Fault Diagnosis In Analog Circuits Based On Neural Networks

With the fast developing of IC, in order to improve product's capability, reduce chip's area and expense, we should integrate digital and analog component into one chip. From the statistic number we can learn that there are 95 percent fault resulted from analog circuit, though only 5 percent of chip's area. Fault diagnosis of analog circuit always a difficult problem on IC industry. Neural networks, which are a typical representation of Computational Intelligence, provide a powerful way to diagnose faults of analog circuits.At the basis of the former accomplishment, I combine the electronic measurement with the artificial intelligence in the system's organizing and designing. Directed toward tolerance and nonlinear of analog circuit, this dissertation based on NN profoundly studies fault feature extraction and fault diagnosis methodology. Thesis's main work concentrates on two aspects as follows:1. On fault feature extraction for analog circuits.(1) Rough setEliminate redundancies and incompatibilities in samples, extract the key information, et al.(2) Principal Component AnalysisFeature parameters of the step response are compressed using principal component analysis (PCA). It has many good properties, such as simplifying the structure of NN, improving the training speed and fault coverage.2. Construct the structure of neural networkAnalysed the structure of neural network, evaluate the diagnosis result with different neural network structures and different parameters, confirm the structure in the practical application. The experimental results show that the proposed method can perform correct diagnosis in the linear analog circuits with tolerances.