作者：马平, 刘南南
作者单位：华北电力大学自动化系, 河北 保定 071003
关键词：声波测温;支持向量机;遗传算法;NOx排放
摘要：
因传统燃烧优化实验控制电厂NOx排放的方法很难满足复杂多变的燃烧工况,为更智能地对NOx排放进行监测和更方便地对其进行优化,对某电厂2#炉300 MW工况下NOx排放优化实验时的DCS内运行数据和声波测温系统内的温度分布数据进行采集。利用最小二乘支持向量机,以炉膛温度信息和其他影响NOx排放的主要因素为输入,以NOx排放浓度为输出建立NOx排放预测模型,在预测模型的基础上利用遗传算法对顶部4层分离燃尽风开度进行直接寻优,达到降低NOx排放的目的。结果表明:加入炉膛温度信息后的NOx排放模型准确度更高,遗传算法优化之后的NOx排放浓度显著降低,优化后参数更符合工程实际。
Modeling and optimization for NOx emission of power stations based on acoustic temperature measurement and LSSVM_GA
MA Ping, LIU Nannan
Department of Automation, North China Electric Power University, Baoding 071003, China
Abstract: Traditional methods to control NOx emissions through combustion optimization experiments can hardly meet complicated and changeable combustion conditions now. For more intelligently monitoring and better optimizing NOx emissions, data within DCS and temperature profile within the acoustic measurement system of furnace 2# in a power station under 300 MW working condition are collected during the optimizing experiment of NOx emissions. That is, a least squares support vector machine is used to create a prediction model for NOx emission based on furnace temperature information and other factors that affect NOx emissions as input value and NOx emission concentrations as output value. Apart from the model, a genetic algorithm is applied to optimize the opening of over fire air of four-layer separation on top so as to reduce NOx emissions. The results show that the NOx emission model is more accurate when furnace temperature information is added, and the NOx emission concentration is significantly reduced and the parameters are more suitable for engineering practice after the opening is optimized through the genetic algorithm.
Keywords: acoustic temperature;support vector machine;genetic algorithm;NOx emission
2016, 42(3): 118-122 收稿日期: 2015-07-23;收到修改稿日期: 2015-08-12
基金项目:
作者简介: 马平(1961-),女,湖南湘潭市人,教授,硕士生导师,研究方向为过程控制、火电厂单元机组控制和优化。
参考文献
[1] 魏辉. 燃煤电站锅炉低NOx燃烧优化运行策略的研究[D].上海:上海交通大学,2008.
[2] 周献. 声学法三维温度场重建算法的研究[D]. 成都:西华大学,2013.
[3] 周昊,朱洪波. 大型四角切圆燃烧锅炉NOx排放特性的神经网络模型[J]. 中国电机工程学报,2002(1):33-37.
[4] 顾民. 神经网络-模糊系统-支持向量机内在联系研究[D].成都:电子科技大学,2004.
[5] 张学工. 关于统计学习理论与支持向量机[J]. 自动化学报, 2000,26(1):32-41.
[6] NELLO C, JOHN S T. 支持向量机导论[M]. 李国正,王猛,曾华军,译. 北京:电子工业出版社,2005:76-88.
[7] SUYKENS J A K. Nonlinear modeling and support vector machines[C]∥ Proceedings of IEEE on Instrumentation and Measurement Technology Conference. Budapest:IEEE,2001:287-294.
[8] 姜静清. 最小二乘支持向量机算法及应用究[D]. 长春:吉林大学,2007.
[9] 魏恩宗. 燃煤锅炉低NOx燃烧技术及其试验研究[J]. 电站系统工程,2005,11(6):363-365.
[10] 韩建新. 燃煤锅炉NOx排放特性及燃烧优化[J]. 内蒙古科技与经济,2009,11(6):79-82.
[11] 徐磊. 基于遗传算法的多目标优化问题的研究与应用[D].长沙:中南大学,2007.
[12] 苟湘,周俊虎. 低NOx改造方案中煤粉再燃风喷口位置的选择[J]. 中国电机工程学报,2008,15(7):1-6.
