基于MI-PCA和ML-AE-ELM的脱硝系统入口NOx质量浓度预测  

Prediction of NOx Mass Concentration at Inlet of Denitration System Based on MI-PCA and ML-AE-ELM

作　　者：靳果[1,2] 屈保中[1,2] 朱清智[1,2] Jin Guo;Qu Baozhong;Zhu Qingzhi(School of Automation Engineering,He’nan Polytechnic Institute,Nanyang 473000,China;He’nan Engineering Research Center of Industrial Robot Application,He’nan Polytechnic Institute,Nanyang 473000,China)

机构地区：[1]河南工业职业技术学院自动化工程学院,河南南阳473000 [2]河南工业职业技术学院河南省工业机器人应用工程技术研究中心,河南南阳473000

出　　处：《兵工自动化》2023年第12期76-82,共7页Ordnance Industry Automation

基　　金：河南省2021年科技发展计划(212102210527);2022年度河南省高等学校重点科研项目(22A120004);2020年南阳市科技计划项目(KJGG207)。

摘　　要：为提高脱硝系统入口氮氧化物(nitrogen oxide,NOx)质量浓度的预测精度,提出一种基于互信息的主成分分析(principal component analysis,PCA)和多层自编码结构的极限学习机(extreme learning machine,ELM)组合算法的预测模型。对输入变量的选取方式进行完善,并加入历史NOx质量浓度,对预测算法的网络结构进行优化。实验结果表明:与其他预测算法模型相比,该模型具有较高的预测效率,在不同工况下均具有较高的预测精度,表现出良好的抗噪能力和泛化能力。In order to improve the prediction accuracy of nitrogen oxides(NOx)mass concentration at the inlet of denitration system,a combination algorithm prediction model of principal component analysis(PCA)and an extreme learning machine(ELM)with multi-layer self-coding structure based on mutual information is proposed.The selection of input variables is improved,and the network structure of the prediction algorithm is optimized by adding the historical NOx mass concentration.The experimental results show that compared with other prediction algorithm models,the proposed model has higher prediction efficiency and higher prediction accuracy under different working conditions,and shows good anti-noise ability and generalization ability.

关 键 词：脱硝系统 NOx质量浓度 互信息 主成分分析 极限学习机 预测模型 

分 类 号：TP272[自动化与计算机技术—检测技术与自动化装置]