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 regulation for unit safety and economy has important significance. Using partial least squares regression combined with actual unit operation data, the reheat enthalpy rise, ie, the endothermic capacity of steam per unit volume, was modeled. The influencing factors of reheat enthalpy rise were analyzed, and intermediate variables that were not introduced but had a direct effect on reheat steam temperature were constructed. In the early stage of modeling, the modeling data were screened through steady state conditions and uniform design, and the modeling accuracy of various data screening methods was compared. The results show that the reheat steam temperature is affected by the comprehensive changes of many factors and has a nonlinear relationship. The modeling of the reheat enthalpy rise can better reflect the nature of the reheat steam temperature change. Selecting different sample data pairs The accuracy and stability of the model have a certain impact; the established partial least-squares regression model can overcome the multiple correlations between variables, to obtain easy to interpret the statistical model, quantify the existing operating data to enable operators to quickly grasp the system characteristics, It also provides a data foundation for the application of advanced control strategies.