Application of neural networks in diagnosing engine faults based on vibration signals

Abstract

This paper investigates and applies artificial intelligence (AI) to improve the monitoring and diagnosis process of electrical engine faults based on vibration signals. The research aims to build a model to collect sample data from engines and utilize three different AI networks in this study, including YOLO (You Only Look Once), Resnet (Residual neural network), and SVM (Support Vector Machine). By applying these models to independently identify faults using the common input signal of vibration, particularly focusing on bearing-related faults in engine systems, the paper concentrates on exploring various faults. The experimental results presented in the paper demonstrate the accuracy of using these networks in diagnosing engine faults and provide important insights into the accuracy and practical applicability of AI networks in the field of industrial equipment maintenance and management.