摘要
针对工业循环冷却水出现的腐蚀与结垢对管道设备的影响,从国内石化公司提取了相关的水质和工艺数据,对数据进行动态主成分分析处理后得出对水质腐蚀及结垢影响较大的6种因素,同时基于非线性自回归模型(NARX)神经网络和模糊逻辑控制方法建立腐蚀速率和黏附速率预测模型,并设计了工业循环水智能决策系统软件。该软件不仅能对水质腐蚀速率及黏附速率进行预测,还能根据预测结果给出相应的处理意见,经实际运行,缩短了运行预测周期,提高了运行效率。
Based on the influences of corrosion and scaling occurred to industrial circulating water on pipeline equipment,the data regarding water quality and technologies are abstracted from a petrochemical company in China,and six kinds of factors that have greater influences on water quality and structures obtained,after the data have been analyzed by dynamic principal component analysis. In addition,the prediction model of corrosion rate and adhesion rate has been established based on NARX neural network model and fuzzy logic control method,and the software of intelligent decision system of industrial circulating water designed. This software can not only predict the corrosion rate and adhesion speed of water quality,but also provide corresponding treatment advice,according to the prediction results. Through practical operation,the predicted operation cycle is shortened and the operation efficiency improved.
作者
冯胜
刘明远
冯旭
Feng Sheng;Liu Mingyuan;Feng Xu(Tianjin Branch of CNOOC Ltd.;Key Laboratory for Control Theory ＆Applications in Complicated Systems,Department of Electrical Engineering,College of Automation,Tianjin University of Science ＆ Technology)
出处
《工业水处理》
CAS
CSCD
北大核心
2018年第3期69-72,共4页
Industrial Water Treatment
关键词
工业循环水
主成分分析
NARX神经网络
智能决策
industrial circulating water
principal component analysis
NARX neural network
intelligent decision
