基于混合推理机制和光谱分析的活塞压缩机磨损故障诊断设计  

Piston Compressor Wear Faults Diagnosis Based on Hybrid Resoniny Mechanism and Spectro-scopic Analysis

作　　者：王潜龙[1] 于庆峰[1] 李小明[1] 冯全科[1] 

机构地区：[1]西安交通大学,陕西西安710049

出　　处：《流体机械》2003年第3期26-28,共3页Fluid Machinery

摘　　要：建立了压缩机组磨损基于规则推理与人工神经网络混合推理的专家系统模型 ,应用该模型可以解决规则不足的问题 ,同时采用黑板机制实现神经网络推理与基于规则推理之间的转换。克服基于规则推理效率低下 ,知识获取困难的障碍 ,实现演绎推理 (基于规则推理 )和归纳推理 (人工神经网络 )的有机配合。The expert system model of compressor group which based on the rule based reasoning and the manual neural networks admixture reasoning is established. We applied the model to solve the absence of rule, at the same time , adopted blackboard framework to realize the conversion between the neural networks reasoning and the rule-based reasoning. These can conquer the obstacle of low efficiency of rule-based reasoning and the difficulty to gain knowledge, and achieve the cooperation between the syllogistic reasoning(rule-based reasoning) and the inductive reasoning(manual neural networks).

关 键 词：混合推理机制 光谱分析 活塞压缩机 磨损 故障诊断 神经网络 

分 类 号：TH457[机械工程—机械制造及自动化]