基于BP神经网络的供水管网分时段宏观模型研究

准确快速地模拟供水管网的运行状态是城市供水系统优化调度的前提和关键。针对城市供水管网建模难的状况，结合南京市供水管网的特点，利用BP神经网络建立了城市供水管网分时段宏观等效模型。经实际运行数据验证，该模型计算得到的压力预测值与实际值非常接近，预测误差能满足优化调度的工程精度要求。可见该模型基本上能反映南京市供水管网的供水量与水压的关系，可用以模拟城市供水管网的运行状态，为管网优化调度提供了有力支持。

Study on Time-division Macroscopic Mode of Water Distribution System Based on BP Neural Network

Exact and rapid simulation of operation status of water distribution system is prerequisite and key for optimal management of urban water distribution system. In order to solve the difficulty in creating a practical model of urban water distribution system, a time-division macroscopic equivalent mode of urban water distribution system was constructed on the basis of the water distribution system of Nanjing City and the BP neural network. The forecasting value of hydraulic pressure in the model is relatively close to measured value. Practical requirements for the optimal management are reached. The model can basically reflect the relationship between the water supply capacity and the hydraulic pressure in Nanjing City. It can be used to simulate operation status of urban water distribution system and provide an effective support for optimal management of water distribution system.