基于GRU-XGBoost的风电场功率短期预测

随着风电技术的不断发展,更多的风电机组并入电网运行。考虑到电网的安全性与稳定性,精确的风电场发电短期预测技术越发重要。在利用自适应噪声的完备经验模态分解(CEEMDAN)风电原始序列信号的基础上,采用GRU-XGBoost模型对非线性、非平稳的功率序列进行建模和预测,以提高模型的预测能力和泛化能力。首先利用CEEMDAN将风电功率原始序列分解为一系列不同时间尺度的分量,将分解后的信号输入GRU神经网络输出预测信号,再输入XGBoost进行校正。通过与多种预测模型进行比较证明此模型拥有更好的预测精度。

Short-term Prediction of Wind Farm Power Based on GRU-XGBoost

With the continuous development of wind power technology, more wind turbines are integrated into the grid. Considering the safety and stability of the power grid, accurate short-term prediction technology for the wind farm power generation is more and more important. Based on adopting the complete empirical mode decomposition(CEEMDAN) of adaptive noise to decompose wind power original sequence signals, the GRU-XGBoost model is used to model and predict nonlinear and non-stationary power sequences to improve the prediction ability of the model and generalization ability. The method CEEMDAN decomposes the original wind power sequence into a series of components of different time scales, inputs the decomposed signal into the GRU neural network to output the prediction signal, and then inputs it into XGBoost for correction. Compared with a variety of predictive models, it proves that this model has better prediction accuracy.