中文摘要：

基于风洞试验中采集的厦门沿海某高层建筑模型的表面风压时程，计算各测点脉动风压的分形维数，偏度系数和峰度系数，找出脉动风压呈现较明显的非高斯特性的4个测点．考虑脉动风时程具有分形的特性，结合相关性变形法，采用具有分形特性的Weierstrass-Mandelbrot函数对非高斯特性明显的4个测点的脉动风压进行了模拟．结果表明，模拟得到的4个测点的非高斯脉动风压时程与风洞试验的非高斯脉动风压时程的概率分布曲线吻合较好，功率谱图形比较相似，分形维数，偏度系数和峰度系数均比较接近．说明本文所提出的用分形方法模拟具有非高斯特性的脉动风压是可行的，弥补了传统的线性滤波器法和谐波叠加法模拟结果均不具有分形特性的缺点，对非高斯脉动风压的数值模拟有一定的参考价值．

英文摘要：

Based on the time history of wind pressure collected from the wind tunnel tests on a high-rise building at Xiamen coast, four points with obvious non-Gaussian characteristics were found out by the calculation of fractal dimension, skewness and kurtosis.Considering the fractal characteristics of time history of fluctuating wind and the correlation distortion method, the non-Gaussian fluctuating wind pressure was simulated by adopting fractal Weierstrass-Mandelbrot function.Comparison of simulation results with ex- perimental results in the aspects of probability distribution, power spectrum, fractal box dimension, skew- ness and kurtosis, showed better consistency.Therefore,it is feasible to simulate the fluctuating wind pres- sure with non-Gaussian characteristics by using fractal method.The method makes up for the shortcomings of traditional linear filter and superposition of harmonic, and provides a reference for the simulation of non- Gaussian fluctuating wind pressure.