基于EEMD与时间序列法的短期风电场功率预测

Short-term Power Forecast of Wind Farm Based on EEMD and Time Series

随着风力发电技术的发展,风电已成为最主要的新能源发电方式。但因风的随机性造成的风场输出功率的随机波动,电网将面对备用容量增多、调度难度增大以及风电场弃风等问题。解决上述问题的有效途径之一就是对风电场输出功率进行准确预测。针对风电场功率时间序列的非线性和非平稳性,分别将EMD和EEMD方法与时间序列的方法相结合应用于风电场功率预测中,提出基于EMD-ARMA和EEMD-ARMA的风功率预测方法。采用某风电场的实际功率数据进行分析预测,预测结果验证了所提方法的正确性和有效性。

With a large number of wind power having been connected into the power grid, it is more important to improve the prediction accuracy of wind power output, in order to increase the ability of accepting wind power, es- tabiish a reasonable power generation plan, and ensure a stable operation of power system. Because the wind power time series is nonlinear and non-stationary, EMD and EEMD are combined into the prediction of wind power out- put, based on which, the method of EMD - ARMA and EEMD - ARMA are proposed. Firstly, using EMD and EE- MD to predict the original wind power series, and the data adaptive to a series of relatively stable component. Sec- ondly, according to the variation of each component, ite can establish the model of internal function, then adding the values together and the sum is the final predictive value. At last, this paper give an example analysis with a certain practical power data of wind farm, the prediction result shows that the proposed method is correct an effec- tive.