基于频域分解的短期负荷预测

Short-term Load Forecasting Based on Frequency Domain Decomposition

针对负荷本身存在随机性和间歇性,提出一种基于频域分解的负荷预测。首先采用频域分解算法将原始负荷分解为日周期和周周期分量;其次分别采用置信度理论和指数平滑法对日周期分量和周周期分量进行预测;最后将负荷的日周期分量和周周期分量重组,实现短期负荷较为准确的预测。根据某地市的2018年国庆期间的负荷数据进行仿真,证明了预测模型的精确度。

In view of the randomness and intermittency of the load itself,this paper presents a short-term load forecasting strategy based on frequency domain decomposition method.Firstly,the original load is decomposed into daily periodic component and weekly periodic component by frequency domain decomposition algorithm.Secondly,the confidence theory and exponential smoothing method are used to predict the daily and weekly periodic components respectively.Finally,the daily and weekly periodic components of load are recombined to achieve more accurate short-term load forecasting.The simulation is based on the load data of the seven days of the 2018 National Day in a certain city,which proves the accuracy of the prediction model.