| Wind buffeting noise is a kind of aerodynamic noise with low frequency and high intensity,which is formed by opening windows at high speed.Affected by the wind buffeting noise,drivers and passengers can feel irritable and tired,especially when the side window is opened.The traditional single sound pressure level evaluation index can not fully describe the impact of wind buffeting noise,and sound quality analysis can describe the multi-dimensional characteristics of noise,which has become a hot topic for researchers.However,the sound quality evaluation of wind buffeting noise is lack of perfect analysis and optimization methods.In the process of wind buffeting noise quality analysis,it is often necessary to obtain some wind buffeting noise data with fixed side window opening.Traditional numerical simulation requires a lot of modeling and simulation work,which is inefficient.At the same time,in the process of wind buffeting noise optimization,it is essential to build a high-precision sound quality prediction model,and the proper prediction algorithm directly affects the optimization effect.Thus,this thesis studies from two parts: improving the productivity of sound quality analysis and improving the effect of sound quality optimization.The main contents of this paper are as follows:1.Based on the laying method and sliding grid method,the numerical simulation of the both side window with continuous opening is realized.The accuracy and efficiency of the dynamic grid method are verified by comparing the real vehicle road test results.2.The objective and subjective evaluation of vehicle wind buffeting noise quality are carried out through the results of numerical simulation.The loudness,sharpness,fluctuation and roughness are calculated,and their changing rules are analyzed.The subjective evaluation experiment of wind buffeting noise annoyance is excuted by using grade scoring method.3.Taking objective evaluation results as input and subjective annoyance as output,a support vector regression(SVR)sound quality prediction model is built.Genetic algorithm(GA)is used to solve the optional parameters of SVR,and a GA-SVR wind buffeting noise sound quality prediction model with higher accuracy is built,which provides a basis for the optimization of wind buffeting noise.4.Based on the generation princle of wind buffeting noise,the subjective annoyance is regarded as the optimization goal.A jet structure is designed to optimize the sound quality of the maximum annoyance of the left rear window when the vehicle speed is 110 km/h.Taking the jet angle,velocity and thickness as the optimization parameters,the multi Island genetic algorithm is used to optimize the jet structure.Through the above optimization process,a better optimization effect of sound quality is achieved. |
