| 引用本文: | 程启明,陈路,程尹曼,张强,高杰.基于EEMD和LS-SVM模型的风电功率短期预测方法[J].电力自动化设备,2018,(5): |
| CHENG Qiming,CHEN Lu,CHENG Yinman,ZHANG Qiang,GAO Jie.Short-term wind power forecasting method based on EEMD and LS-SVM model[J].Electric Power Automation Equipment,2018,(5): |
|
| 摘要: |
| 原始风速信号具有的间歇波动性特征给风电场的功率预测带来了挑战，采用集合经验模态分解(EEMD)法将原始风速信号分解为频域稳定的子序列，有效地提高了预测精度，避免了传统经验模态分解(EMD)存在的模态混叠现象。提出一种改进型果蝇优化算法(FOA)，将风速子序列重构参数和最小二乘支持向量机(LS-SVM)参数作为优化目标建立风速预测模型，扩大了参数搜索范围，提高了优化收敛速度；通过风速风功率转化关系可以求得风电场的功率值。实验结果验证了所提方法相比于EMD和LS-SVM预测方法具有更高的预测精度。 |
| 关键词: 微电网 功率预测 风电场 模态分解 支持向量机 相空间重构 果蝇优化算法 |
| DOI：10.16081/j.issn.1006-6047.2018.05.004 |
| 分类号:TM615 |
| 基金项目:国家自然科学基金资助项目(61573239)；上海市重点科技攻关计划项目(14110500700)；上海市电站自动化技术重点实验室项目(13DZ2273800)；上海市自然科学基金资助项目(15ZR1418600) |
|
| Short-term wind power forecasting method based on EEMD and LS-SVM model |
|
CHENG Qiming1, CHEN Lu1, CHENG Yinman2, ZHANG Qiang1, GAO Jie1
|
|
1.Shanghai Key Laboratory Power Station Automation Technology Laboratory, College of Automation Engineering, Shanghai University of Electric Power, Shanghai 200090, China;2.North Power Supply Branch, Shanghai Electric Power Company, Shanghai 200041, China
|
| Abstract: |
| Since the intermittent fluctuation characteristics of original wind speed signal have brought challenges to wind power forecasting, the EEMD(Ensemble Empirical Mode Decomposition) method is adopted to decompose the wind speed signal into stable sub-sequence signals on the frequency domain, which effectively improves the forecasting accuracy and avoids the modal aliasing phenomenon of the traditional EMD(Empirical Mode Decomposition) method. An improved FOA(Fruit fly Optimization Algorithm) is proposed, which takes the parameters of wind speed sub-sequence reconstruction and LS-SVM(Least Squares Support Vector Machine) as the optimal objective to establish the wind speed forecasting model, expanding the search range of the parameters and improving the convergence speed. The wind power can be obtained according to the transformation relationship between wind speed and wind power. Experimental results show that the proposed method has higher forecasting accuracy than the EMD or LS-SVM method. |
| Key words: microgrid power forecasting wind farms mode decomposition support vector machines phase space reconstruction fruit fly optimization algorithm |
