Study On Fault Diagnosis System Of Diesel Engine Based On Support Vector Machine

Diesel engine fault diagnosis technology is one of the more popular study direction for domestic and foreign experts in recent years, By analyzing and processing the state and historical data of the diesel engine to analyze and predict the future diesel engine fault conditions,in actual work diesel engine's malfunction is a small sample situation, so the paper lead into support vector machine, support vector machine show excellent performance, it advance fast in recent years, the paper do some basic research, wish help the Promotion of support vector machine in the area of diesel engine fault diagnosis technology.In this thesis introduces the following work1,First study the form of diesel engine failure and failure mechanism, detailed analysis of the main structure of the diesel engine, diesel engine with a chart format, the common cause of failure and failure are enumerated2,Establish signal analysis based on Vector machine, statistical learning theory and support vector machine, detailed analysis Machine learning problems, Empirical risk minimization, vapnik-chervonenkis dimension and Structural risk minimization principle, optimal hyperplane surface, kernel function and slack variable3,Use analytical methods to analyze the collected signal, introduction Support Vector Machine Fault Diagnosis in the theory and application, focusing on the full support vector machine multi-class support vector machines and multi-class support vector machine combined the two classification methods, using LIBSVM software and programming of the SVM in MATLAB Fault diagnosis on diesel engine fuel system applications.4,Established a marine diesel engine fault diagnosis system, I develop fault diagnosis system,the paper use vb language realize diesel engine fault diagnosis system and detailed show the realize process.