A testability modeling method for analog circuit fault prediction

Abstract

This paper puts forward a testability modeling method for analog circuit fault prediction. It firstly gets the grey correlation entropy of each test point in the analog circuit. Then it treats each grey correlation entropy as a correlation coefficient to form the dependency matrix of testability. After that, according to the dependency matrix we get, the paper uses the method of PSO (Particle Swarm Optimization) to optimize test points, which will optimize the prediction indexes of the analog circuit. Last, the method will be applied in analog circuits. Application suggests that the method is very helpful for analog circuit fault prediction.