摘要
针对传统指数平滑法在风速预测中的不足,搭建基于改进指数平滑法和马尔科夫修正模型的风速组合预测模型。通过梯度优化法快速追踪最优平滑系数α,加快计算效率和提高预测精度;利用马尔科夫模型修正残差,进一步增加预测精度。实验结果表明,该模型在计算效率上,比动态遍历指数平滑法提高近80%;而在预测精度上,比传统指数平滑法和灰色预测法分别提高了27%和32%。该模型对于风速的预测是准确、有效的,具有一定的实用价值。
For the deficiency of traditional exponential smoothing method in forecasting wind speed,this paper proposed an wind speed forecasting method based on the improved exponential smoothing method and Makov correction model.Firstly,optimal smoothing coefficientα was fastly traced by gradient optimization method to accelerate computational efficiency and improve forecasting accuracy.Then the Markov model was applied to modify residuals and further improve the forecasting accuracy.Experimental results show that the proposed model increases the computational efficiency by 80% compared with the dynamic traversal exponential smoothing method,and improves the prediction accuracy respectively by 27% and 32% compared with the traditional exponential smoothing algorithm and the grey prediction method.The proposed method for wind speed prediction is accurate,and effective,which has great practical value.
出处
《电力科学与技术学报》
CAS
北大核心
2016年第1期85-89,共5页
Journal of Electric Power Science And Technology
关键词
风速预测
传统指数平滑法
平滑系数
梯度优化
马尔科夫模型
wind speed forecasting
traditional exponential smoothing method
smoothing coeffi cient
gradient optimization
Markov model
