)

基　　金：广州市科技攻关项目(2004Z3-E0351)

摘　　要：为解决传统谐波合成法模拟多变量随机脉动风场效率低下的问题,首先将互谱密度函数矩阵分块,然后采用递归优化算法进行矩阵的Cholesky分解,同时引入矩阵乘法算法来代替传统算法的叠加过程,从而减少了互谱密度矩阵中元素生成的数量,加快了矩阵分解和脉动风场合成的速度.工程实例表明,上述优化算法可有效减少整个模拟过程的计算时间,提高模拟效率.In order to improve the efficiency of the conventional wave superposition method in simulating the multivariable stochastic wind fields, an optimized recursive algorithm is employed to conduct the Cholesky factorization by blocking the cross-spectral density function matrix into several sub-matrices, and a matrix multiplication algorithm is introduced to replace the harmonic superposition process of the conventional method. Thus, the number of the elements for the factorization in the cross-spectral density matrix is reduced and the Cholesky faetorization and the wind field synthesis are speeded up. The engineering practice indicates that the proposed optimized algorithm remarkably saves the simulation time and improves the simulation efficiency.

关 键 词：随机风场 优化 算法 数值模拟 矩阵分块 

分 类 号：P315.69[天文地球—地震学] TU351[天文地球—固体地球物理学]