文献类型：期刊文献

英文题名：Intelligent control of the electrochemical nitrate removal basing on artificial neural network (ANN)

作者：Meng, Guangyuan[1];Fang, Liqiang[1];Yin, Yao[2];Zhang, Zhijie[1];Li, Tong[3];Chen, Peng[1,6];Liu, Yongdi[1];Zhang, Lehua[1,4,5,6]

机构：[1]East China Univ Sci & Technol, Natl Engn Lab Ind Wastewater Treatment, Shanghai 200237, Peoples R China;[2]Shanghai Municipal Engn Design Inst Grp Co Ltd, Shanghai 200092, Peoples R China;[3]Bozhou Univ, Continuing Educ Ctr, Bozhou 236800, Peoples R China;[4]East China Univ Sci & Technol, State Environm Protect Key Lab Environm Risk Asses, Shanghai 200237, Peoples R China;[5]Shanghai Inst Pollut Control & Ecol Secur, Shanghai 200092, Peoples R China;[6]130 Meilong Rd, Shanghai 200237, Peoples R China

年份：2022

卷号：49

外文期刊名：JOURNAL OF WATER PROCESS ENGINEERING

收录：;WOS:【SCI-EXPANDED(收录号:WOS:000865362700003)】；

基金：This work was supported by the National Key Research and Devel-opment Program of China (Developing Highly Efficient Means of Water Utilization, 2019YFC0408202) , the National Natural Science Foundation of China (NSFC) (21876050) , and Shanghai Rising-Star Program (19QB1405300) .

语种：英文

外文关键词：Electrochemistry; Nitrate removal; Artificial intelligence; Back propagation algorithm; Artificial neural network

摘要：Artificial intelligence technologies were confirmed as a useful tool for wastewater treatment, but its application in the electrochemical nitrate removal had less been reported. In this work, the artificial neural network in machine learning was used to construct the model, which combines the methods of electrochemistry and arti-ficial intelligence to achieve the prediction and intelligent control of nitrate removal. The control system consists of a prediction module with an artificial neural network (ANN) algorithm model and a control module. First, initial nitrate concentration, pH, time and current density were considered as input. An ANN algorithm using 7 hidden layers and a negative feedback regulation mechanism was developed to optimize the model and predict the nitrate removal rate. Results indicates that the proposed prediction model (4-7-1) yields a better coefficient of determination and lower root mean square error. The optimal set-points of the current density in the elec-trochemical process can be obtained according to the water quality change and qualified effluent quality using the ANN model. Also, the proposed intelligent control strategy can eliminate the influence of water quality change on nitrate removal and reduce energy consumption by 15.0 % compared to the post strategy. This work demonstrated the potential of artificial intelligence in the electrochemical process of nitrate removal.