Diagnosis Technology Of The S&r Of Automobile Seats Based On Feature Extraction

In-vehicle noise is one of the important factors influencing customers’ judgment of car quality and durability.Reducing in-vehicle noise is the main focus of car manufacturers in improving the customer satisfaction and market leadership in the highly competitive automotive market.With the advancement of vehicle sound and vibration engineering technology,the overall noise level of the vehicle has continued to decrease,causing occupants to pay more attention to intermittent noise such as squeak,and rattle(S&R)of the vehicle.Traditionally,manufacturers’ evaluation of vehicle products from the perspective of S&R relies heavily on physical testing and engineering experience,that is,subjective evaluation tests to detect abnormal noises and propose control schemes.Although this method can comprehensively evaluate abnormal sounds and can well reflect the owner’s real car experience,it is too dependent on the experience level of the engineer.This method also has poor consistency,high time cost,and low repeatability A series of questions.In order to solve these problems,we have developed a set of objective diagnostic solutions for abnormal sounds in car seats.In practice,the sound signal at the right ear is collected and processed to find the abnormal sound characteristics to achieve the objective detection objective,and the localization of the abnormal sound source is performed through the comparison of multiple near-field position psychoacoustic parameters.It is verified that the method is more accurate and efficient than subjective evaluation and can greatly reduce labor costs.The main research contents of this article are as follows:(1)This paper proposes a method for generating universal seat abnormal noise load spectrum and verifies its effectiveness.Select three types of pavement suitable for seat abnormal noise detection,drive 9 types of vehicles to collect the load spectrum of each pavement,study the characteristics of the load spectrum and find that the frequency range of the same pavement load spectrum is basically the same.General load spectrum of the test.Comparing the results of the bench test and the road test under the excitation of the universal load spectrum,the validity of the universal load spectrum is verified,so that it is not necessary to collect road spectrum for each vehicle in the project,reducing costs and improving efficiency.(2)A time-frequency analysis method based on EMD and WVD was established to detect abnormal seat sounds and effectively suppress the problem of cross terms generated by WVD transformation.Analyze the acoustic signal collected at the right ear,and detect the characteristic information of the abnormal sound on the time-frequency spectrum chart.According to these characteristics,the purpose of abnormal sound detection is achieved and the time point at which the abnormal sound occurs.Combined with the results of the subjective evaluation,the abnormal noise is gradually eliminated,and the abnormal noise characteristics of the time-frequency spectrum graph after disappearance prove the accuracy of the detection.(3)Propose seat abnormal sound source localization based on psychoacoustic parameters of near-field acoustic signals.After the time period of the abnormal noise is determined by time-frequency analysis,the near-field position N10 loudness and S10 sharpness of each time period are compared to determine the seat abnormal sound occurrence position during this time period,and the abnormal sound is completed.Source positioning.Later,the success rate of positioning was verified with the support of a large number of experiments.