作　　者： (蒋磊）; (李春祥）; (邓莹）;

机构地区： 上海大学土木工程系,上海200444

出　　处： 《上海大学学报（自然科学版）》 2017年第4期600-608,共9页

摘　　要： 提出一种基于线性预测和Z转换(linear prediction and Z-transform,LPZ)谱分析法的非高斯脉动风压模拟算法.运用Johnson变换系统实现高斯随机过程到非高斯白噪声的转换;再使用LPZ对其进行数字滤波,进而得到所需的非高斯脉动风压.采用基于LPZ谱分析法对单变量非高斯随机信号和非高斯脉动风压进行了数值模拟.通过对模拟的非高斯信号和脉动风压的统计参数(峰度和偏度)与目标统计参数,以及模拟的功率谱与目标功率谱进行比较,验证了基于LPZ谱分析法的非高斯脉动风压模拟算法的有效性. This paper presents an algorithm based on linear prediction and Z-transform （LPZ） spectral analysis to simulate non-Gaussian fluctuating wind pressure. The Gaussian process is converted into a non-Gaussian white noise process by Johnson translator system. The non-Gaussian white noise process is then filtered with LPZ spectral analysis to obtain fluctuating wind pressure. The univariate non-Gaussian random signals and non-Gaussian fluctuating wind pressure are simulated using the algorithm. By comparing the statistical parameters including skewness, kurtosis and power spectral density of the non-Gaussian white noise process and fluctuating wind pressure with their target values, it is confirmed that the algorithm based on LZP spectral analysis can effectively simulate non-Gaussian fluctuating wind pressure.

关 键 词： 非高斯脉动风压模拟 线性预测和 转换谱分析法 变换系统 数字滤波