基于偏最小二乘回归的锅炉再热汽温建模  

Modelling of Utility Boiler Reheat Steam Temperature Based on Partial Least Squares Regression

作　　者：刘吉臻[1] 曲亚鑫[1] 田亮[1] 常太华[1] 

机构地区：[1]华北电力大学自动化系,北京市昌平区102206

出　　处：《中国电机工程学报》2011年第11期99-105,共7页Proceedings of the CSEE

基　　金：国家863专项经费项目(2007AA04Z163);国家自然科学基金重点项目(51036002);北京市教育委员会共建项目专项项目~~

摘　　要：锅炉再热汽温调节对于机组安全性和经济性有着重要的意义。利用偏最小二乘回归结合机组实际运行数据对再热期望焓升-即单位流量蒸汽吸热能力进行建模。分析了再热期望焓升的影响因素,构建了现场没有引入但会对再热汽温造成直接影响的中间变量。在建模前期通过稳态工况以及均匀设计的方法对建模数据进行筛选,并比较了多种数据筛选方法的建模精度。研究结果表明,再热汽温受多个因素综合变化影响,且呈非线性关系,对再热期望焓升进行建模能够更好地反映再热汽温变化的本质;选择不同的样本数据对模型的精度和稳定性有一定的影响;建立的偏最小二乘回归模型能够克服变量间的多重相关性,得到易于解释的统计学模型,量化已有运行数据,使运行人员快速掌握系统特性,同时也为先进控制策略的应用提供了数据基础。Boiler reheat steam temperature control is of great significance for unit safety and economy. Considering the relative merits of historical data modelling and field test data modelling, the uniform design method was selected to filter the historical data, and a reheater heat absorption model combined with partial least squares regression was established. Compared with other data selection methods, the method was proved to valid. Also, a detailed analysis had been made, which quantifies the operating experience, could help the commissioning staff to master the system characteristics, and provide data foundation for the application of advanced control strategies.

关 键 词：锅炉 再热汽温 均匀设计 偏最小二乘回归 

分 类 号：TK223[动力工程及工程热物理—动力机械及工程]