Study On Indentification Method Of Modal Parameter Of Close Models Based On The Fuzzy Clustering

The experimental modal analysis is a kind of effective method in analyzing thedynamic property of the system. In the process of solving practical engineering problems,the experimental modal analysis of the system is a very important method for thestructural design and function estimation. Modern mechanical products design trend to abigger, faster, and more reliable, people have much higher requirement on the product’ssafety, which demand more excellent performance on the precise of modal parameteridentification, Especially for multipoint and close models.This paper analyzes the estimation error during the process of the test data acquisition,and puts forward the measures to void. The coherent functions is used to analysis thequality of the excitation signal and the response signal, According to the noise pollution’scharacteristics in the test environment, we take the different estimation methods for thefrequency response function, and get the high accuracy of frequency response functiondata. In the modal parameter identification methods, this paper presents in detail the highprecision admittance circle fitting of single-mode parameter identification According tothe characteristics of the multipoint and close model and through to the processing to thetest data we use the orthogonal polynomial of the frequency normalization method andimprove the morbidity phenomenon of the transition matrix. It is easily to produce thepole position shift and false pole phenomenon because of the increase in the number of thetest point and so on So we introduce the fuzzy clustering algorithm(FCM) on Fuzzymathematics, and get the more accurate system poles, and improve the accuracy of themodal parameter identification.In order to facilitate the modal analysis, this paper also develop the experimentalmodal analysis software package which realize the modularization of the test datapretreatment and modal parameter identification using the visual basic 6.0 programminglanguage which has friendly interface. Finally we simulate the data which has thecharacteristics of close models and high damping by Matlab/Simulink , and prove thecorrectness of the algorithm, and show the adaptive to the experiments data by the actual modal experiment.