Study On Fault Diagnosis Expert System Of Engine Based On Ant Colony-Neural Network

The engine is the power source for an automobile. It has a complicated structure and frequently operates under very adverse conditions, therefore the risk of engine faults is high. Generally engine faults account for more than 40% of all faults in an automobile. It is essential that the source of the fault be diagnosed accurately. It is also the pivotal task of fault diagnosis of the whole automobile. This paper is proposing a fault diagnosis expert system of engine based on ant-colony neural networks that can diagnose faults efficiently, accurately, rapidly and provide a new way of fault diagnostics for engines.This paper expatiates on the basic principles of fault diagnosis expert system, ant colony algorithm and neural network. To overcome the shortcoming of slow convergent speed and easy convergence to the local minimum points of BP algorithm, the way is proposed that ant colony algorithm is used to train the network and set up the model of ant colony-neural network. It is applied to the conventional expert system and makes the convergent speed and precision the fault diagnosis improved. The design of software of expert system regards Windows2000 as the platform, adopt language of Visual Basic and Access database to develop. It realizes such functions as fault diagnosis, knowledge acquisition, maintenance management, studying auxiliary, etc. by the friendly interface between human and computer. It is applied to diagnosis the fault such as engine cannot start and the effect is nicer.