基于时间序列缺失数据的GAM短期电力负荷预测

GAM Short-Term Electricity Load Forecasting Based on Time Series Missing Data

可靠、准确的区域用电负荷预测对于电力系统的运行和规划至关重要,短期负荷预测有助于经济运行、调度电力。考虑到时间序列数据从采集、传输到存储的过程中可能出现数据缺失这一情况,为此本文基于广义相加模型,建立短期电力负荷预测模型。首先对数据进行规范化和数据插补预处理,然后在GAM模型中调用合适的平滑函数、设置最佳参数,在R平台上对时间序列数据进行短期负荷预测,最后实验得出短期内的用电高峰和低谷时段。

Reliable and accurate regional load forecasting is very important for power system operation and planning.Short-term load forecasting is helpful for economic operation and power dispatching.Considering that the time series data may be missing in the process of collection,transmission and storage,this paper establishes a short-term power load forecasting model based on generalized additive model.Firstly,data normalization and data interpolation are preprocessed,and then appropriate smoothing function is called in GAM model and the optimal parameters are set.Short-term load prediction is carried out for time series data on R platform.Finally,the peak and trough periods of short-term electricity consumption are obtained experimentally.