基于水力模型耦合机器学习的城市污水干管风险评估  

Risk Evaluation of Urban Sewage Mains Based on Coupling of Hydraulic Model and Machine Learning

作　　者：周军 柯臻玮 霍广炼 周午阳 朱世泰 王广华 ZHOU Jun;KE Zhenwei;HUO Guanglian;ZHOU Wuyang;ZHU Shitai;WANG Guanghua(Guangzhou Urban Drainage Co.,Ltd.,Guangzhou 510308,China;Guangzhou Municipal Engineering Design&Research Institute Co.,Ltd.,Guangzhou 510060,China)

机构地区：[1]广州市城市排水有限公司,广东广州510308 [2]广州市市政工程设计研究总院有限公司,广东广州510060

出　　处：《净水技术》2024年第11期132-140,共9页Water Purification Technology

基　　金：广东省水利科技创新项目(2020-34);广东省建设科技项目(2023-K4-130852)。

摘　　要：污水干管承担了主要的污水转输任务,失效后对环境影响大,为保障污水系统安全稳定运行,需要开展污水干管风险评估。以广州市中心城区的污水干管为研究对象,基于最新的管网现状结构数据和运行数据,耦合使用机器学习、水力模型等多种方法用于评估,简化用于评价的数据种类,筛选出7类指标作为风险评估指标,找出“高风险”污水干管,采用增设连通管等方法降低风险。结果表明:机器学习模型“非高风险”管段预测精度为96.84%,“高风险”管段预测精度为85.95%,可认为模型精度满足使用要求;与改造前的管段进行对比,有88.92%的“高风险”管段转变为“非高风险”管段,保障了污水干管的安全运行。The sewage mains undertake the main task of sewage diversion,which has a great impact on the environment after the failure.In order to ensure the safe and stable operation of sewage pipe network,it is necessary to carry out the risk assessment of sewage main pipe.The main sewage pipe in the central urban area of Guangzhou was taken as the research object,various methods such as machine learning and hydraulic model were coupled.The latest structural data and operation data of the pipe network were utilized,and the types of data used for assessment were simplified.Seven types of indices were selected as risk assessment indices,and"high risk"main sewage pipes were identified,and methods such as adding connecting pipes were adopted to low the risks.The results showed that the prediction accuracy of the machine learning model was 96.84%for the"non-high risk"section and 85.95%for the"high risk"section,which could be considered to meet the requirements of use.Compared with the pipe before the connection,88.92%of the"high risk"pipe section was transformed into a"non-high risk"pipe section,so as to ensure the safe operation of the sewage main pipe.

关 键 词：污水干管 风险评估 机器学习 水力模型 运行安全 

分 类 号：TU992[建筑科学—市政工程]