A Study On Internal Combustion Engine Abnormal Fault On-line Diagnosis Based On Neural Network And Fuzzy Technology

Internal combustion engine is a widespread power equipment, the operation of which influences production of many departments. Making fault diagnosis of internal combustion engine may enhance safety and reduce pollution, fuel consumption and maintenance costs. Due to the fact that the structure of internal combustion engine is complex, fault diagnosis cannot be made in a single way and human intelligence diagnosis technique making use of information interinfiltrate and neural network and fuzzy logic and expert system must be adopted in order to obtain high accuracy and real time.The research of this paper is based on the latest research of neural network and fuzzy theory to resolve internal combustion engine thermal process intelligence fault diagnosis. The major research contents are as follows.First, the state of art internal combustion engine fault diagnosis both at home and abroad has been summarized, and the existing problems have been pointed out.Second, owing to the wide application of the expert system, there is also a variety of models applied for combining neural network, fuzzy technology and expert system in different fields. This paper has made a research on internal combustion engine fault diagnosis-oriented combining model theory and implementing method.Third, based on the research of neural network theory and problem of the BP algorithm, a new method of the accelerating convergence rate and resolving part minima dot has been proposed.Fourth, after the characters of neural network, fuzzy technology and expert system were analyzed three models about combining expert system and neural network and their respective applied condition have been proposed. The methods of using fuzzy knowledge in expert system and combining neural network and fuzzy logic have also been discussed.Fifth, on the basis of analysis on the characteristics of internal combustion engine fault diagnosis problem, internal combustion engine fault diagnosisstrategy based on abnormal thermal process is systematically studied. Fault tree method is applied to internal combustion engine fault diagnosis.Sixth, because internal combustion engine fault diagnosis is both crucial and difficult problem, and the current method only obtains part miscellaneous information, this paper deals with inspecting temperature, pressure, flow volume and thermal process, and realizes fault sampling. It enhances accuracy, reliability and practicability.Finally, on the basis of the theoretical research results of this paper, the 8th chapter has designed and developed a practical system "Based on neural network and fuzzy technology internal combustion engine thermal process fault diagnosis expert system".