Multi energy load forecasting for integrated energy systems

Abstract

The integrated energy system (IES) has developed rapidly as a form of energy management that realizes the multi-energy complementary characteristics of the energy Internet. IES has gradually evolved into one that integrates renewable energy, electric energy, and thermal. The spatial coupling characteristics of users' multiple energy sources and the randomness and variability in the time series make it more difficult to predict the multi-energy time series accurately. To address the challenge of accurately predicting multi-energy time series for a large number of users in integrated energy systems, this paper presents a novel analytical approach that considers multi-dimensional data coupling. Leveraging the Autoformer model, the research proposes a method for quantitatively analyzing the coupling characteristics of multi-energy time series, thereby extracting the energy consumption behavior of multienergy users. An iterative training method is employed to adjust the model parameters adaptively, with the ultimate objective of providing a comprehensive prediction and theoretical basis for the composition of the user's multi-energy time series prediction model