Fault Feature Selection Based on CA/ILS and Its Application in Chemical Process Fault Diagnosis

摘要：

In large scale chemical process industry,fault feature selection is an important part of fault diagnosis system.If all variables are collected for fault diagnosis,it results in poor fault classification because there are too many irrelevant variables,which also increase the dimensions of data.A novel optimization algorithm (CA/ ILS) that integrates cultural algorithms (CA) with iterated local search(ILS),is proposed to select the fault feature variables for fault diagnosis.In CA/ ILS algorithm,the population space of cultural algorithm adopts discrete particle swarm algorithm (DPSA).At the initial stage of the search,the belief space of Cultural Algorithms is applied to improve the performance of DPSA and guide the search toward the right direction.At late stage of the search,when the most promising regions of solutions are fixed,ILS replaces DPSA and the knowledge from the belief space is applied to enhance local search.The test of optimizing functions shows the proposed algorithm is valid and effective.The simulations on Tennessee Eastman process (TEP) show the developed CA/ ILS algorithm can find the essential fault feature variables effectively and exactly.And with fault feature selection,more satisfied performance of fault diagnosis is obtained.

关键词： cultural algorithm iterated local search fault feature selection discrete particle swarm algorithm.

作者: Hai-yan Huang Xing-sheng Gu

作者单位: Department of Automatization,East China University of Science and Technology,Shanghai,China

会议类型: 国际会议

会议名称: International Conference on Modelling,Identification and Control(模拟、鉴定、控制国际会议)

会议地点: 上海

会议语种:英文

在线出版日期: 2008-06-29（万方平台首次上网日期，不代表论文的发表时间）