摘要
文中提出了一种基于预测的城市供水管网运行评估方法。基于精细化管理需求,为提高城市供水系统效率,响应节能供水号召,在此背景下,使用机器学习方法以及集成学习思想,首先利用了融合多纠偏机制的混合需水量预测模型对供水区域未来1 d的居民总需水量进行预测,然后提出了基于序列到序列的编解码网络结构对管网各个节点的水压进行分钟级预测。最后基于预测信息、节假日信息等提出了管网测压点的压力动态阈值的优化方法,实现了对城市供水管网整体运行状态的准确评估,为城市供水系统的水量调配提供调度指导,通过压力波动数据可以看出,研究所提供的阈值计算方法提高了调度决策水平。与此同时,文中以上海某区域为背景,通过需水量预测以及管道压力预测的方法,验证了所提出的模型在实际工作中,相较于传统模型具有更高的预测精度。
This paper proposed a prediction-based method for evaluating the operation status of urban water supply networks.In response to the call for energy-saving water supply and in order to improve the efficiency of the urban water supply system,machine learning methods and integrated learning ideas were used in the context of fine-tuned management needs.Firstly,a mixed water demand forecasting model that integrates multiple correction mechanisms was used to predict the total water demand of residents in the water supply area in the next day,and then a sequence-to-sequence codec network structure was proposed to predict the water pressure of each node of the pipeline network in minutes.Finally,based on the forecast information and holiday information,the pressure dynamic threshold of the pressure measuring point of the pipe network was proposed,which realized the accurate assessment of the overall operation status of the urban water supply pipe network,and provided scheduling guidance for the water distribution of the urban water supply system.It can be seen from the pressure fluctuation data that the threshold calculation method provided by the research has improved the level of scheduling decision-making.At the same time,taking a certain area in Shanghai as the background,through the methods of water demand prediction and pipeline pressure prediction,it is verified that the proposed model has higher prediction accuracy than the traditional model in actual work.
作者
赵平伟
蓝杨
冯偲慜
王景成
ZHAO Pingwei;LAN Yang;FENG Simin;WANG Jingcheng(Shanghai Chengtou Water<Group>Co.,Ltd.,Shanghai 200002,China;School of Electronic Information and Electrical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China)
出处
《净水技术》
CAS
2023年第2期147-155,共9页
Water Purification Technology
关键词
需水量预测
水压预测
压力动态阈值
供水管网状态评估
智能调度
water demand forecast
water pressure prediction
dynamic pressure threshold
water supply network status assessment
smart controling
