基于模糊神经网络的环管反应器故障诊断  

Fault Diagnosis of Loop Reactor Based on Fuzzy Neural Network

作　　者：闫志国[1] 郑明[1] 宣爱国[1] 吴元欣[1] 

机构地区：[1]武汉工程大学绿色化工过程省部共建教育部重点实验室湖北省新型反应器与绿色化学工艺重点实验室,湖北武汉430073

出　　处：《化学与生物工程》2011年第12期24-27,共4页Chemistry & Bioengineering

基　　金：国家自然科学基金资助项目(20906073);武汉市科技局科技攻关项目(200860423193)

摘　　要：环管反应器是丙烯聚合工艺中的主要反应设备之一。环管反应器运行过程中故障产生原因与征兆参数之间的影响关系比较模糊、复杂,不能准确地用数学关系式表述。针对此种情况,总结出环管反应器温度升高时可能的故障原因及征兆参数,并用模糊理论进行处理,再运用神经网络建立了环管反应器故障诊断系统。故障诊断实例表明,该系统具有较高的可靠性。Loop reactor is one of the primarily equipments of propylene polymerization. And relationship be- tween loop reactor operation faults and their symptoms are fuzzy, complicated and random, which are difficult to be clearly represented by math equations. For this problem, the probably causes of the fault and symptoms of loop reactor temperature rise was summarized and treated using the fuzzy theory,then a loop reactor fault diag- nosis system was set up using neural network. The practical application result showed the fault diagnosis system of loop reactor based on fuzzy neural network had a high reliability.

关 键 词：环管反应器 模糊理论 神经网络 故障诊断 

分 类 号：TQ053.6[化学工程] TP306.3[自动化与计算机技术—计算机系统结构]