基于LSTM的多变量电力负荷短期预测

Multivariate Electricity Load Short-term Forecasting Based on LSTM

电力系统负荷预测是电力系统规划与调度的一项重要内容。若电力负荷预测准确,短期内可为电网内部机组启停、调度、运营提供参考。文章基于华南某城市的用电负荷相关数据,使用LSTM深度学习模型进行多变量预测,并在实证分析中对线性回归、决策树和xgboost三种机器学习算法进行比较,验证了该模型的准确性。

Power system load forecasting is an important part of power system planning and dispatching. If the electricity load forecasting is accurate, it can provide a reference for unit startup, shutdown, dispatching and operation in the power grid in the short term.Based on the power load related data of a city in South China, this paper uses LSTM deep learning model to make multivariable prediction,and compares three machine learning algorithms of linear regression, decision tree and xgboost in the empirical analysis to verify the accuracy of this model.