摘要
中长期径流模拟可为水资源合理配置提供科学依据,对流域高质量发展具有重要意义。基于改进的自适应噪声完备集成经验模态分解算法(ICEEMDAN)和奇异谱分析(SSA)的两阶段分解策略,利用鲸鱼算法(WOA)优化的长短期记忆网络模型(LSTM),构建了月径流模拟模型ICEEMDAN-SSA-WOA-LSTM。将其应用于石头河水库月径流模拟,并与单次分解的ICEEMDAN-WOA-LSTM、SSA-WOA-LSTM和未分解的WOA-LSTM模型进行对比分析。结果表明,ICEEMDAN-SSA-WOA-LSTM模型模拟效果最佳,率定期和验证期3项评价指标均优于其他模型,验证期均方根误差为1.278 m^(3)/s、平均绝对误差为0.893 m^(3)/s、纳什效率系数为0.985。两阶段分解策略月径流模拟模型可显著提高月径流模拟精度,可用于全年入库径流模拟。
The medium and long-term runoff simulation can provide a scientific basis for the rational allocation of water resources,which is of great significance to the high-quality development of the basin.Based on the improved complete ensemble empirical mode decomposition with adaptive noise(ICEEMDAN)and singular spectrum analysis(SSA)two-stage decomposition strategy,the monthly runoff simulation model ICEEMDAN-SSA-WOA-LSTM was constructed by using the long short-term memory network model(LSTM)optimized by the whale optimization algorithm(WOA).It was applied to the simulation of monthly runoff in Stone River Reservoir and compared with the single decomposed ICEEMDAN-WOA-LSTM,SSA-WOA-LSTM and the undecomposed WOA-LSTM models.The results show that the ICEEMDAN-SSA-WOA-LSTM model has the best simulation effect,and the three evaluation indexes in the calibration period and validation period are better than other models,with the root mean square error of 1.278 m^(3)/s,the average absolute error of 0.893 m^(3)/s,and the Nash efficiency coefficient of 0.985 in the validation period.The two-stage decomposition strategy model can significantly improve the accuracy of monthly runoff simulation and can be used for year-round incoming runoff simulation.
作者
李鑫
王双银
黄毓林
樊镕鑫
马雪燕
LI Xin;WANG Shuang-yin;HUANG Yu-lin;FAN Rong-xin;MA Xue-yan(College of Water Resources and Architectural Engineering,Northwest A&F University,Yangling 712100,China)
出处
《水电能源科学》
北大核心
2023年第9期6-10,共5页
Water Resources and Power
基金
陕西省水资源与河库调度中心关中地区灌溉水源水质状况调查研究项目(SXRCZB2021-ZC-CS1008)。
关键词
径流模拟
二次分解
鲸鱼优化算法
长短期记忆网络
石头河水库
runoff simulation
secondary decomposition
whale optimization algorithm
long short-term memory neu-ral network
Stone River Reservior
