Multi-disciplinary NVH Optimization Of A MPV Door-in-white Door

More and more attention has been paid to the NVH performance of vehicle by producers and consumers.As an important panel,door provides a closed space with the car-body.Optimization design of door-in-white structure can improve door performance and reduce interior-vehicle noise.Aiming at the problem of NVH sound insulation in the medium and low frequency noise of a MPV door-in-white.The main work of this article as follows:?1?According To the CATIA three-dimensional structure model of the door-in-white,i establish a finite element model in Hypermesh,and determine the connection of each plate?including flanging,solder joint,adhesive,bolt,door lock,window glass connection?and simulation of the glass.The free mode and door window frame of the established finite element model are carried out by Optistruct.The finite element analysis of stiffness and strength under mechanical load,two kinds of sinking and four torsional conditions is carried out.According to the industry standard,the performance of white door is analyzed and evaluated according to the analysis results.The performance parameters provide a reference standard for the next step of optimization design.?2?According to the acoustic finite element method and sound solid coupling theory,the acoustic transmission loss of rectangular plate is deduced.After comparing PML and AML technology,i simulate the reverberation room and the anechoic chamber in Virtual.lab to do door-in-white strongly coupled experiment.Then i got the sound insulation curve of the 100Hz-1000 Hz range is obtained,and found the 300Hz-400 Hz frequency of the low sound pressure.Also interval and plate are obtained,and the sound pressure cloud,displacement and amplitude of cloud and displacement curves are obtained.Doing of those are prepared for the next chapter of the test design and optimization.?3?Comparing and analyzing several test design techniques and approximate model methods,the thickness of six pieces of door-in-white panel: inside and outside plate and middle and upper plate Property₀.65,right lower guide slot Property₁.0,window frame upper Property₀.8,right and right underpad Property₁.8 in window frame,glass glass₄.3,and inner panel thickening Property₁.3,etc.Those six panel thickness are as a design factor,i use the optimized Latin square test design to collect sample data in the input space to get 50 levels of test design matrix.Using the responses of the first order bending mode of the door-in-white,the front stiffness of the window frame,the stiffness of the window frame,the stiffness of the weight,the stiffness of the weight,the 4 strength of the torsional working condition,the 4 strength of the torsional working condition,the quality of the white car door and the volume mean square in the 300Hz-400 Hz range,as the output,i calculated those in the software Hypermesh,Optistruct and Virtual.lab 13.6.Finally,in the software Isight,i use polynomial response surface model to establish an approximate model of six factors,and analyze the error accuracy of the approximate model.The analysis results show that the fitting accuracy of the approximate model is high,and it can be used as an alternative model of finite element for optimization calculation.?4?According to the established response surface approximation model,the optimized Latin square is used again for 200 sampling in the design range of six factors,and the main effect map,Pareto diagram,and interaction effect diagram are obtained by using the experimental design post processing tools.The six plate thickness parameters which have the most influence on the output variables are found,and six input x1,x2,x3,x4,x5 and x6 are optimized as the optimization variables,and the square root of the sound insulation of door-in-white is optimized in the 300Hz-400 Hz range.The first order bending modal f1?32Hz,the front stiffness wf?6mm,the middle stiffness wm?6mm,the sinking rigidity of the window frame sink1?3.5mm,the sinking stiffness of the 750 N working condition sink2?4mm,and the quality of the door-in-white m?0.028 ton,strength of torsional working condition tor4?235MPa,as constraints,then use the multi-Island genetic algorithm to optimize the plate thickness.The optimization results show that in the case of a slight increase in the quality of the white door,the average square root of the volume increases 9.81 d BA within the range of 300Hz-400 Hz,and the other constraints are within the range.The optimized input parameters are brought into the model to calculate the real value,and the approximate value of the response surface model is compared.At the same time,the effectiveness of the response surface model and the optimization result is proved.