Research on Intelligent Engine Fault Detection Method Based on Machine Learning

To diagnose the engine fault quickly and effectively, we purposed a method to engine diagnosis, based on Twin Support Vector Machine. This method utilized five exhaust gas parameter values of HC, CO, CO2, O2, NOX and normalized them. Then it took these data as feature vector for test and train in Twin Support Vector Machine classifier, so as to achieve the purpose of identifying fault categories. The experimental results shows that twin support vector machine have better effect than Neural Network or Support Vector Machine, and the training speed is faster. In the case of small sample data, the accuracy rate of fault diagnose can reach 97.6%, which can effectively describe the complex relationship between the changes of vehicle exhaust components and engine default.