结合人工神经网和符号推理的多重故障诊断

介绍了基于符号推理（基于规则推理、基于模型）的故障诊断方法和基于人工神经网的故障诊断方法，并分析了各自存在的局限性。提出了结合人工神经网和符号推理的多重故障诊断方法，给出了该方法的定义和算法。由于该方法能结合基于符号推理方法和基于人工神经网的方法的两者优势，因而能更有效地实现故障诊断。

Multiple Failure Diagnosis Based on Mixed Models

Failure diagnostic approach based on symbolic inference and failure diagnostic approach based on Artificial Neural Network(ANN) are introduced, and their limitations are analysed. A new approach which combines ANN and symbolic inference is presented, and its definitions and algorithms are given. As this approach combines advantages of both failure diagnostic approaches which are based on ANN and symbolic inference, it can realize failure diagnosis more effectively.