考虑外生变量的广义自回归条件异方差日前电价预测模型

Exogenous Variable Considered Generalized Autoregressive Conditional Heteroscedastic Model for Day-Ahead Electricity Price Forecasting

利用广义自回归条件异方差模型预测电价,并在该模型中引入周用电比率作为外生变量,以增加模型对外界影响的响应。采用上述方法对美国PJM电力市场2004年12月份的日前电价进行预测,结果表明该方法对高峰时段电价的预测精度明显高于与之对比的其他模型,整体预测精度也好于对比模型。

The authors forecast day-ahead electricity price by generalized autoregressive conditional heteroscedasticity （GARCH） model and as exogenous variable the daily price ratios of different weekday types are led into this GARCH mode to intensify the response of the proposed model to external influences. Using the proposed model, the day-ahead electricity prices of PJM （Pennsylvania-New Jersey-Maryland） electricity market in December 2004 are forecasted. Forecasting results show that the forecasted day-ahead electricity prices in peak hours are evidently better than those by other models taken for contrasts, and its global forecasting precision is also better than those by other models.