Multivariate load forecasting of integrated energy system based on GBLA

Pei He, Xiaodong Wang, Yangming Guo

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

There may be complex and strong coupling relationship between various loads in the integrated energy system. Compared with the single and independent forecasting of various loads, the direct development of multivariate load forecasting can further explore the internal relations between loads and improve the forecasting accuracy. This paper presents a prediction model based on GRA-Bi-LSTM-Attention(GBLA). GRA was used to quantitatively analyze the coupling between the influencing factors. Bi-LSTM was used to capture the nonlinear relationship in the multi-load time series and enhance the short-term memory ability. The attention mechanism was introduced to distribute the weight of the prediction results, so as to realize the joint prediction of multi-load. Finally, based on the load data set of typical integrated energy systems, the validation work is carried out, and the comparison analysis is made with other prediction models. The results show that the prediction method proposed in this paper has better performance.

Original languageEnglish
Title of host publication2023 IEEE 2nd Industrial Electronics Society Annual On-Line Conference, ONCON 2023
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350357974
DOIs
StatePublished - 2023
Event2nd IEEE Industrial Electronics Society Annual On-Line Conference, ONCON 2023 - Virtual, Online, United States
Duration: 8 Dec 202310 Dec 2023

Publication series

Name2023 IEEE 2nd Industrial Electronics Society Annual On-Line Conference, ONCON 2023

Conference

Conference2nd IEEE Industrial Electronics Society Annual On-Line Conference, ONCON 2023
Country/TerritoryUnited States
CityVirtual, Online
Period8/12/2310/12/23

Keywords

  • Attention
  • Bi-LSTM
  • component
  • GRA
  • integrated energy system
  • load forecasting

Fingerprint

Dive into the research topics of 'Multivariate load forecasting of integrated energy system based on GBLA'. Together they form a unique fingerprint.

Cite this