| 基于核极端学习机的短期风电功率预测研究 |
| |
| 引用本文： | 华北电力大学 新能源电力系统国家重点实验室,北京,华北电力大学 控制与计算机工程学院,河北 保定.基于核极端学习机的短期风电功率预测研究[J].热能动力工程,2017,32(1):95-100. |
| |
| 作者姓名： | 华北电力大学 新能源电力系统国家重点实验室 北京 华北电力大学 控制与计算机工程学院 河北 保定 |
| |
| 作者单位： | 随着风力发电的大规模发展和并网运行，风电场输出功率的精确预测对电力系统的运行具有重大意义。针对风力发电功率具有非线性和非平稳的特性，利用经验模态分解和核极端学习机结合的方法对短期风力发电功率预测进行研究。通过经验模态分解把风电功率时间序列分解成为一系列相对平稳的子数据序列，对每个子数据序列采用核极端学习机算法分别进行模型建立与预测，把每个预测模型得到的子数据序列预测值相加获得最终的风电功率预测值。基于此方法的某风电场输出功率实例数据预测仿真结果表明，该方法的预测模型能更好地跟踪风电功率的变化，预测误差比单独KELM方法减小7.6%，比EMD-SVM方法减小1.7%，能够在一定程度上提高风电功率预测的准确性。 |
| |
| 摘 要： | 随着风力发电的大规模发展和并网运行,风电场输出功率的精确预测对电力系统的运行具有重大意义。针对风力发电功率具有非线性和非平稳的特性,利用经验模态分解和核极端学习机结合的方法对短期风力发电功率预测进行研究。通过经验模态分解把风电功率时间序列分解成为一系列相对平稳的子数据序列,对每个子数据序列采用核极端学习机算法分别进行模型建立与预测,把每个预测模型得到的子数据序列预测值相加获得最终的风电功率预测值。基于此方法的某风电场输出功率实例数据预测仿真结果表明,该方法的预测模型能更好地跟踪风电功率的变化,预测误差比单独KELM方法减小7.6%,比EMD-SVM方法减小1.7%,能够在一定程度上提高风电功率预测的准确性。
|
| 关 键 词： | 风力发电 功率预测 经验模态分解 核极端学习机 |
Short Term Wind Power Prediction based on Extreme Learning Machine with Kernels |
| |
| State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources,North China Electric Power University,Beijing,Chin,Post Code: and School of Control and Computer Engineering,North China Electric Power University,Baoding,Chin,Post Code:.Short Term Wind Power Prediction based on Extreme Learning Machine with Kernels[J].Journal of Engineering for Thermal Energy and Power,2017,32(1):95-100. |
| |
| Authors: | State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources North China Electric Power University Beijing Chin Post Code: and School of Control and Computer Engineering North China Electric Power University Baoding Chin Post Code: |
| |
| Affiliation: | With the large scale development of wind power and grid connection operation, accurate prediction of wind power is important for the power system. A method combined with empiricalmode decomposition and extreme learning machine with kernels was proposed to cope with the nonlinearity and nonstationarity of wind power data for short term wind power prediction. The empirical mode decomposition method was utilized to decompose the signal of wind power into series of stable sequences. The extreme learning machine with kernels was used to model and predict each sequence data. Eventually, the prediction results of each sequence data were added to obtain the final wind power prediction result.The simulation results show that forecasting model with the proposed method in this study can track the change of wind power better,and the prediction error is 7.6% less than the KELM method and 1.7% less than the EMD SVM method,which can effectively improves the prediction accuracy of wind power prediction. |
| |
| Abstract: | LIU Chang liang |
| |
| Keywords: | wind power power prediction empirical mode decomposition extreme learning machine with kernels |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《热能动力工程》浏览原始摘要信息 |
|
点击此处可从《热能动力工程》下载全文 |
|
