The BP Information Model Based on AI-ESTATE

摘要

This paper proposes a new fault diagnosis strategy for analog circuit after analyzing diagnosis method based on the BP neural networks and AI-ESTATE standard. At present, The artificial intelligence algorithms such as SVM and BP has been widely adopted in analog circuit fault diagnosis, but it is not included in the IEEE AI-ESTATE standard. So there is no normalization of BP information model, and the interchange formats of the BP information model is application-dependent. Therefore, it has no portability and cannot share diagnostic knowledge and data across different platform or application in analog circuit fault diagnosis. For these problems, the BP information model based on AI-ESTATE standard is proposed. And it not only extends the standard of AI-ESTATE, but also presents a normalization of the BP information model in analog circuit fault diagnosis. The simulation result shows that the proposed BP information model can share data among different platform, while maintain the same diagnostic rate compared with the traditional method.