基于多周期GARCH-M模型的短期电价预测

Short-term Electricity Price Forecasting Based on Multi-cycle GARCH-M Model

电价的分布特性是电力市场风险管理和电力金融产品定价的重要依据。建立了一个采用虚拟变量和正弦函数来刻画现货电价序列多周期性特征的GARCH-M模型。该模型易于定阶、待估参数少,可同时处理电价序列的趋势变化、多周期、异方差及其与负荷之间的非线性相关性,具有一定的实用价值。对PJM电力市场历史数据的分析表明,电价分布的异方差和负荷的平方对电价均值具有显著的影响,电价序列具有周、半月、月、季、半年等多重周期和明显的波动集聚性。

The distribution properties of power prices are the important information for the risk management of power markets and the pricing of power financial derivatives. The paper proposes a GARCH-M model in which the multicycle properties of the power price series in the spot market are described by dummy variable and sine function. This model not only helps to select the order with fewer estimated parameters, but is also able to deal with the changing trend, muhicycles and nonlinear relationships between load and power price of the power price series simultaneously, therefore has value of practical application. The numerical example based on the historical data of the PJM market shows that the heteroscedasticity and the load squares of the price priced have a significant impact on the mean power prices, and there exist volatility clustering and weekly, semi-monthly, monthly, quarterly and semi-annual periods.