基于多边际条件下的水位流量预测模型在防洪和水资源调配中的应用  

Application of stage discharge prediction model based on multi marginal conditions in flood control and water resources dispatching

作　　者：钟惠钰 陈棨尧 ZHONG Huiyu;CHEN Qiyao(Taihu Basin Authority of Ministry of Water Resouces Suzhou Administration,Suzhou 215000,China)

机构地区：[1]太湖流域管理局苏州管理局,江苏苏州215000

出　　处：《江苏水利》2025年第3期7-11,共5页Jiangsu Water Resources

摘　　要：为支撑数字孪生太浦闸先行先试建设任务,构建数字孪生水利工程“四预”闭环链条,研究上线了基于机器学习的数字孪生太浦闸多维动态水位流量预测模型,解决了工程调度运行与安全运行的相关业务难题。该模型基于对原有BP神经网络预测模型优化改进,耦合太湖流域水质水量模型,对上下游水位预报成果进行反演推算,实现预测工程未来24h泄流能力,为流域防洪和水资源调度提供决策参考。通过优化影响因子权重,调整训练样本,闸门精准控制率提高10%以上,调度执行精准率达到100%,在倒流“四预”上取得突破,极大地保障了工程安全运行。在2022年防御台风和向下游地区压咸保供工作中应用成效显著。In order to support the pilot construction task of digital twin Taipu gate and to build the"four forecast"closed-loop chain of digital twin water conservancy project,the multi-dimensional dynamic stage discharge prediction model of digital twin Taipu gate based on machine learning has been developed and launched,which has resolved relevant operational challenges in project scheduling and safe operation.The model is based on the optimization and improvement of the original BP neural network prediction model,coupling the water quality and quantity model of Taihu Basin,and reversing the upstream and downstream water level forecast results,so as to predict the discharge capacity of the project in the next 24 hours,and provide reference for decision-making for flood control and water resources dispatching in the basin.By optimizing the weight of the influence factor and adjusting the training sample,the precision control rate of the gate increased by more than 10%,and the precision rate of scheduling execution reached 100%,a breakthrough was made in the“four forecasts”of reverse flow,which greatly guaranteed the safe operation of the project.In 2022,the application of typhoon defense and the work of pressing salt and supply to downstream areas has achieved remarkable results.

关 键 词：数字孪生 机器学习 水位流量预测 防洪 水资源调度 高质量发展 

分 类 号：TV21[水利工程—水文学及水资源]