基于LSTM网络的短期负荷预测

Short-term Load Forecasting Based on LSTM Network

提出一种基于深度学习的LSTM的短期电力负荷预测方法,以减小负载预测模型所需的数据量的维度大小。基于某地电力公司的历史负荷数据,对比LSTM神经网络预测结果与BP神经网络算法预测结果,验证了所提算法的预测精度更高。

A short-term power load forecasting method based on the deep learning Long Short-Term Memory(LSTM)is proposed to reduce the dimension size of the data volume required by the load forecasting model.Historical load data from a local power company is used for simulation experiments.The forecasting results of LSTM neural network are compared with the forecasting results of BP neural network algorithm,and the forecasting accuracy of the proposed algorithm is higher.