Author(s): Wei Cui; Zhinan Ding; Xiangpeng Mu; Yuling Lei; Hui Liu And Wenxue Chen
Linked Author(s): Wei Cui
Keywords: South-to-North Water Diversion Project Control gate BP neural network One-dimensional unsteady gradually varied flow Water level rise prediction
Abstract: When accidents occur in open-channel water conveyance projects, it is often necessary to quickly close the control gates. Accurately estimating the upstream water level rise is crucial for developing a reasonable gate closure plan. Focusing on a specific section of the Middle Route of the South-to-North Water Diversion Project as the study object, a water level estimation model based on the BP neural network is constructed. Due to the lack of rapid gate closure sample data in the project, a one-dimensional unsteady gradually varied flow simulation model is developed. After calibration with measured data, the model is used for training the neural network. Through simulation analysis, the main factors influencing the upstream water level rise were identified, namely the gate closure amplitude, gate closure duration, and operational water level, which serve as the input variables for the neural network model. Based on tests from 25 cases, the BP neural network model’s maximum positive deviation in water level estimation is 0.034 m, and the maximum negative deviation is -0.020 m, with an accuracy sufficient to meet the needs of engineering applications.
Year: 2025