基于深度学习的家庭电力短期负荷预测

Short Term Load Forecasting of Household Power Based on Deep Learning

为了更好地实现电力利用率最大化,短期电力负荷预测受到了广泛研究,这些研究为智能电网系统和家庭能源成本的降低提供了可能性。但目前的研究大多集中在预测整个家庭的负荷水平上。文章作者提出了一种新的方法来预测单个电子设备的负载。提出的方法使用了一个具有长—短期记忆的递归深层神经网络。实验结果表明,提出的方法具有较好性能,能够满足家庭日常需求。

In order to better maximize power utilization, short-term power load forecasting has been extensively studied. These studies provide the possibility of reducing smart grid systems and household energy costs. However, most of the current researches focus on predicting the load level of the entire family. Therefore, in this article, a new method is proposed to predict the load of a single electronic device. The proposed method uses a recursive deep neural network with long-short-term memory. Experimental results show that the proposed method has good performance and can meet the daily needs of families.