论文部分内容阅读
大型回转机械的故障诊断是一件十分复杂的工作。本文提出一种基于双向联想记忆模型的故障诊断专家系统,运用专家经验,综合故障信息进行深层推理。双向联想记忆知识存贮的矩阵形式及输入与输出层间的异联想功能为专家系统的知识获取与表征、不确定推理等提供了全新的途径,它在大型回转机械故障诊断专家系统中的实践充分显示了其高度的容错性、鲁棒性、实时性和自适应性。
Large rotary machinery fault diagnosis is a very complex task. This paper presents a fault diagnosis expert system based on bi-directional associative memory model, using expert experience and comprehensive fault information for deep reasoning. The matrix form of bi-directional associative memory knowledge storage and the heteroskedastic function between input and output layers provide a new way for expert system’s knowledge acquisition and characterization, uncertain reasoning and so on. Its practice in large-scale rotary machinery fault diagnosis expert system Fully demonstrated its high degree of fault tolerance, robustness, real-time and self-adaptability.