基于IVMD-KPCA-LSTM的超短期风电功率预测

Ultra-short-term Wind Power Prediction Based on IVMD-KPCA-LSTM

为提高电力系统的可靠性和稳定性,以应对风电波动性带来的挑战,充分考虑制约风力发电的五种气候因素(风速、风向、气压、温度、湿度),首先,利用IVMD方法对气候因子序列进行分解,获得不同时间尺度下数据信号的变化,降低气候因子序列的非平稳性;其次,利用KPCA提取特征序列的关键影响因素,消除原始序列的相关性和允余性,降低模型输入的维数;最后,利用LSTM网络对多变量特征序列进行动态建模,实现风电功率预测。采用龙源电力集团真实风力发电数据预测数据集2号风机数据进行验证,实验结果表明,基于IVMD-KPCA-LSTM的超短期风电功率预测模型相较于单一的LSTM模型在预测精度上有了显著的提升,RMSE下降了48.6%,MAE下降了34.4%,MAPE下降了81.6%。

In order to improve the reliability and stability of the power system and cope with the challenges brought by the volatility of wind power,the five climatic factors that restrict wind power generation(wind speed,wind direction,pressure,temperature,humidity)were fully considered.First,IVMD was used to decompose the climate factor sequence to obtain the changes in data signals at different time scales and reduce the non-stationarity of the climate factor sequence;secondly,We used KPCA to extract the key influencing factors of the characteristic sequence,eliminated the correlation and redundancy of the original sequence,and reduced the dimension of the model's input;finally,the LSTM network was used to dynamically model the multivariable feature sequence to achieve wind power prediction.The Longyuan Electric Power Group's real wind power generation data prediction data set No.2 wind turbine data was used for verification.The experimental results show that the ultra-short-term wind power prediction model based on IVMD-KPCA-LSTM has significantly improved prediction accuracy compared with a single LSTM model.With the improvement,RMSE dropped by 48.6%,MAE dropped by 34.4%,and MAPE dropped by 81.6%.