Sound Quality Prediction And Control Technology Of Exhaust Noise

With the progress of society and the development of science and technology,the vehicle noise level standard can not meet the requirements of the consumers,the research and improvement of vehicle sound quality has become the mainstream direction in the field of vehicle noise.Vehicle exhaust noise is one of the main noise sources of vehicles,and it is of great significance to study the sound quality to improve the NVH characteristics of the vehicle and the competitiveness of the vehicle market.The purpose of this paper is to study the influence of structural parameters of exhaust muffler on exhaust sound quality,and explore the design method to improve the sound quality of exhaust noise.Therefore,the exhaust noises in steady and unsteady conditions of several domestic passenger cars were chosen to be research objects,to establish sound quality prediction models of steady and unsteady exhaust noises,and based on the model,the simulation of acoustic characteristics of the muffler and the sound quality design of the exhaust noise were carried out.Firstly,in this paper,the subjective evaluation test of sample noises was carried out by the paired comparison method,and the satisfaction scores of the sample noise were obtained.At the same time,the psychoacoustic parameters of the steady and unsteady samples were calculated based on the Zwicker steady-state and time-varying algorithm,and the correlation between the objective parameters and the satisfaction scores were analyzed.Then,in MATLAB,the BP neural network was optimized by genetic algorithm(GA),which was used to establish the sound quality prediction model of steady exhaust noise,and the model has high accuracy,which can be applied to research and predict sound quality of steady exhaust noise.Secondly,methods such as the wavelet decomposition and the WVD distribution optimized by RNR(regularization nonsteady-state regression)were introduced.And the sound quality parameter SQP-RW(Sound quality parameter base on RNR-WVD)proposed in this paper was established.It was concluded that the parameter SQP-RW has the ability to reflect the fluctuation characteristics of signals,but simply using this parameter as the input of the prediction model can not reflect the influence of loudness,sharpness and so on,and this distributed to the insufficient precision of the model.The roughness and fluctuation that characterizing the satisfaction of the unsteady noise insufficiently were replaced by the SQP-RW value,and the SQP-RW was combined with loudness,sharpness,A sound level and kurtosis to be the input of the model,and the GA-wavelet neural network was introduced,a prediction model of vehicle unsteady-state exhaust sound quality with high accuracy was built.In order to design and improve the sound quality of exhaust noise,the simulation method of sound quality was proposed and the feasibility was verified.Meanwhile,the influence of structural parameters such as expansion ratio,cavity length and perforation rate on the sound quality of exhaust noise was studied based on the sound quality prediction model.Finally,the optimization design of the three vacity target muffler was carried out.After optimization,the sound quality satisfaction is increased by 12.24% compared with that before optimization.Meanwhile,the pressure loss of the muffler is not increased,but it has decreased.It proves the feasibility of the optimization plan.The research can provide a reference for the sound quality research of vehicle noise and the optimization design of muffler for improving the exhaust sound quality.