The Research And Implementation Of PCB Fault Diagnosis Based On Magnetic Image

PCB fault diagnostic technology based on radiated magnetic-field image is a new type of non-destructive diagnostic techniques of circuit board, which have been developed in recent years, to rapidly detect and locate the circuit board fault and improve the reliability of the circuit board. This paper presents an extensive research on related technologies of the circuit board fault diagnosis, and designs a circuit board fault diagnosis system based on magnetic image, At last, makes a detailed analysis and testing. The main work in the paper involves:Firstly, the paper briefly introduces the basic principles, method and realization process of the diagnosis of the circuit board based on the magnetic field image; In view of the characteristics of the magnetic Image, the paper gives a detailed research of the digital image processing technology on Image filtering, image enhancement, etc; On the basis of the in-Depth analysis of the wavelet transform's mechanism, advantages and disadvantages,we completes the magnetic image of multi-level wavelet decomposition, and achieves the extraction of wavelet energy feature; A new diagnostic rule is developed, that is, the improved BP algorithm, for training the neural network and the circuit board fault diagnosis.Secondly, from the main task of the circuit board fault diagnosis system design, based on the related technologies, combining with the two software's characteristics of MATLAB and Access, the paper designs a PCB fault diagnosis system based on radiated magnetic-field image. This system combines the use of wavelet analysis and BP neural network method, using wavelet analysis to extract the wavelet energy feature, and then using BP neural network to learn and discriminate. At the same time, gives some important fault Diagnosis software code. Finally, simulation experiment is completed, taking a specific circuit board in a variety of fault mode. Using Ansoft software, we can get the different circuit board failure modes of the magnetic field image; using wavelet packet, we can decompose and extract the wavelet energy feature of the magnetic image, then respectively as the sample feature and the test feature vector of the circuit board fault diagnosis system to demonstrate the feasibility and reliability of the system using a specific example. The results show that the designed circuit boards fault diagnosis system based on field mapping can effectively diagnose the circuit board, so the paper provides a new way for the circuit board fault diagnosis.