Operational Modal Analysis Based On Transmissibility Measurements

Accurately identifying the dynamic characteristics of the system is essential for the reduction of structural damage,vibration and noise.Operational Modal Analysis(OMA)is an important technique to obtain the dynamic parameters of the system,which has great prospects in the fields of ground vehicles,aircraft,ships,bridges and buildings.However,when the classical OMA methods based on the assumption of random and stationary white noise excitation are applied to identify the parameters of a system that is both subjected to environmental random excitations and harmonic excitations,these harmonic frequencies and their forced response mode shapes may be wrongly identified as physical modes,which will cause great interference to the system’s modal parameter extraction.To solve this problem,the method of transmissibility based OMA(TOMA)was proposed,which takes the advantage of the property that the transmissibility function is only related to the mode shapes of measurement points and becomes independent of the excitation at modal frequencies.This method can suppress the false modes caused by harmonic excitations and it expands the application scope of OMA.In this paper,further researches were carried out on the avoidance of reference selection and the improvement of parameter identification performance on the basis of TOMA.The main work and the conclusions are as follows:Firstly,a peak picking method based on the average of multiple reference measurement points,which is based on the mean of the reciprocal of the second singular values of the transmissibility function matrices was proposed to identify the modal parameters.A 5-DOF mass-spring-damping system and a free beam were used as numerical example and test case respectively,to validate this method from two aspects of loading conditions and measurement noise.The results showed that this method can accurately obtain the modal frequencies and mode shapes of systems regardless of under white noise or non-white noise loading conditions,and has good anti-noise interference ability,but it cannot obtain the damping ratios.Subsequently,the pseudo-frequency response function(p FRF)method based on transmissibility functions,which uses the least squares complex frequency domain(LSCF)method to curve-fit the p FRF to obtain the system poles,was introduced.The application results of the above numerical example and test case showed that the p FRF method can effectively suppress the false modes caused by harmonic excitations,and can accurately identify the modal frequencies,damping ratios and mode shapes of the system under concentrated excitation loading conditions.However,the identification error of damping ratios is large and the stability of parameters identification is easily affected by the selection of the reference measurement point and loading condition under incoherent distributed excitation loading conditions.Moreover,the ability of anti-noise interference needs to be enhanced.Finally,with the p FRFs as elements,a pseudo-frequency response function matrix(p FRFM)was constructed to be the initial data for parameter identification using the LSCF method.The application results of the above simulation example and test case showed that this method can not only accurately identify the modal frequencies,damping ratios and mode shapes of the system,but also effectively suppress the harmonic modes.Meanwhile,this method is without manual selection of the reference measurement point and loading condition,and both the stability of parameters identification and the ability of anti-noise interference outperform the p FRF method.