摘要
地下水水位埋深是影响河套灌区生态环境的主要因素,开展地下水位埋深预测研究对灌区远景发展规划与用水管理具有现实指导意义。本文采用基于粒子群算法的BP神经网络模型(PSO-BP),对河套灌区永济灌区地下水位埋深进行了预测模拟,相对于传统BP模型纳什效率系数NSE (0.791), PSO-BP模型NSE(0.887)提高了12%。表明,BP神经网络可以有效处理地下水位埋深与其影响因素之间的复杂非线性问题,同时粒子群算法可以提高模型的预测精度。
The groundwater level depth is the main factor affecting the ecological environment in Hetao irrigation district.Carrying out the study on prediction of the groundwater level depth has practical guiding significance in the long-term development planning and water use management for the irrigation area.In this paper,the authors,based on PSO-BP in particle swarm optimization algorithm,carried out predicting and simulating groundwater level depth in irrigation area in Hetao irrigation district,comparing PSO-BP model with traditional BP model,the NSE is increased from 0.791 to 0.887,the extent of improvement is about 12%.The results show that by means of BP neural network,the complicated and nonlinear problems between the groundwater level depth and the factors affecting on it can be effectively solved,also using the particle swarm optimization algorithm can improve the model prediction accuracy.
作者
李皓璇
仲委
王宁
侯效锋
Li Haoxuan;Zhong Wei;Wang Ning;Hou Xiaofeng
出处
《吉林水利》
2020年第8期12-14,27,共4页
Jilin Water Resources
基金
2019年扬州大学大学生科创基金项目“永济灌域地下水位埋深预测研究”(X20190486)。
关键词
地下水位埋深预测
BP神经网络
粒子群算法
永济灌域
prediction for groundwater level depth
BP neural network
particle swarm optimization algorithm
Yongji irrigation area
