引用本文: | 崔明建,孙元章,柯德平,王树鹏.基于原子稀疏分解理论的短期风电功率滑动预测[J].电力自动化设备,2014,34(1): |
| CUI Mingjian,SUN Yuanzhang,KE Deping,WANG Shupeng.Short-term wind power forecasting based on atomic sparse decomposition theory[J].Electric Power Automation Equipment,2014,34(1): |
|
摘要: |
采用一种具有很强的非平稳信号跟踪、预测能力的原子稀疏分解(ASD)法,作为人工神经网络(ANN)的前置分解方法,将风电功率序列分解为原子分量和残差分量,对原子分量进行自预测,残差分量进行ANN预测,再通过追加最新的风电功率实时数据来更新ASD的结果,进而滑动预测下一个时刻的风电功率。以实际风电场数据进行验证,结果证明了该模型可以有效地处理风电功率非平稳性,产生更为稀疏的分解效果,显著地降低了绝对平均误差、均方根误差计算值的统计区间。 |
关键词: 风电 预测 原子稀疏分解 人工神经网络 模型 |
DOI: |
分类号: |
基金项目:国家重点基础研究发展计划(973计划)资助项目(2012-CB215101) |
|
Short-term wind power forecasting based on atomic sparse decomposition theory |
CUI Mingjian1, SUN Yuanzhang1, KE Deping1, WANG Shupeng2
|
1.School of Electrical Engineering,Wuhan University,Wuhan 430072,China;2.School of Mathematics and Physics,China University of Geosciences,Wuhan 430074,China
|
Abstract: |
ASD(Atomic Sparse Decomposition),which has excellent ability to track and forecast unstable signal,is applied as the pre-decomposition of ANN(Artificial Neural Network) to decompose the wind power series into atomic component and residual component. The former is self-forecasted while the latter is forecasted by ANN. The latest real-time data of wind power are added to update the result of ASD for forecasting the wind power of next instant. The model is verified by the practical data of a wind farm,which shows that,the instability of wind power is effectively dealt with to produce more sparse decomposition effect,significantly reducing the statistical intervals of absolute mean error and root mean square error. |
Key words: wind power forecasting atomic sparse decomposition ANN models |