基于机器学习方法的葛洲坝水电站下游水位预测研究  被引量:2

Research on downstream water level prediction of Gezhouba Hydropower Station based on machine learning methods

在线阅读下载全文

作  者:刘晓阳 郭乐[1,2] LIU Xiaoyang;Guo Le(Three Gorges Cascade Dispatch&Communication Center,Yichang 443002,China;Hubei Key Laboratory of Intelligent Yangtze and Hydroelectric Science,Yichang 443002,China)

机构地区:[1]三峡水利枢纽梯级调度通信中心,湖北宜昌443002 [2]智慧长江与水电科学湖北省重点实验室,湖北宜昌443002

出  处:《水利水电快报》2022年第10期19-22,36,共5页Express Water Resources & Hydropower Information

摘  要:为准确预测葛洲坝水电站下游水位,以人工神经网络、随机森林和支持向量机3种机器学习方法为基础,以2018~2020年葛洲坝水电站历史数据为样本集,建立葛洲坝水电站下游水位单点预测模型,并利用k折交叉验证对3种方法的预测精度进行评价。结果表明:3种方法在测试集上均具有良好表现;其中支持向量机预测模型表现最优,均方误差MSE平均值为0.0071,决定系数R^(2)平均值为0.98,预测精度可满足生产需求。研究成果对葛洲坝水电站后期调度方案编制具有指导意义。In order to accurately predict the downstream water level of Gezhouba Hydropower Station.We constructed downstream water level prediction model of Gezhouba hydropower station based on historical data of the power station in the years of 2018 ~2020 by three machine learning methods of artificial neural network, random forest and support vector machine.Then we evaluated the prediction accuracy of the three models by K-fold cross validation.The results showed that three models had good performance on the test set.Among them, the support vector machine prediction model performs best, with an average MSE of 0.0071 and an average R^(2) of 0.98,and the prediction accuracy could meet the production demand, which had great guiding significance for the preparation of the future dispatching plan of Gezhouba Hydropower Station.

关 键 词:水位预测 人工神经网络 随机森林 支持向量机 葛洲坝水电站 

分 类 号:TV747[水利工程—水利水电工程]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象