摘要
为了减少故障特征集的维数,降低流程系统故障诊断知识库的复杂程度,本文将基于二进制粒矩阵的粒约简算法引入到基于SDG模型的故障诊断中.以离心泵与液位系统为例,用粒语言来描述和表达SDG故障诊断模型中的元素,建立反映故障-征兆因果关系的决策表,进而对冗余属性及属性值进行约简,有效地约简了SDG诊断规则,提高了故障诊断的效率.
In this paper,Bit Granular Matrix-based granular reduction algorithm is introduced to the SDG model based fault diagnosis.With centrifugal pumps and liquid level system as an example,first,the element of SDG fault diagnosis model is described by granular language,a decision table which reflects the causality of faults and signs is established.Redundant attributes and attribute values are reduced by knowledge discovery algorithm of granular computing.The diagnosis rules of SDG model are reduced by the method which improves the fault diagnosis effectiveness.
出处
《北京交通大学学报》
CAS
CSCD
北大核心
2012年第5期73-77,共5页
JOURNAL OF BEIJING JIAOTONG UNIVERSITY
基金
国家自然科学基金资助项目(60795032)
关键词
故障诊断
符号有向图
二进制粒矩阵
粒约简
fault diagnosis
signed directed graph（SDG）
bit granular matrix（BGrM）
granular reduction
作者简介
张志军（1967-），男，山西太原人，副教授．研究方向为智能信息处理与模式识别．email：zhijunzhang888@126．com
谢刚（1972-），男，山西五台人，教授．email：xiegang@tyut．edu．cn．
