支持向量机在污水处理厂出水COD预测中的应用  

作　　者：李健 刘坚 于广平 郭清达[2] 袁沐坤[2] 

机构地区：[1]广州工业智能研究院,广东广州511458 [2]中国科学院沈阳自动化研究所,辽宁沈阳110169

出　　处：《环境保护与循环经济》2023年第8期18-21,26,共5页environmental protection and circular economy

基　　金：国家自然科学基金重大项目“城市污水处理过程智能优化运行基础理论及关键技术”(61890935);河北省重点研发计划项目“基于焦化水处理数字化的智能管控关键技术”(21315101D);广东省重点领域研发计划项目“复杂制造环境下的人机物协同控制与决策理论方法”(2020B0101050001);广州市“岭南英杰工程”后备人才项目;羊城创新创业领军人才支持计划资助项目(2019006)。

摘　　要：随着城市化的进行和经济的快速发展,城市居民用水量呈逐年递增趋势。城市污水处理过程中由于水力停留时间过长等,难以实时掌握污水水质情况。化学需氧量(COD)是衡量污水水质的重要指标之一,通过有效监测出水COD的波动,可实现污水处理过程在线管控。通过收集某城镇污水处理厂进水、出水历史数据(COD、pH、溶解氧、氨氮等),利用支持向量机SVM(Support Vector Machine)、K近邻算法(K-Nearest Neighbor,KNN)等人工智能算法建立数学预测模型,实现了对出水COD指标的有效预测。With the progress of urbanization and rapid economic development,urban residents'water consumption has become an increasing trend year by year,urban sewage treatment process due to the hydraulic retention time is too long and other factors,it is difficult to grasp the sewage quality in real time.Chemical oxygen demand(COD)is one of the important indicators to measure sewage quality,through effective monitoring of the fluctuation of effluent COD,the sewage treatment process can be realized online control.By collecting the historical data of inlet and outlet water(COD,pH,dissolved oxygen,ammonia nitrogen,etc.)of a sewage treatment plant in a town,and using artificial intelligence algorithms such as Support Vector Machine(SVM)and K-Nearest Neighbor(KNN)algorithm to establish mathematical prediction models,the effective prediction of effluent COD indicators is realized.

关 键 词：污水处理 人工智能算法 SVM COD预测 

分 类 号：X703[环境科学与工程—环境工程]