多目标进化算法在电站锅炉燃烧优化控制系统设计中的应用  

Multiobjective evolutionary algorithm applied in design of combustion optimization control system of utility boilers

作　　者：饶苏波 

机构地区：[1]广东省粤电集团有限公司,广州510630

出　　处：《广东电力》2006年第4期11-15,共5页Guangdong Electric Power

摘　　要：构造了一种基于最小二乘支持向量机和多目标进化算法的锅炉燃烧优化控制系统,通过从电厂分散控制系统(DCS)上采集数据,利用最小二乘支持向量机对锅炉燃烧特性建模,并通过样本的机器学习,提出了以锅炉效率与NOx排放量为组合的锅炉燃烧多目标优化模型,采用基于Pareto最优概念的多目标进化算法实现运行工况寻优,根据模糊集理论在Pareto解集中求得满意解,获得锅炉燃烧优化调整方式。Power plant operation is confronted with two requirements to reduce its operation cost and to lower its emission. This paper presents a research on optimized system design for high-efficiency and low-emission combustion of utility boilers. Multiobjective evolutionary algorithm （MOEA） is employed to solve these multiple and conflicting objectives and perform a search to determine the optimum solution of the least square support vector machine （LS-SVM） model, which is used to set up a boiler combustion response property model for NO, emission and efficiency, so as to obtain currently optimum combustion adjustment mode of boiler. Simulation and theoretical analysis show that the proposed optimal method may meet the two requirements to reduce operation cost and to lower emission.

关 键 词：燃烧优化 NOX排放 支持向量机 多目标算法 

分 类 号：TK323[动力工程及工程热物理—热能工程]