Study on vehicle internal organization fault and performance detection based on mechanical vibration feature extraction

Abstract

Vehicle mechanical fault diagnosis is of great significance to ensure the safe and stable operation of vehicle mechanical equipment. The method of vehicle mechanical fault feature extraction based on independent component analysis is studied. Firstly, multi-domain features such as time-domain features, frequency domain features, and EEMD energy features of vibration signals are extracted, and a multi-objective optimization function containing the correlation, monotonicity, and robustness of each feature vector is introduced as the degradation feature evaluation function. Through the independent component analysis of the vehicle mechanical fault signals of different working conditions, the independent components of the signals of various working conditions are obtained, and these independent components contain some inherent characteristics of the vibration signals of the working conditions. Through the vehicle, internal organization fault, and performance testing test, the accuracy and effectiveness of the proposed method are proved, with high prediction accuracy and robustness, which has important guiding significance for engineering practice.