摘要
为了优化循环流化床锅炉(CFB)的床温系统控制,使用支持向量回归(SVR)人工智能方法进行建模预测;为提高预测结果的精准性,引入互信息法则(MI)及主成分分析法(PCA)作为输入特征选择手段,同时使用粒子群算法(PSO)进行参数寻优。研究结果表明,MI法则进行数据预处理并利用PSO算法优化后的SVR模型能够精确地预测在不同运行工况下的床温变化,且预测误差较小,拥有较强的泛化能力。
In order to optimize the bed temperature control of circulating fluidized boiler(CFB),support vector regression(SVR)artificial intelligence method is used for modeling and prediction.To improve the accuracy of the prediction results,while using mutual information law(MI)and principal component analysis(PCA)as input feature selection methods,particle swarm optimization(PSO)is used for parameter optimization.The research result shows that after data preprocessing by the MI law and optimized by the PSO algorithm,the SVR model can accurately predict the bed temperature changes under different operating conditions,and the prediction error is little with strong generalization ability.
作者
黄纯颖
曾庆敏
陈玲红
吴学成
岑可法
HUANG Chun-ying;ZENG Qing-min;CHEN Ling-hong;WU Xue-cheng;CEN Ke-fa(State Key Laboratory of Clean Energy Utilization,Zhejiang University,Hangzhou 310027,China)
出处
《能源工程》
2022年第3期11-17,共7页
Energy Engineering
关键词
循环流化床
床温
支持向量回归
粒子群优化
特征选择
circulating fluidized bed
bed temperature
support vector regression
particle swarm optimization
feature selection
