基于AR算法的小波技术在随机场模拟中的应用  

Application of wavelet technique based on AR method in random field simulation

作　　者：徐林[1] 叶继红[1] 

机构地区：[1]东南大学混凝土与预应力混凝土教育部重点实验室

出　　处：《振动与冲击》2010年第8期64-72,共9页Journal of Vibration and Shock

基　　金：国家自然科学基金重大研究计划培育项目(90715005);教育部新世纪优秀人才支持计划(NCET-07-0186);教育部博士点基金(200802860007)

摘　　要：采用线性滤波器法中的AR法模拟生成了地震动场和风场的时程信号。由于AR算法具有很高的计算效率,但模拟精度却不理想,故引入小波分析中的小波分解方法和小波包分解方法对AR算法模拟的信号进行调整,并对这两种方法改进的效果进行对比。算例分析表明由改进的AR法得到的样本信号频谱与目标谱吻合良好,且发现小波包分解方法调整信号具有比小波分解方法更好的精度。The time histories signals in earthquake motion field and wind field were simulated with the help of AR (autoregression)method.In view of the deficiency of low accuracy by use of AR method,the signals simulated were modified by wavelet decomposition as well as wavelet packet decomposition methods.The improvement measures were discussed in detail.Several examples show that the simulated spectra by the improved AR method are in good agreement with the corresponding target spectra.Furthermore,the signals modified by the wavelet packet decomposition method are of better precision than those modified by the wavelet decomposition method.

关 键 词：AR法 数值模拟 地震动场 风场 小波分解 小波包分解 

分 类 号：TU311.3[建筑科学—结构工程]