Research On Evaluation And Optimization Of The Vehicle Door-slamming Sound Quality

With the rapid development of social science and technology and the automobile industry,the demand for the vehicle comfort has been gradually improved.The research on the traditional automobile acoustic is to reduce the sound pressure level of the car,which can’t meet the modern requirements of people on the vehicle comfort.The automobile enterprises are increasingly focused on the research and application of sound quality,and the sound quality has become a focus of the most important research contents in the field of noise control.In order to further study the influence of the vehicle sound quality on the acoustic comfort for the consumer,this paper takes the vehicle door-slamming sound quality as an example.In this paper,the followings are the mainly content:1.This paper collects sound signals of 25 different levels of the vehicle door-slamming by acoustic artificial,and then chooses suitable psychoacoustic parameters,including loudness,sharpness,roughness and fluctuation.The evaluation results of sound quality are obtained by combining with the mathematical model of psychoacoustic parameters.In order to get the subjective evaluation value,the research on subjective evaluation experiment is carried out based on the vehicle door-slamming sound signal.The research on correlation between the subjective evaluation value and objective evaluation parameters is to demonstrate the contribution degree of the objective parameters.2.Based on BP neural network,a theoretical model for evaluating the quality of the vehicle door-slamming sound is established.In order to achieve accuracy,prediction model of the BP neural network is built by applying the MATLAB neural network toolbox to evaluate the subjective evaluation value and objective evaluation parameters of the vehicle door-slamming sound quality.The advantages of neural network and multiple linear regression method are analyzed in prediction accuracy and model stability.The results show that the subjective evaluation value of vehicle door-slamming sound quality of prediction model of neural network is more accurate,so it can reduce the evaluation time and the cost of evaluation for the vehicle door-slamming sound quality subjective evaluation.3.The research on optimization design of a vehicle door-slamming sound quality is carried out based on neural network prediction model.The maximum point of relative displacement of the door closing process is obtained by the analysis of the finite element model of the vehicle.The point is simulated and optimized.Door-slamming tests were carried out based on simulation.By comparing the simulation results with experimental results,it verifies that the simulation is accurate.The subjective evaluation of optimized vehicle is obtained by subjective evaluation of door-slamming sound quality prediction model.The reliability and validity of the prediction model of the vehicle door-slamming sound quality is verified.