数据挖掘技术在热电厂过程控制与优化中的应用研究.pdfVIP

第 19 卷第2 期 电 站 系 统 工 程 Vol.19 No.2 2003 年3 月 Power System Engineering Mar. , 2003 文章编号：1005-006X(2003)02-0048-03 数据挖掘技术在热电厂过程控制与优化中的应用研究 清华大学 卢 勇 徐向东 陈 明 摘 要：在电厂各种热力设备运行中，越来越多的现场数据被DAS 和DCS 系统存储到实时数据库中，在这些数据背后往往蕴涵着丰 富的知识，而这些知识的发掘和总结用人工分析的方式是无法实现的，因此提出利用数据挖掘技术服务于电厂热力设备运行控制和优化。 首先阐述了数据挖掘技术的相关知识，简述了其特征和方法；之后，结合目前电厂热力设备运行中较为重要的两个方面（动态过程监测 与控制、监督层静态优化）进行了详细论述，提出了一些新的控制策略和优化方法。最后总结了方法的优越性和可行性。 关键词：数据挖掘；数据库知识发现；多模型自适应控制；静态优化技术 中图分类号：TK32 文献标识码：A Use of Data Mining Techniques in Process Control and Optimization of Thermal Power Plant LU Yong, XU Xiang-dong Abstract: The scale of plant data collected in real-time database by DAS and DCS system in power plant is becoming larger and larger. There are abundant and valuable knowledge hidden behind these data. However, the knowledge discovered from these data through artificial analysis is impractical, hence a method utilizes the data mining technology to serve as a facility for process control and optimization of a power plant is proposed. The basic concept and knowledge of data mining is explained at first, then the application of data mining in two aspects, i.e. dynamic process control and supervisory steady state optimization, which are the most important issue of automatic operation of thermal power equipment is expounded. Finally, a system, which integrated the control and optimization module with the knowledge and rule-base obtained from operation database by data mining method, is proposed. Key words: data mining; knowledge discovery in database; multi-model adaptive control; steady state optimization 目前新建电厂或技术改造后的老电厂都安装了不同规 一个利用模式识别方法以及其他统计和数学方法，通过仔细 模、结构的DCS 系统，与老式控制系统相比，DCS 系统最