数据驱动的BFG/Coal混烧锅炉热效率建模与优化操作  

Thermal Efficiency Modeling of a BFG(Blast Furnace Gas)/Coal Blended Combustion Boiler Driven by Data and Its Optimized Operation

作　　者：王建国[1] 程华利 肖前平 马世伟[1] 

机构地区：[1]上海大学机电工程与自动化学院上海电站自动化技术重点实验室,上海200072

出　　处：《热能动力工程》2014年第5期544-547,598,共4页Journal of Engineering for Thermal Energy and Power

基　　金：国家自然科学基金资助项目(61171145)

摘　　要：基于实际操作数据提出了一种数据驱动的混烧锅炉热效率建模与优化操作策略。首先,过程输入变量转换为具有物理统计意义的3个派生变量;然后引入NNG(非负绞杀)变量选择算法和ARIMA(自回归积分滑动平均)校正方法,给出了一种自适应建模方法;最后提出了基于模型的混烧锅炉的优化操作策略,并将之应用于锅炉优化操作。研究结果表明,采用基于模型的优化操作策略的混烧锅炉具有较大的节能空间,热效率的平均提升量为0.32%。Based on the practical data for operation,put forward was tactics for thermal efficiency modeling of a blended combustion boiler driven by data and its optimized operations. Firstly,the process input variables were converted into three derivative variables of physical statistical significance. Then,the non-negative garrote( NNG) variable selective algorithm and the self-regressive integral moving average( ARIMA) correction algorithm were introduced and a self-adaptive modeling method was given. Finally,the tactics for optimizing the operation of blended combustion boilers based on the model in question was put forward and used for optimization of operation of boilers. It has been found that the blended combustion boilers using the tactics for optimizing the operation based on the model enjoy a relatively big energy-saving space and the average increase of the thermal efficiency is 0. 32%.

关 键 词：多燃料锅炉 热效率 NNG 变量选择 数据驱动 

分 类 号：TK223.7[动力工程及工程热物理—动力机械及工程]