Study On The Intelligent Fault Diagnostic System For Diesel Engine

A diesel engine is a type of complicated reciprocating power machine. Its complicated structure and components make fault diagnosis very difficult. Therefore, study in diesel engine diagnostic techniques have being concentrated on. Diesel engine diagnostic techniques, based on many subjects, are synthesis techniques, which can help identify the real-time technique state of the recognition unit and predict the abnormal breakdown through analyzing and processing. How to collect the fault tag information more effectively and quickly and how to establish the fault criterion have become the essential parts of the research work.At present, performance parameter, vibration noise parameter and oil analysis parameter are used to monitor the engine conditions. However, each individual monitoring method can’t manage or utilize information from multiple sources and dimensions synthetically. As a result, any motoring means, to some degree, lacks accuracy, reliability and practicability. In this paper, the writer just suggests that the engine performance parameter, vibration noise parameter and oil analysis should be taken into consideration together and then corresponding measures be adopted in order to bring the engines into full play.The paper begins with an introduction to the purpose of the present study and its significance. And then it provides an overview of current situations, existing problems and developing tendency of diesel engine diagnostic techniques. The paper goes on with the primary coverage of the research work. After that, Cummins Diesels are taken as examples. Based on the bench testing system about condition monitoring, design theory research on fault diagnosis testing system has been done. Testing experiments on Cummins Diesel engines have been made, during which quantities of data have been collected. Then, based on the analysis of the test data, characteristic parameters with high sensitivity are chosen as main factors to score models, which have got better corresponding relationships with engine conditions. The weight of each factor is fixed in the comprehensive evaluation and membership function between each factor and different engine state is built on the basis of mathematics statistics and fuzzy theory. Thus, based on integrated decisions in fuzzy inference, models to evaluate engine conditions are set up. The accuracy and practicability of the model is tested with specific data. Meanwhile, neural network theory is combined with fuzzy theory; algorithm is improved with the application of BP neural network and Elman neural network. Engine condition classifier design is just based on neural networks. At the end of the paler, with the combination of fuzzy clustering analysis techniques with artificial neural networks ,the diesel engine fault diagnosis on fuzzy clustering analysis is studied and the fault diagnosis system on neural network is developed which is based on the rough set theory.