论文部分内容阅读
针对中长期径流预报在水库中长期运行方案制定及调度决策形成中的作用,基于传统和智能预报方法各自的优势,利用均生函数模型记忆时间序列的内在规律,采用偏最小二乘方法对预报因子进行降维处理,建立了结合均生函数的神经网络预报模型,并利用神经网络模型修正预报结果。实例计算表明,该模型不仅可提取径流序列的特征,且预报精度也较单一的均生函数模型和神经网络模型有所提高。
Based on the advantages of both traditional and intelligent forecasting methods, this paper uses the inherent law of memorial function to model the long-term run-time prediction in reservoir and the decision-making process. Based on partial least squares Factor reduction processing, the establishment of a neural network prediction model combined with a homogeneous function, and the use of neural network model to correct the forecast results. The case study shows that this model can not only extract the features of runoff series, but also improve the prediction accuracy compared with a single mean function model and neural network model.