Edge computing-enabled energy efficiency prediction of immersion cooling system for supercomputing centers

Within the context of cleaner production, enhancing energy efficiency and sustainability has emerged as a central focus in supercomputing center development. To address the challenges in predicting energy consumption, this study proposes an architecture of edge computing-enabled energy efficiency pr...

Full description

Saved in:
Bibliographic Details
Main Authors: Shuaiyin Ma, Yichun Cao, Yang Liu
Format: Article
Language:English
Published: Elsevier 2025-07-01
Series:Cleaner Engineering and Technology
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2666790825001375
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849724250805501952
author Shuaiyin Ma
Yichun Cao
Yang Liu
author_facet Shuaiyin Ma
Yichun Cao
Yang Liu
author_sort Shuaiyin Ma
collection DOAJ
description Within the context of cleaner production, enhancing energy efficiency and sustainability has emerged as a central focus in supercomputing center development. To address the challenges in predicting energy consumption, this study proposes an architecture of edge computing-enabled energy efficiency prediction of immersion cooling system for supercomputing centers. This architecture combines time-series generative adversarial network (TimeGAN) for data augmentation with the neural basis expansion analysis for time series (N-BEATS) for prediction, providing a robust solution for accurate energy consumption prediction. TimeGAN enhances the training dataset by generating high-quality synthetic time-series data, effectively mitigating issues of data sparsity and imbalance. N-BEATS, with its modular architecture and strong adaptability to temporal data, ensures precise predictions by capturing both global trends and local variations in energy usage. Experimental results demonstrate the effectiveness of the proposed architecture, exhibiting superior performance across key metrics when compared to traditional models. Specifically, the RMSE of TimeGAN-N-BEATS reduced by more than 8 %, the MSE decreased by over 18 %, and the R2 reached 97.31 %, outperforming baseline models such as long short-term memory and gated recurrent unit, and their attention-enhanced variants. This study highlights the potential of integrating generative and predictive models to optimize energy efficiency in liquid cooling systems, offering valuable insights for sustainable supercomputing operations.
format Article
id doaj-art-d423c2c7109846a5b3a82c0776a863fa
institution DOAJ
issn 2666-7908
language English
publishDate 2025-07-01
publisher Elsevier
record_format Article
series Cleaner Engineering and Technology
spelling doaj-art-d423c2c7109846a5b3a82c0776a863fa2025-08-20T03:10:47ZengElsevierCleaner Engineering and Technology2666-79082025-07-012710101410.1016/j.clet.2025.101014Edge computing-enabled energy efficiency prediction of immersion cooling system for supercomputing centersShuaiyin Ma0Yichun Cao1Yang Liu2School of Computer Science and Technology, Xi'an University of Posts and Telecommunications, Xi'an, 710121, China; Shaanxi Key Laboratory of Network Data Analysis and Intelligent Processing, Xi'an University of Posts and Telecommunications, Xi'an, 710121, China; Xi'an Key Laboratory of Big Data and Intelligent Computing, Xi'an, 710121, China; Shaanxi Union Research Centre of University and Enterprise for 5G+ Industrial Internet Communication Terminal Technology, Xi'an University of Posts and Telecommunications, Xi'an, 710121, China; Corresponding author. School of Computer Science and Technology, Xi'an University of Posts and Telecommunications, Xi'an, 710121, China.School of Computer Science and Technology, Xi'an University of Posts and Telecommunications, Xi'an, 710121, ChinaDepartment of Management and Engineering, Linköping University, SE-581 83 Linköping, Sweden; Industrial Engineering and Management, University of Oulu, 90570 Oulu, Finland; Corresponding author. Department of Management and Engineering, Linköping University, SE-581 83 Linköping, Sweden.Within the context of cleaner production, enhancing energy efficiency and sustainability has emerged as a central focus in supercomputing center development. To address the challenges in predicting energy consumption, this study proposes an architecture of edge computing-enabled energy efficiency prediction of immersion cooling system for supercomputing centers. This architecture combines time-series generative adversarial network (TimeGAN) for data augmentation with the neural basis expansion analysis for time series (N-BEATS) for prediction, providing a robust solution for accurate energy consumption prediction. TimeGAN enhances the training dataset by generating high-quality synthetic time-series data, effectively mitigating issues of data sparsity and imbalance. N-BEATS, with its modular architecture and strong adaptability to temporal data, ensures precise predictions by capturing both global trends and local variations in energy usage. Experimental results demonstrate the effectiveness of the proposed architecture, exhibiting superior performance across key metrics when compared to traditional models. Specifically, the RMSE of TimeGAN-N-BEATS reduced by more than 8 %, the MSE decreased by over 18 %, and the R2 reached 97.31 %, outperforming baseline models such as long short-term memory and gated recurrent unit, and their attention-enhanced variants. This study highlights the potential of integrating generative and predictive models to optimize energy efficiency in liquid cooling systems, offering valuable insights for sustainable supercomputing operations.http://www.sciencedirect.com/science/article/pii/S2666790825001375Edge computingLiquid cooling systemN-BEATSSupercomputing centersTimeGAN
spellingShingle Shuaiyin Ma
Yichun Cao
Yang Liu
Edge computing-enabled energy efficiency prediction of immersion cooling system for supercomputing centers
Cleaner Engineering and Technology
Edge computing
Liquid cooling system
N-BEATS
Supercomputing centers
TimeGAN
title Edge computing-enabled energy efficiency prediction of immersion cooling system for supercomputing centers
title_full Edge computing-enabled energy efficiency prediction of immersion cooling system for supercomputing centers
title_fullStr Edge computing-enabled energy efficiency prediction of immersion cooling system for supercomputing centers
title_full_unstemmed Edge computing-enabled energy efficiency prediction of immersion cooling system for supercomputing centers
title_short Edge computing-enabled energy efficiency prediction of immersion cooling system for supercomputing centers
title_sort edge computing enabled energy efficiency prediction of immersion cooling system for supercomputing centers
topic Edge computing
Liquid cooling system
N-BEATS
Supercomputing centers
TimeGAN
url http://www.sciencedirect.com/science/article/pii/S2666790825001375
work_keys_str_mv AT shuaiyinma edgecomputingenabledenergyefficiencypredictionofimmersioncoolingsystemforsupercomputingcenters
AT yichuncao edgecomputingenabledenergyefficiencypredictionofimmersioncoolingsystemforsupercomputingcenters
AT yangliu edgecomputingenabledenergyefficiencypredictionofimmersioncoolingsystemforsupercomputingcenters