基于PCC LSTM MTL的综合能源系统多元负荷预测

Multi-energy Load Forecasting for Integrated Energy Systems Based on PCC LSTM MTL

为了保证综合能源系统(IES)的运行效率和可靠性,能源需求的准确预测至关重要。提出了一种基于Pearson相关系数(PCC),长短期记忆(LSTM)神经网络和多任务学习(MTL)的多元负荷预测方法。首先,运用PCC选取与冷热电负荷相关性较大的影响因素作为模型的输入;然后,通过LSTM建立MTL的共享层,实现多元负荷的联合预测;最后,结合亚利桑那州立大学的实测多元负荷数据来测试所提模型的预测精度。结果表明:所提模型具有更高的预测精度。

To ensure the operational efficiency and reliability of the Integrated Energy System(IES),accurate forecasting of energy demand has become a crucial task.Therefore,this paper proposes a multi-energy load prediction method based on Pearson Correlation Coefficient(PCC),Long and Short-Term Memory(LSTM)neural network,and Multi-Task Learning(MTL).First,the PCC is used to select the influencing factors that have a greater correlation with the cooling,heat,and electrical load as the input of the model.Then a shared layer of multi-task learning is established through the LSTM to realize the prediction of multi-energy load.Finally,combined with the measured multivariate load data of Arizona State University,the prediction accuracy of the proposed model is tested.The results show that the proposed model has higher prediction accuracy.