Research On Sound Quality Of Vehicles Based On Transmission Path Analysis

With the rapid development of the times,the domestic passenger car population is becoming increasingly saturated,and people’s demand for cars is more rational.Today,car riding comfort has become an important consideration for people buying cars.This puts forward higher requirements for the adjustment of the NVH performance of vehicles,and the issue of sound quality has increasingly received the attention of researchers from major OEMs and universities.Establishing a model of the influence of noise sources on the sound quality of interior noise,is of great significance to improving the comfort of the interior acoustic environment and the competitiveness of vehicle products.According to this demand,the purpose of this paper is to study the influence of the structure and air noise transmission path on the sound quality of the vehicle by means of transmission path analysis.The global search capability of the Genetic Algorithm(GA)is used to optimize the parameters of the main contribution paths,thereby effectively improving the sound quality level in the car.Based on the basic principles of transmission path synthesis,this paper first tested the excitation signals and transfer functions of each path and synthesizes TPA(Transfer Path Analysis),to establish a vehicle noise transfer model that can identify the noise contribution of each path.By sampling multiple cars,a database of noise samples inside the car were established.Objective psychoacoustic parameters such as loudness and sharpness of the sample were calculated,and the organization staff scored the sample based on the evaluation index of irritability.Based on this,three kinds of algorithms that can be used for prediction were introduced: multiple linear regression,BP(Back Propagation)neural network,and GA-BP neural network.A sound quality prediction model with objective psychoacoustic parameters as input and irritability as output was established.By comparing the prediction accuracy of the three models,a more accurate GA-BP neural network model was determined for the sound quality prediction in this paper.Subsequently,the sound quality contribution factor(SQCF)was established to clearly reflect the sound quality contribution of each path.Through the horizontal comparison of the SQCF values of each path in the same working condition and the longitudinal comparison of the SQCF values of different working conditions in the key path,the left and right mount Y-direction structural paths of the engine mount system under low-speed conditions,and the air path of the engine’s rear surface radiated noise under high-speed conditions were determined as key optimization paths.Then,for the critical structural path,based on the analysis of the uncoupled motion of the suspended elements corresponding to the two structural paths that have a greater impact,through a two-level optimization strategy,first set the target irritability value,and then used the genetic algorithm to search for the corresponding optimal path transfer function.Finally,by matching with the static stiffness of the suspension element,the corresponding optimization value and transfer function of the static stiffness of the suspension element were obtained.Aiming at the critical air path,the sound insulation design optimization was carried out at the front panel of the prototype car by experimental optimization.By perfecting the laying position and thickness of the sound insulation material,the target transfer function was approximated.The research results show that the optimization for a single critical path can improve the overall level of sound quality in the car,while reducing the contribution factors of sound quality under their respective key conditions.The comprehensive optimization results of the three paths effectively improved the overall sound quality of the car under each research condition.It shows that the research in this paper can provide an effective reference and new ideas for the development of interior sound quality.