| In recent years, with the increasing number of the vehicle population in the world and with the increasingly improved structure of automobile, traditional technologies of auto repair have become limited. In view of this, many specialists and experts have started to study intelligetized diagnostic technology of auto failures.This paper makes a detailed study on the development of diagnostic technology of auto failures both home and abroad; introduces the theories on the technology, the history and features of neural network as well as the connections between neural network and diagnostic technology of auto failures; points out the idea that the diagnostic technology based on neural network is a prospective technology. What’s more, the paper also states the internet mode, the common learning rules and the working process of neural network.From view of application, this paper analyzes the BP network model, BP network rules, BP network structure designs, and the selection rules for initial value as well as the diagnostic process. The paper also presents the disadvantages of applying BP networks to failure diagnosis and its corrective methods. An introduction to ELMAN network is also made in the paper.With MATLAB7.0.1data-calculation software as the assistance, this paper mainly studies the failure diagnosis emulation of auto engines, which includes the access to the sample set, the design of the network, the pretreatment of information. It also makes a study on the trouble shooting of the misfire of an engine with the technology of BP network and ELMAN network. A comparison between BP network and ELMAN network is made and studied, which reaches a conclusion that the error amount of BP network is smaller than that of ELMAN network. But due to the feedback effect of ELMAN network, its error curve is smooth and tends to be a straight line, neural network can respond automotive fault diagnostic work fully. This article also use the powerful data calculation functions of MATLAB and the powerful interactive features of VB, using ActiveX technology between VB and MATLAB to achieve the seamless integration, completing the development of fault diagnosis system, With the help of VB and SQL’s integration to achieve the fault system database modifications and updates. |
