Research Of Load Identification On Whole Vehicle NVH

Traditional approaches of improving the performance of vehicle noise,vibration and harshness(NVH)mainly relies on the unilateral modification of the system property in the late-stage of the whole vehicle.However,the two perspectives including system property and load excitation are seldom considered to improve the NVH performance.Additionally,due to the lack of force excitation data,only the virtual unit excitation or the random excitation rather than the actual exciting loads is considered to improve the NVH performance.These approaches lead to the large amplitude of the vibration and noise in the late-stage of the upgrade and replacement of the vehicles,from which the actual measurement thus cannot meet the design requirements.Load identification,served as an approach to reliably provide the load data,plays a significant role in the improvement of NVH performance in recent years.However,the dynamic loads cannot be tested directly through the force transducer.While the structural response and the system property can be accurately measured and simulated,which reminds us that the load can be indirectly identified through these measurements.Since the load is identified inversely,the general problem of the inverse process should be solved.Most importantly,the ill-posedness as well as the uncertainty problem studied emphatically in this paper is also should be simultaneously resolved in the load identification.This dissertation conducts a research on the ill-posedness and uncertainty of the load identification,aims at contributing some useful researches on providing some effective solutions for the improvement of NVH performance of the whole vehicle.As a result,the following studies are carried out in this paper:(1)General regularization methods applicable to the time domain are studied.Firstly,dynamic loads are expressed as a series of impact functions.Based on the Duhamel integration,convolution relationship among structural response,dynamic load and Green's kernel function is established to describe the forward problem of load identification.Secondly,the ill-posedness of the inverse problem is analyzed.Thirdly,the regularization method and the criterion of selecting the best regularized parameter are introduced to improve the ill-posedness of the inverse problem based on the spectrum analysis.Finally,the load is identified stably and reliably.(2)Based on the interval analysis,a novel load identification method in frequency domain is proposed for the uncertain structure.Similarly,the product relationship among the structural response,frequency response function and dynamic load are constructed for the uncertain structure in model of forward problem.Based on the perturbation theory,the problem of load identification for the uncertain structure transform into the assumption of the midpoint load and the perturbation load,which are two kinds of deterministic problems.Then,Moore-Penrose method is addressed to solve the midpoint load and the TTLS regularization method is applied to solve the perturbation load.Finally,the upper bound and lower bound of the load are identified effectively.(3)An acoustic-pressure based load identification method for the acoustic-structure interaction(ASI)system with non-probabilistic parameters is proposed.The forward problem of the load identification is constructed through the ASI,and further the formulation among acoustic pressure response,Green's function and load excitation is deduced.The load acting on the structure in the ASI system is reversed through the regularized method.Considering the non-probabilistic uncertainty widely existing in the ASI system,the bound of the load is expressed into the assumption of midpoint load and perturbate load according to the Taylor expansion.Finally,the bounds of the load are identified.(4)Based on one MPV vehicle,the experiment for the load identification of the powertrain is carried out.The process of load identification is constructed.The tests for the transfer function from the excited point to the response point and the operational response are obtained successfully.Then,by using the TSVD technique,the dynamic load of the powertrain is identified.