Application Research Of A Radial Basis Function Neural Network To Fault Diagnosis Of Electronic Ejection Engine

This paper introduces the process of the equipment developing and analyses the direction of theory research on automobile fault diagnosis, based on a large of the abroad and domestic information. First, the purpose and significance about this subject is discussed and it is dictated definitely that if neural network can be used in simplifying data stream of automobile diagnosing instruments the difficulties of automobile fault diagnosis will be decreased and maintainers' operating efficiency will be enhanced greatly. Second, this thesis presents some elementary knowledge about radial basis function neural network(RBF) and electronic control parts of EFI engine (electronic fuel injection engine) . Last but not least, taking incorrect idle speed of Jetta ATK engine as example, the sample sets of the symptoms and troubles have been designed and tested with VAG1552. Simultaneously, using a practical Neural Network Toolbox in MATLAB environment, the simulation calculations for fault diagnosis of EFI engine have been performed. In order to verify the accurate and the degree of fault mode, a large number of test calculations have been done through computerization. The simulation experimental results demonstrate that this diagnosis method for EFI engine is more feasible and successful than BP(Back-propagation).In addition, the mathematical model, structure and its main functions of a fault diagnosis simulating system for EFI engine are presented. And the trouble diagnosis approach, which based on RBF neural network, is introduced. The system is simple in construction and perfect in function. It is of a practical importance to develop the EFI engine performance fault analysis system and extend the functions for automobile diagnosing instruments, and it is also able to train maintainers to deal with troubles and to make out repairing strategies accordingly for the purpose of enhancing the utility and working life of an EFI engine.If this program can be solidified in diagnosing instrument, it will further develop to the direction of "foolization"...