基于改进EMD分解的时变结构密集模态的瞬时参数识别

Identifying Parameters of Time Varying Structures with Closely Spaced Modes Based on Improved EMD Method

提出基于改进经验模态分解(EMD分解)识别含有密集模态的时变结构的瞬时参数的方法。通过波组信号前处理和正交化经验模态分解方法(OEMD)解决传统的EMD无法分解2个近频模态的固有模态函数(IMF)和IMF分量之间不正交这两个问题,将该方法应用于时变结构密集模态的瞬时参数的识别中,给出基于此方法识别时变结构密集模态参数的步骤,并通过一个含有密集模态的3自由度时变结构算例验证了该方法的正确性、有效性以及识别密集模态的优势。

In identifying the parameters with the empirical mode decomposition( EMD) method,there exists the deficiency of the mode separation of two intrinsic mode function( IMF) components with near frequency and the non-orthogonality between IMF components,which make the parameters identification results not accurate and stable enough. To solve these problems,we propose the improved EMD method which combines the wave group( WG) signal processing method with the orthogonal empirical mode decomposition( OEMD). Based on the improved EMD method,the parameter identification procedures for time varying structures with closely spaced modes are proposed and applied to the numerical simulation with a three degrees-of-freedom time varying dynamic model. The simulation results demonstrate that the proposed method has very good effectiveness,efficiency and has the advantages of identifying the parameters of time varying structures with closely spaced modes.