Study On Intelligent Fault Diagnose System For Work Process Three-way Catalytic Converter In Vehicle

It is a complicated process for the working process of three-way catalytic converter with the features of multi-variable and non-linear, which concerns chemical reactions, heat and mass transfer, fluid flow, faults that deviating from normal condition are always occurred. But a good fault diagnosis expert system can do real-time fault diagnosis and alarm for three-way catalytic converter of vehicle according to the working status of it. It's conducive to improve the reliability and security of three-way catalytic converter by operation and maintenance personnel making the corresponding measures in time, Thereby improve the working life of three-way catalytic converter and ensure good quality of emissions in the after-treatment process.In order to achieve fast and efficient diagnosis on the happened or potential fault of the three-way catalytic converter working process, in this paper, develop a intelligent fault diagnosis expert system for three-way catalytic converter that based on neural network with the application of neural network control theory and expert system technology, the main work and innovation of this paper are as follows:(1) A air flow measurement model of intake manifold in gasoline engine of vehicle was established by fitting the function linked neural networks (FLNN) of rotation angle of throttle plate and the discharge coefficient of throttle plate, the results show that the intelligent measurement error of mass air flow(MAF) in gasoline engine of vehicle was reduced with the increase of the intake manifold pressure, the error is less than 5.0%, less expensive and more accurate.(2) To make sure the three-way catalytic converter of vehicle worked in a reliable and optimizing environment. The cluster controller was applied to control the air and gasoline flow of gasoline engine of vehicle based on the fusion of neurons– fuzzy reasoning.(3) Sub-networks of fault and sub-networks of fault module was established by the dual production rules, knowledge base was built by the trained connection power and the threshold matrix with the Leverberg_Marquardt algorithm that based on numerical optimization, and the knowledge diagnosis method of doing deep knowledge diagnosis by the results of neural networks (using shallow knowledge) was introduced.(4)Mixed reasoning on the faulty condition of the key parameters in three-way catalytic converter of vehicle was done according to the direction of forward and backward mixed inference by inaccurate inference method and heuristic search strategy, and the inference engine of neural network fault diagnosis expert system of three-way catalytic converter of vehicle was established.A neural network fault diagnosis expert system of three-way catalytic converter of vehicle was established according to the working process of three-way catalytic converter with Visual Basic6.0, ACCESS and other development tools. The results of a more than a half year commission of this system show that, the rate of diagnosing accuracy of this intelligent fault diagnosis system is higher than 85%.