知识和数据驱动的污水处理反硝化脱氮过程协同优化控制  

Knowledge-data-driven Cooperative Optimal Control for Wastewater Treatment Denitrification Process

作　　者：韩红桂 王玉爽 刘峥 孙浩源[1,2,3] 乔俊飞 HAN Hong-Gui;WANG Yu-Shuang;LIU Zheng;SUN Hao-Yuan;QIAO Jun-Fei(Faculty of Information Technology,Beijing University of Technology,Beijing 100124;Beijing Key Laboratory of Computational Intelligence and Intelligent System,Beijing 100124;Engineering Research Center of Digital Community,Ministry of Education,Beijing 100124)

机构地区：[1]北京工业大学信息学部,北京100124 [2]计算智能与智能系统北京市重点实验室,北京100124 [3]数字社区教育部工程研究中心,北京100124

出　　处：《自动化学报》2024年第6期1221-1233,共13页Acta Automatica Sinica

基　　金：国家自然科学基金(62125301,62021003,62103010,62303024);国家重点研发计划(2022YFB3305800-5);中国博士后科学基金(2022M720319);北京市自然科学基金(KZ202110005009);青年北京学者项目(037);北京市博士后工作经费资助项目(2023-zz-91)资助。

摘　　要：为有效提升城市污水处理过程的脱氮效果,提出一种知识和数据驱动的反硝化脱氮过程协同优化控制(Knowledge-data-driven cooperative optimal control,KDDCOC).所提方法主要有以下两个方面:首先,建立一种基于自适应知识核函数的协同优化控制目标模型,动态描述出水水质(Effluent quality,EQ)以及泵送能耗(Pumping energy consumption,PE)、关键变量的协同关系;其次,提出一种知识引导的协同优化算法(Knowledge guide-based cooperative optimization algorithm,KGCO),快速准确求解硝态氮(Nitrate nitrogen,SNO)优化设定值,提高KDDCOC的响应速度.KDDCOC利用比例−积分−微分(Proportional-integral-derivative,PID)控制器对硝态氮优化设定值进行跟踪,将提出的KDDCOC应用于城市污水处理过程基准仿真模型1号(Benchmark simulation model No.1,BSM1),实验结果表明,该方法能够提高出水水质,降低运行能耗,有效改善脱氮效果.In order to effectively improve the performance of wastewater treatment denitrification process,a knowledge-data-driven cooperative optimal control(KDDCOC)is proposed.The main work of this paper includes the following two points:First,a cooperative optimal control objective model,based on adaptive knowledge kernel function,is designed to dynamically describe the cooperative relationship among effluent quality(EQ),pumping energy consumption(PE),and key variables;Second,a knowledge guide-based cooperative optimization algorithm(KGCO)is proposed to quickly and accurately obtain the optimal set-points of nitrate nitrogen(SNO).Then,the response speed of KDDCOC is improved.A proportional-integral-derivative(PID)controller is used to track the optimal setpoints of nitrate nitrogen.The proposed KDDCOC is applied to the benchmark simulation model No.1(BSM1)of wastewater treatment process.The experimental results indicate that KDDCOC can improve the effluent quality and the efficiency of denitrification,reduce the energy consumption.

关 键 词：污水处理反硝化脱氮过程 知识和数据驱动 协同优化控制 自适应知识核函数 知识引导的协同优化算法 

分 类 号：X703[环境科学与工程—环境工程]