基于EMD和小波变换的结构风振高阶参振模态识别

柯世堂, 赵林, 邵亚会, 葛耀君

同济大学土木工程防灾国家重点实验室



摘要:

为了确定结构随机理论求解中的高阶参振模态数目,采用经验模式分解(EMD)与小波变换相结合的方法分析结构气弹模型自激响应数据信号的时-频-谱联合特性,从原始信号中分解出固有模态函数(IMF),再对各个IMF进行小波变换提取信号特征参数,从而识别出结构风振随机计算所需的高阶参振模态截止频率,并将识别结果与直接采用随机理论对不同参振模态的计算结果进行对比验证.结果表明:该方法能够准确地识别出结构风振高阶参振模态,并能在任意敏感的频段捕捉到信号变化的局部特征,能更清晰地刻画信号能量随时间、频率的分布.通过数值模拟和对结构响应的随机理论计算论证了该方法的有效性和实用性.

关键词:  经验模式分解  固有模式函数  小波变换  谱分析  参振模态  风洞试验  随机理论

DOI：10.11918/j.issn.0367-6234.2011.04.016

分类号:TP391.41

基金项目:国家自然科学基金资助项目(50978203);科技部重大科技项目(2008ZX06004-001)



Identification of high-order modes of vibration based on EMD and wavelet transform for wind-induced response

KE Shi-tang, ZHAO Lin, SHAO Ya-hui, GE Yao-jun

State Key Laboratory for Disaster Reduction in Civil Engineering,Tongji University, 200092 Shanghai,China

Abstract:

To identify the number of modes of vibration in the random theory results for wind vibration,the empirical mode decomposition(EMD) method and wavelet transform method are used to analyze the joint characteristic for self-induced response signals of aeroelastic model.First,the intrinsic mode functions(IMF) are directly decomposed from original signal with EMD,and then the signal characteristic parameters of every IMF can be extracted with wavelet transform,which are compared with the results by traditional spectrum analysis method and wavelet transform on original signal.The combination of EMD and wavelet transform method can identify the high-order mode of vibration;the main characteristics of signals in arbitrary frequence band can be obtained;and the distribution of signal energy with time and frequence can be displayed more clearly.The numerical simulation and the analysis of response signal date show that this method is effective and practical.

Key words:  EMD  IMF  wavelet transform  spectrum analysis  mode of vibration  wind tunnel test  random theory