摘要
建立更准确的NO_(x)生成浓度预测模型对于燃煤机组减少NO_(x)排放,降低脱硝成本具有重大意义。搭建NO_(x)生成模型基于机组相关变量,同时依赖模型结构设计,设计模型结构的参数称为超参数。进行合理的数据处理与超参数设定,能够有效提升NO_(x)预测模型精度与泛化性。该文提出一种基于树状结构Parzen估计器优化长短期记忆(tree-structure parzen estimator optimized long short-term memory neural network,TPE-LSTM)神经网络的NO_(x)生成浓度预测模型。基于某330 MW燃煤机组的历史运行数据,获取NO_(x)生成相关变量参数,将模型结构参数与NO_(x)相关变量参数的时间序列窗口长度以及主成分数量相互耦合,组成一类新的超参数;通过优化改进后的超参数取值,构建基于长短期记忆(long short-term memory,LSTM)神经网络的NO_(x)生成浓度预测模型;将所提出的超参数优化后的NO_(x)预测模型与基于未优化的LSTM模型、采用粒子群优化的LSTM(particle swarm optimization optimized LSTM,PSO-LSTM)模型对比,预测结果表明,TPE-LSTM预测模型具有较好的模型精度与泛化能力。
Developing an efficient prediction model for NO_(x) generation is of great significance for reducing NO_(x) emissions and denitrification costs in coal-fired units.The NO_(x) model is built based on related variables and depends on the design of the model structure,with the parameters of the model structure referred to as hyperparameters.Setting these hyperparameters appropriately can substantially enhance the accuracy and generalization capabilities of NO_(x) prediction models.This paper presents a NO_(x) generation prediction model based on tree-structure parzen estimator optimized long short-term memory neural network(TPE-LSTM).Using historical operational data from a 330 MW coal-fired unit,model structural parameters are combined with time series data window length and the number of principal components of NO_(x) generation related variables,thus creating a new type of hyperparameter.The improved hyperparameters are then optimized to construct NO_(x) generation prediction model based on long short-term memory(LSTM)neural networks.Comparing the proposed hyperparameter optimized NO_(x) prediction model with the unoptimized LSTM model and the typical optimization algorithm particle swarm optimization optimized LSTM(PSO-LSTM)model,the prediction results reveal that the TPE-LSTM prediction model demonstrates superior accuracy and generalization abilities.
作者
陈东升
梁中荣
郑国
何荣强
屈可扬
甘云华
CHEN Dongsheng;LIANG Zhongrong;ZHENG Guo;HE Rongqiang;QU Keyang;GAN Yunhua(School of Electric Power Engineering,South China University of Technology,Guangzhou 510640,Guangdong Province,China;Zhanjiang Electric Power Co.,Ltd.,Zhanjiang 524099,Guangdong Province,China)
出处
《中国电机工程学报》
北大核心
2025年第7期2710-2718,I0022,共10页
PROCEEDINGS OF THE CHINESE SOCIETY FOR ELECTRICAL ENGINEERING
基金
国家自然科学基金项目(52376108)
广东省省级科技计划项目(2022A0505050004)。
