电力负荷时间序列预测的应用与研究

时间序列预测是电力负荷预测的重要方法。但是它对天气、节日等影响电力负荷变化的因素不敏感,对非平稳序列的处理需要多次差分运算达到基本平稳后才能进一步建立预测模型。因此,提出基于属性分类的时间序列预测方案。该方案把电力负荷按影响因素进行分类,预测时按预测对象的属性来选取预测样本。基于属性分类的电力负荷时序预测方案把时刻、天气、节日等因素考虑到了预测过程中,弥补了电力负荷时序预测的缺陷。实验证明该方法提高了电力负荷时序预测的速度和准确度。

The Application and Research of Electric PowerLoad Forecasting Technology Based on the Time Series Model

Power load forecasting based on time series was an important way of power load forecasting,but it wasn't sensitive to weather,holidays and other factors that would change power load,and needed to do multiple differential operation when it processed of a non-stationary series.In order to resolve the problems,an improved method of power load forecasting based on time series is introduced.The new method will first assort power value based on properties,including weather,holidays and so on.Then sample time seri...