A new Developed Modal Parameter Identification Method Based on Empirical Mode Decomposition and Natural Excitation Technique. (2017)

• Record Type: Journal Article
• Authors: Song, Xingyu; Ma, Hongwei; Wang, Kun
• Is Part Of: Procedia engineering. Volume 199(2017)
• Page Start: 1020
• Page End: 1025
• Publication Date: 2017
• Subjects: Natural Excitation Technique -- Cross-Correlation Function -- Empirical Mode Decomposition -- frequency -- damping ratio; Engineering -- Congresses; Engineering -- Periodicals; Engineering; Conference proceedings; Periodicals; 620.005
• DOI: 10.1016/j.proeng.2017.09.270
• Languages: English
• ISSNs: 1877-7058

Abstract

Abstract: Natural excitation technique(NExT) is a kind of frequently used method. Traditional natural excitation technique uses cross-correlation function that responses between two points to replace impulse response function, identifying parameters with time domain modal identification method. Its shortcomings lie in that the cross-correlation function got through this method reflects influences of each mode, thus significant errors are seen in calculation results and even false modes are generated. For this reason, this article combines empirical mode decomposition with intermittency check criterion with natural excitation technique and puts forward the natural excitation technique based on empirical mode decomposition. To do this, first, the tested dynamic response signals, v1 and v2 of the structural two points are to decomposed by EMD, in which the mixed modes are eliminated with empirical mode decomposition with cut-off frequency and Intrinsic Mode Functions IMF_v1(i) and IMF_v2(i) can be tested respectively. Intrinsic Mode Functions IMF_v1(i) and IMF_v2(i) of the two points that are corresponding to the same modal response are used as modal identification signals and NExT is used to gain cross-correlation function between the two points of this mode. Then the cross-correlation function is taken as the initial input signal to identify parameters with modal identification method in time domain. Numerical simulation and engineering examples show that the proposed method could accurately identify the modal frequency and modal damping ratio of the structure.