考虑短期负荷影响的DeepESN电力市场实时电价预测研究

精准的电价预测有助于宏观调控的实施。但能源结构转型导致大规模可再生能源并网，因此会导致电价降低和产生波动，降低时序预测序列的相关性，加大实时电价的预测难度。针对这一问题，采用自相关函数和最大信息数计算电价自身和电价与电量关联性，为模型输入提供依据，并在此基础上应用具有深度储备池特性的深度回响网络进行实时电价预测。研究结果表明:电价与电量、电价自身具有较强相关性，应考虑自身与电量因素；深度回响网络能够显著提升预测模型的精度，并且具有较强的鲁棒性。

Real-time Electricity Price Forecasting of Electricity Market Using DeepESN Considering Short-term Load Impact

Accurate price forecasting is helpful to the implementation of macro-control. However， the transformation of energy structure leads to large-scale renewable energy integration， causing the reduction and fluctuation of electricity price， reducing the sequence correlation of time series forecasting， and increasing the difficulty of real-time electricity price forecasting. To solve this problem， this paper uses autocorrelation function and the maximum number of information to calculate the electricity price itself and the correlation between the electricity price and electricity quantity， and provides basis for model input. On this basis， deep echo state network with deep reserve pool characteristics is used for the real-time electricity price forecasting. The results show that: there is a strong correlation between the electricity price and the electricity quantity and electricity price itself， which should be taken into account； The deep echo state network can significantly improve the accuracy of forecasting model and has strong robustness.