基于GA-LSSVM直吹式制粉系统煤粉细度软测量  

Soft Measurement of Pulverized Coal Fineness in Direct Blowing Pulverized System Based on GA-LSSVM

作　　者：朱万进 岳峻峰[2] ZHU Wanjin;YUE Junfeng(National Energy Group Xuzhou Power Generation Co.,Ltd.,Xuzhou,Jiangsu 221100,China;Jiangsu Fangtian Power Technology Co.,Ltd.,Nanjing,Jiangsu 211102,China)

机构地区：[1]国家能源集团徐州发电有限责任公司,江苏徐州221100 [2]江苏方天电力技术有限公司,江苏南京211102

出　　处：《上海电力大学学报》2023年第5期472-478,共7页Journal of Shanghai University of Electric Power

摘　　要：针对燃煤电厂中煤粉细度难以测量的问题,提出了一种基于遗传算法优化最小二乘支持向量机(LSSVM)超参数的煤粉细度软测量方法。通过制粉系统特性试验,综合分析出影响磨煤机煤粉细度参数的主要因素。采用遗传算法优化支持向量机超参数方法构建煤粉细度软测量模型。将该方法用于特性试验,发现检验样本的均方根误差为1.7%,证明煤粉细度软测量方法有效,满足煤粉细度模型精确预测的需求。Aiming at the problem that it is difficult to measure pulverized coal fineness in coal-fired power plants,a soft measurement method of pulverized coal fineness based on genetic algorithm optimization of least square support vector machine super parameters is proosed in this paper.The main factors affecting pulverized coal fineness parameters of pulverized coal mill are analyzed comprehensively through the characteristic experiment of pulverized coal system.The soft sensor model of pulverized coal fineness is constructed by the superparameter method of support vector machine optimized by genetic algorithm.The method is applied to a practical case,and the root-mean-square error of the test sample is 1.7%,which proves the validity of the soft measurement method of pulverized coal fineness,and meets the demand of accurate prediction of pulverized coal fineness model.

关 键 词：煤粉细度 参数优化 最小二乘支持向量机 软测量 

分 类 号：TK37[动力工程及工程热物理—热能工程]