【摘要】 风场模拟中广泛应用的AR模型主要有标量过程的AR模型和向量过程的AR模型.通过取数学期望的方法,可以得到标量过程AR模型模拟空间相关性风场所采用的参数,但该参数并未考虑时滞的情况.向量过程AR模型建立在联合平稳的假设之上,可以很好地综合考虑各点的空间相关性,且回归系数矩阵包含时滞的影响,可以考虑较长时间范围内各点风速时程的时间相关性,从理论来说,要相对完善一些.本文算例表明,向量过程AR模型所模拟的风场精确度总体上要高于标量过程AR模型所模拟的风场,但也存在所模拟风场各点的自相关函数偏差大和总体运算时间多等缺点.更多还原

【Abstract】 There are two kinds of AR models,i.e.scalar and vector process,applied into wind field simulation widely.In the case of AR model of scalar process,spatial correlation parameter of AR model for simulating wind field is derived from mathematic exception and time-lag is not considered in the parameter of AR model of scalar process.While in the case of AR model of vector process,various points’ space correlation is considered well because hypothesis of jointly stationary random processes is adopted,and its regressive parameter matrix includes effects of time-lag.Therefore,from the theoretical point of view,AR model of vector process is perfect.Numerical examples show that the more accuracy of the wind field simulated by AR model of vector process can be obtained,while there are few shortcomings in this model,e.g.bigger error for auto-correlation and more running time.更多还原