Vibration load and transfer path identification of vehicle using inverse matrix method based on singular value decomposition

Abstract

Based on the theory of inverse transfer matrix, a novel method for simultaneous load identification of vehicle vibration is presented in this paper. Some response, excitation, reference points (called key points) and their transfer paths, which have severe effects on the vibration of a whole vehicle, are defined. The transfer functions among the key points are measured by experiments, and thereby a transfer function matrix of vehicle vibration is established. To solve ill-conditioning problem in the transfer function matrix, the methodology of singular value decomposition is introduced into matrix inversion in the excitation load identification. To reduce the identification error, four transfer function matrices with different reference points and condition numbers are selected and discussed. The results show that the more the reference points are, the smaller the condition number of transfer function matrix is, the higher the accuracy of excitation load identification. The transfer function matrix with minimum condition number is used to identify the excitation loads at the vehicle key points. Experimental verifications suggest that the newly proposed method is effective and feasible for excitation load identification of vehicle vibration. Using the identified excitation loads, furthermore, the vibration causes of the steering wheel and seat rail are obtained, which is helpful for improving vibration performance of the sample vehicle. In applications, the excitation load identification method proposed in this paper may be applied not only to other types of vehicle, but also to other complex electromechanical products for load identification in engineering.