| 引用本文: | 李守涛,王军涛,于 明,等.基于GA-RBF神经网络的位山闸引水能力预测研究[J].灌溉排水学报,2021,(6):100-104. |
| LI Shoutao,WANG Juntao,YU Ming,et al.基于GA-RBF神经网络的位山闸引水能力预测研究[J].灌溉排水学报,2021,(6):100-104. |
|
| 摘要: |
| 【目的】对位山闸改建后的引水能力进行预测。【方法】利用RBF神经网络非线性拟合能力强和遗传算法寻优能力强的优点，建立一种基于遗传算法（GA）优化RBF神经网络隐层各参数的位山闸引水能力预测模型，模型输入变量为闸门开数、闸前、闸后水深和季节因子（受汛期影响，汛期内季节因子为1，非汛期内季节因子为2），输出变量为实测过闸流量，利用现状水情数据组成的样本集对该模型进行训练和检验，检验训练后的模型平均误差为1.64%，证明预测效果较好。【结果】汛期时引水能力能满足设计要求；非汛期时，引水能力随闸后水头降低而增大，考虑闸后输沙渠下挖改造方案，基本也能满足设计要求。【结论】GA-RBF模型在位山闸引水能力预测上适应性强，预测精度高，有一定的推广应用价值。 |
| 关键词: 位山闸；GA-RBF；引水能力；预测 |
| DOI：10.13522/j.cnki.ggps.2020161 |
| 分类号: |
| 基金项目: |
|
| Using GA-RBF Neural Network Model to Calculate the Diversion Capability of the Weishan Sluice |
|
LI Shoutao, WANG Juntao, YU Ming, YAO Jingwei, ZHAO Guoping, FAN Yumiao
|
|
1. Yellow River Institute of Hydraulic Research, Zhengzhou 450045, China; 2. Hohai University College of Water Conservancy and Hydropower Engineering, Nanjing 210098, China; 3. Weishan Irrigation District Management Office of Liaocheng City, Liaocheng 252053, China; 4. Yellow River Affairs Bureau of Liaocheng City, Liaocheng 252000, China
|
| Abstract: |
| 【Background and Objective】Sluice is an important component in hydraulic projects to control water flow, and its performance and diverting capacity are affected by its operating parameters and ambient environment. Understanding how its diversion capacity varies with these factors is important but not trivial. The objective of this paper is to present a new model to estimate the change in diversion capacity of sluice in response to its operating parameters and environmental factors. 【Method】We took the Weishan sluice in Shandon province as an example. The change in its flowing capacity with factors including the number of its openings, water depth at the back and front of the sluice was analyzed using the radial basis function neural network. Optimization of the network, including the number of its hidden layers, was solved using the genetic algorithm. The model was trained first using measured data, and it was then used to evaluate the performance of the sluice under different operating and environmental conditions.【Result】The trained model was accurate, and compared with the measured data, its average error was 1.64%. Results calculated from the model showed that in flooding seasons, the diversion capacity of the sluice meets the design requirement, while during dry seasons its diversion capacity increased with the reduction in the water depth at the back of the sluice. Considering the dredge to be carried out in the downstream sediment conveyance, the calculated flowing capacity meets the design requirements.【Conclusion】The GA-RBF model was adaptable and accurate for calculating the diversion capacity of the Weishan sluice. |
| Key words: Weishan Sluice; GA-RBF; diversion capacity; prediction |
