Research On Fault Detection Method And System Of Abnormal Sound In Automobile Intelligent Seat

As an important part of the car,the performance of the electric seat directly affects the safety and experience of the driver and passenger.Therefore,it is very important to test its quality before leaving the factory.Motor is the power component of regulating electric seat,which is easy to be damaged.It can be diagnosed by abnormal sound detection technology,which has a wide range of market demand.Therefore,in view of the problems caused by the noisy production environment of enterprises,such as the variable signal-to-noise ratio of acquisition signals,this research proposes a fault diagnosis method for abnormal sound of automobile electric seat,and develops a detection system to verify the proposed algorithm.The main work is as follows:A fault diagnosis system for abnormal sound of automobile electric seat had been developed.First of all,the system design requirements,the development of the system overall plan,the system by the upper computer and the lower mechanism.The lower machine includes:mechanical module,electrical control module,signal acquisition and transmission module,to complete the electric seat motion control and noise signal acquisition.The upper computer had been realized based on Lab VIEW software,including: initialization module,noise detection module,user login module,manual automatic module,signal communication module,system control module,printing and scanning module,to realize noise signal detection and result output.Based on the experimental data collected by the developed system,a fault diagnosis method for abnormal sound of automobile electric seat had been proposed.Firstly,the characteristics of the measured signal had been analyzed,and the least squares fitting method had been used to remove the trend item,and the high-frequency noise had been removed by the Vibration Mode Decomposition(VMD).Then,11 features including peak-peak value,mean value,mean square value,standard deviation,effective value,peak factor,pulse factor,waveform factor,margin factor,skewness factor and kurtosis factor had been extracted,and the feature selection had been carried out based on Principal Component Analysis(PCA).Finally,the classifier had been trained based on the Random Forest(RF)and the Probabilistic Neural Network(PNN)to achieve intelligent recognition of the anomalous signals.The results had shown that the performance of random forest classifier is the best in the recognition of different sounds,with an average accuracy of 95.11%.The proposed algorithm had been verified by the developed fault diagnosis system of electric seat.The method of graphic programming and MATLAB had been used to realize the preprocessing,feature extraction and well-trained classifier of different sound signals.Through the actual test,the proposed algorithm can realize the accurate identification of electric seat fault.