高比例风电下电力市场短期电价预测

Short-term Electricity Price Forecasting for High Proportion Wind Power Market

为应对近年来国内风电渗透率不断增加、大量清洁能源并入电网给日前电价预测带来的挑战,提高高比例风电接入情况下电力市场短期电价预测精度,将高比例风电情况下的风电出力与负荷数据进行融合得到了一项改进的输入特征变量,代表风电与负荷共同对电价的影响程度。采用最大信息系数法分析各特征变量与电价之间的相关性,并结合长短期记忆神经网络(long-short term memory,LSTM)与注意力机制(Attention)的特点构建了LSTM-Attetion预测模型,然后对不同输入条件下的预测结果进行对比分析,数据结果显示,引入该输入特征变量后模型的预测精度都有明显提升。经过进一步算例实验后表明,所提出的特征变量相比风荷比而言,能够有效提高高比例风电情况下电价预测精度,适用于许多经典算法。

In order to deal with the challenges brought by the increasing penetration of domestic wind power and the integration of a large number of clean energy into the power grid in recent years,and improve the accuracy of short-term electricity price prediction in the power market under the condition of high proportion of wind power access,the wind power output and load data under the condition of high proportion of wind power to obtain an improved input characteristic variable was integrated,which represents the degree of influence about wind power and load on electricity price.The maximum information coefficient method was used to analyze the correlation between each characteristic variable and the electricity price,and the LSTM-Attetion prediction model was constructed by combining the characteristics of the long-short term memory(LSTM)and the attention mechanism(Attention),then,the prediction results under different input conditions show that the prediction accuracy of the model has been significantly improved after introducing the input characteristic variable.After further experiments,comparing with the wind load ratio,the accuracy of electricity price prediction under the high proportion of wind power can be effectively improved with the characteristic variables proposed,and are suitable for many classical algorithms.