月径流时间序列预测模型研究与应用OA
Research and Application of Monthly Runoff Time Series Prediction Model
为提高月径流时间序列预测精度,提出季节性趋势分解(STL)-小波包变换(WPT)-海马优化(SHO)算法/白鹭优化算法(ESOA)/冠状病毒优化算法(COVIDOA)-极限学习机(ELM)月径流时间序列预测模型,通过云南省龙潭站月径流时序数据对模型进行验证.利用 STL 将原始月径流序列分解为趋势项、季节项和残差项,利用 WPT 将残差项分解为低频、高频 2 个部分,共得到 4 个子序列分量;利用 SHO、ESOA、COVIDOA寻优 ELM 超参数;建立 STL-WPT-SHO/ESOA/COVIDOA-ELM 模型对各分量逐一预测,将各分量的预测值线性集合即为最终预测结果,构建STL-WPT-SHO-B 等21种模型作对比分析.
To improve the prediction accuracy of monthly runoff time series,the STL-WPT-SHO/ESOA/COVIDOA-ELM monthly runoff time series prediction model is proposed,and the model is verified by the monthly runoff time series data of Longtan station in Yunnan Province.Firstly,STL is used to decompose the original monthly runoff series into trend term,seasonal term and residual term,and then WPT is used to decompose the residual term into low-frequency and high-frequency parts,a total of four subsequence components are obtained;Secondly,the super parameters of ELM were optimized by SHO,ESOA and COVEDOA;Finally,the STL-WPT-SHO/ESOA/COVIDOA-ELM model is established to predict each component one by one,and the linear set of the predicted values of each component is the final prediction result.At the same time,21 models such as stl-wpt-sho-b are constructed for comparative analysis.
董欣林;崔东文
文山州清水河水利枢纽工程管理局,云南 文山 663000云南省文山州水务局,云南 文山 663000
建筑与水利
月径流预测模型对比分析验证
monthly runoff forecastmodelcomparative analysisverification
《云南水力发电》 2026 (9)
10-14,5
国家高分辨率对地观测系统重大科技专项(89-Y50-G31-9001-22)
评论