基于多模块合成的故障诊断智能系统推理机的研制  

Development of a Inference Engine for a Fault Diagnosis Intelligent System Based on Multi-module Integration

作　　者：谈理[1] 刘谨[1] 叶赛蓬[1] 梅丽婷[1] 

机构地区：[1]上海大学

出　　处：《中国机械工程》2006年第6期621-625,共5页China Mechanical Engineering

摘　　要：针对连续性生产线的特点,指出其故障诊断专家系统推理机不宜选用单一模块的形式。结合某钢铁厂的棒材生产流水线故障诊断专家系统研制,在分析连续性生产线的故障特点的基础上,介绍了多模块合成推理机的构造。推理机由基于神经网络的实时诊断模块、基于模糊理论的自动诊断模块和基于规则的人机交互诊断模块有机结合而构成。Aimed at the features of continuous production line, this paper presented that it was unsuitable to use unitary module for the inference engine of a fault diagnosis expert system. Combined with the development of fault diagnosis expert system for steel bar production line in a steel mill factory, based on analyzing the fault features of continuous production line, the composition of inference engine with multi-module was introduced in detail. The inference engine is composed of real-time diagnosis module based on neural network, auto-diagnosis module based on fuzzy and human-computer interaction module based on inference rules.

关 键 词：故障诊断 多模块合成 推理机 人工智能 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]