摘要
针对煤电运行数据挖掘程度不足、煤耗控制效果不佳等问题,设计了一种基于GRA-GWORF算法的煤电运行数据分析及优化算法。该算法以煤电运行数据为输入,利用GRA算法筛选影响煤耗的关键可控变量,再通过GWO改进RF算法的子决策树数量与选取特征数量两个参数。同时将完成筛选的关键可控变量数据作为参数优化后RF算法的输入,进而实现煤电运行方式的优化。算例分析结果表明,采用所提GRA-GWO-RF算法改善后,平均煤耗降低了13%以上,而RF算法仅下降约10%。因此相比于RF算法,该算法能够更为准确地优化煤电机组的运行方式,并提高其经济性。
Aiming at the problems of insufficient coal power operation data mining and poor coal consumption control,a coal power operation data analysis and optimization algorithm based on GRAGWO-RF algorithm is designed.The algorithm takes coal power operation data as input,uses GRA algorithm to screen key controllable variables that affect coal consumption,and then uses GWO algorithm to optimize the number of sub decision trees and the number of selected features of RF algorithm.The selected key controllable variable data is taken as the input of RF algorithm after parameter optimization,and finally the optimization of coal power operation mode is realized.The results of the example analysis show that the average coal consumption is reduced by more than 13%after the optimization of the GRA-GWO-RF algorithm proposed in this paper,while the average coal consumption is reduced by only about 10%after the optimization of the RF algorithm.It can be seen that compared with RF algorithm,GRA-GWO-RF algorithm can more accurately optimize the operation mode of coal power units and improve the economy of coal power units.
作者
张磊
熊健
金龙云
刘秉祺
方鑫宇
ZHANG Lei;XIONG Jian;JIN Longyun;LIU Bingqi;FANG Xinyu(State Grid Jilin Electric Power Co.,Ltd.,Changchun 130021,China;Beijing QU Creative Technology Co.,Ltd.,Beijing 100025,China)
出处
《电子设计工程》
2024年第21期65-69,共5页
Electronic Design Engineering
基金
国家电网公司科技项目(52230021001800)。
关键词
决策树
数据挖掘
煤耗
随机森林
灰色关联度
灰狼优化算法
decision tree
data mining
coal consumption
random forest
grey correlation degree
grey wolf optimization algorithm