工业烟气SCR脱硝过程智能控制模型的研究进展  

Research progress on intelligent control model of industrial gas SCR denitrification process

作　　者：章新宇 任文锦 余正伟 龙红明 陈良军 张旭 ZHANG Xinyu;REN Wenjin;YU Zhengwei;LONG Hongming;CHEN Liangjun;ZHANG Xu(School of Metallurgical Engineering,Anhui University of Technology,Ma’anshan 712046,Anhui,China;Shaanxi Hydrogen Energy Research Institute Co.,Ltd.,Xi’an 712046,Shaanxi,China)

机构地区：[1]安徽工业大学冶金工程学院,安徽马鞍山243011 [2]陕西氢能研究院有限公司,陕西西安712046

出　　处：《烧结球团》2024年第6期53-67,共15页Sintering and Pelletizing

基　　金：国家自然科学基金青年科学基金资助项目(52404333)。

摘　　要：随着工业的快速发展和环保要求的提升,氮氧化物(NO_(x))排放的控制成为大气污染物治理的重要议题。传统的选择性催化还原(SCR)等脱硝工艺虽然已经较为成熟,但大多仍采用粗犷的人工控制或精细化程度较低的自动、半自动化控制方式,存在排放波动大、资源消耗大、劳动强度大等诸多缺陷。近年来,随着人工智能、机器学习及大数据分析等技术的飞速发展,智能化技术逐步被引入到工业烟气SCR脱硝过程控制中用以克服传统控制方式的不足,取得了一定的成效。本文综述了近年来烟气脱硝过程智能控制模型的研究进展,重点介绍了基于大数据、人工智能、机器学习等先进技术的烟气脱硝优化控制方法;对采用BP神经网络、深度学习(如LSTM、CNN)等方法进行NO_(x)浓度预测和脱硝控制优化以及BP神经网络的优化算法(如结合灰狼算法、粒子群算法等)的研究进行综述,并对比基于PID的模糊控制与二自由度策略等自适应控制模型的优势与控制效果,探讨脱硝过程智能控制技术在工业中的应用现状及面临的挑战;最后,展望了脱硝过程智能控制技术的发展趋势与研究方向,以期为工业烟气污染物治理提供参考。With the rapid development of industry and improvement of environmental protection requirements,the control of nitrogen oxide(NO_(x))emissions has become an important issue in air pollution control.Although the traditional denitrification processes such as SCR have been relatively mature,most of them still use rough manual control or automatic and semi-automatic control methods with a low refinement,which have many defects including large emission fluctuations,large resource consumption and high labor intensity.In recent years,with the rapid development of artificial intelligence,machine learning and big data analysis technologies,intelligent technologies have been gradually introduced into the control of industrial gas SCR denitrification process to overcome the shortcomings of traditional control methods,and have achieved certain results.The research progress of intelligent control models for gas denitrification process in recent years is reviewed,the optimal control methods for gas denitrification based on such advanced technologies as big data,artificial intelligence,and machine learning are introduced;NO_(x) concentration prediction and denitrification control optimization are carried out through BP neural network,deep learning(such as LSTM,CNN)and other methods;the research on optimization method of BP neural networks(such as combining gray wolf algorithm,particle swarm algorithm,etc.)is reviewed;the advantages and control effects of PID-based fuzzy control and adaptive control models such as two-degree-of-freedom strategy is compared;the application status and challenges of intelligent control technology for denitrification process in industry are discussed;finally,the future development trend and research direction of intelligent control technology for denitrification process are prospected,in order to provide reference for industrial gas pollution control.

关 键 词：工业烟气 脱硝技术 人工智能 大数据 优化控制 

分 类 号：X701.7[环境科学与工程—环境工程]