Towards an Explainable Artificial Intelligence approach for smart grid systems

Abstract As global energy demands escalate, effective management of electrical grids and reducing carbon emissions have become critical objectives. This paper proposes a novel system which employs Explainable Artificial Intelligence (XAI) to enhance the operational efficiency of smart grids by predi...

Full description

Saved in:
Bibliographic Details
Main Authors: Mahmoud Alfayan, Hani Hagras
Format: Article
Language:English
Published: Springer 2025-04-01
Series:Discover Artificial Intelligence
Online Access:https://doi.org/10.1007/s44163-025-00261-5
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850139110472155136
author Mahmoud Alfayan
Hani Hagras
author_facet Mahmoud Alfayan
Hani Hagras
author_sort Mahmoud Alfayan
collection DOAJ
description Abstract As global energy demands escalate, effective management of electrical grids and reducing carbon emissions have become critical objectives. This paper proposes a novel system which employs Explainable Artificial Intelligence (XAI) to enhance the operational efficiency of smart grids by predicting energy consumption and optimising resource allocation accordingly. The integration of XAI addresses the complexities of grid management by providing transparency into AI-driven predictions, thus fostering user trust and facilitating informed decision-making. The main contributions of this work include evaluating the status of smart energy grid management and presenting a pathway for integrating XAI to enhance these practices. Using a real-world dataset with variables related to operations, environment, and time, our system employs a type-2 fuzzy logic XAI based system to generate clear, interpretable predictions of energy demand. The transition from a Type-1 to a Type-2 fuzzy logic system resulted in enhanced prediction performance, as evidenced by a reduction in the root mean square error (RMSE) from 8.7734 to 5.9422, resulting in 32.2% enhancement in RMSE. The proposed system demonstrated comparable performance compared to conventional black-box models, including neural network. It also shows stakeholders how to interpret these predictions across a wide range of consumption scenarios, which emphasizes the importance of predictions in making the best use of resources. Ultimately, this paper presents a first step for the employment of XAI within smart grid management.
format Article
id doaj-art-a544c50f968f48edbf60ded9f8b54f45
institution OA Journals
issn 2731-0809
language English
publishDate 2025-04-01
publisher Springer
record_format Article
series Discover Artificial Intelligence
spelling doaj-art-a544c50f968f48edbf60ded9f8b54f452025-08-20T02:30:24ZengSpringerDiscover Artificial Intelligence2731-08092025-04-015112010.1007/s44163-025-00261-5Towards an Explainable Artificial Intelligence approach for smart grid systemsMahmoud Alfayan0Hani Hagras1School of Computer Science and Electronic Engineering, University of EssexSchool of Computer Science and Electronic Engineering, University of EssexAbstract As global energy demands escalate, effective management of electrical grids and reducing carbon emissions have become critical objectives. This paper proposes a novel system which employs Explainable Artificial Intelligence (XAI) to enhance the operational efficiency of smart grids by predicting energy consumption and optimising resource allocation accordingly. The integration of XAI addresses the complexities of grid management by providing transparency into AI-driven predictions, thus fostering user trust and facilitating informed decision-making. The main contributions of this work include evaluating the status of smart energy grid management and presenting a pathway for integrating XAI to enhance these practices. Using a real-world dataset with variables related to operations, environment, and time, our system employs a type-2 fuzzy logic XAI based system to generate clear, interpretable predictions of energy demand. The transition from a Type-1 to a Type-2 fuzzy logic system resulted in enhanced prediction performance, as evidenced by a reduction in the root mean square error (RMSE) from 8.7734 to 5.9422, resulting in 32.2% enhancement in RMSE. The proposed system demonstrated comparable performance compared to conventional black-box models, including neural network. It also shows stakeholders how to interpret these predictions across a wide range of consumption scenarios, which emphasizes the importance of predictions in making the best use of resources. Ultimately, this paper presents a first step for the employment of XAI within smart grid management.https://doi.org/10.1007/s44163-025-00261-5
spellingShingle Mahmoud Alfayan
Hani Hagras
Towards an Explainable Artificial Intelligence approach for smart grid systems
Discover Artificial Intelligence
title Towards an Explainable Artificial Intelligence approach for smart grid systems
title_full Towards an Explainable Artificial Intelligence approach for smart grid systems
title_fullStr Towards an Explainable Artificial Intelligence approach for smart grid systems
title_full_unstemmed Towards an Explainable Artificial Intelligence approach for smart grid systems
title_short Towards an Explainable Artificial Intelligence approach for smart grid systems
title_sort towards an explainable artificial intelligence approach for smart grid systems
url https://doi.org/10.1007/s44163-025-00261-5
work_keys_str_mv AT mahmoudalfayan towardsanexplainableartificialintelligenceapproachforsmartgridsystems
AT hanihagras towardsanexplainableartificialintelligenceapproachforsmartgridsystems