Design And Realization Of The System Fault Diagnosis

Test equipment was used to complete a variety of integrated circuits, circuit boards and other components of the automatic detection, to ensure product quality and normal use of great significance. Fault detection and diagnosis technology is an integrated technology, it involves many disciplines door. Such as modern control theory, signal processing, pattern recognition, computer engineering, artificial intelligence, electronics technology, applied mathematics, mathematical statistics and the corresponding applied science. Its process of development, has experienced equipment failures from the early detection of a variety of simple means, to rely on the experience accumulated in the diagnosis process, the further development of the current knowledge-based intelligent fault diagnosis technology, which drew extensively on a variety of related disciplines an effective way to and means. Set up test equipment for Intelligent Fault Diagnosis System for the existence of structurally complex test equipment, diagnostic mission fragmented and immature diagnosis theory, diagnostic knowledge difficult to access, diagnosis and other factors of uncertainty. These factors make its Intelligent Fault Diagnosis System is extremely tedious and difficult job.This article proposed the some problem the traditional maintenance and repair department can not resolve by using multimeter, oscillographs, or portable computers and problems of the simple tools and shortage of peoples' experience in the formerly fault detecting,considering the actual application,the article gone deep into the analysis of the fault diagnose technique. The works are mainly on following aspects.Firstly, the paper described the design of the fault detection and Diagnosis. It collected and coordinated the malfunction phenomenas and detecting experiences. Secondly, sample of the fault detection based on the neural network, especially on the BP neural network, was studied. The BP model of the COM was built up, after the training by the BP improved studying algorithms which momentum was introduced into, the sample was convergent successfully and had high identification presion. Finally, software of the singlechip were studied and designed detailedly, including COM,network communication and CAN bus communication. Preparation, including network card, COM, CAN card, such as board-level fault detection software.