成分分解方法预测月度电力负荷

Monthly Load Forecasting Using Component Decomposition Method

为了提高月度负荷预测精度,提出了基于X-12-ARIMA季节调整模型的月度负荷预测方法。首先在季节调整前,消除原始负荷中离群值、工作日、闰年等效应的影响,然后对经季节调整后的趋势循环序列应用H-P滤波方法进行成分分解,再针对分解后得到的长期趋势、循环周期、季节因子、不规则成分序列的特点选择了适合的预测模型进行预测并得到最终结果。通过甘肃地区188个月的负荷数据进行检验,结果表明该预测方法是可靠有效的。

In this paper, a monthly load forecasting method based on X-12-ARIMA model with seasonal adjustment is proposed to improve the performance of load forecasting. Firstly, the effects of factors such as outliers, work days and leap years are eliminated from the original load data. Secondly, H-P filtering method is used to decompose the sequenc- es of trend and cycle obtained after seasonal adjustment. According to the characteristics of the component sequences of long term trend, cycle, seasonal factors and irregularity, an appropriate forecasting model is selected, and the final result is obtained. Through an empirical test on the load data in Gansu area for 188 months, it is indicated that the pro-posed method is reliable and effective.