SafetyGPT: An autonomous agent of electrical safety risks for monitoring workers’ unsafe behaviors

Workers’ unsafe behavior is one of the major causes of accidents in electric power production. Intelligent monitoring of workers’ unsafe behaviors can effectively prevent the expansion of safety risks, thereby blocking the development process of risks to accidents. Electric power production processe...

Full description

Saved in:
Bibliographic Details
Main Authors: Wei Li, Fuqi Ma, Zhiyuan Zuo, Rong Jia, Bo Wang, Abdullah M Alharbi
Format: Article
Language:English
Published: Elsevier 2025-07-01
Series:International Journal of Electrical Power & Energy Systems
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S0142061525002236
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849321712177381376
author Wei Li
Fuqi Ma
Zhiyuan Zuo
Rong Jia
Bo Wang
Abdullah M Alharbi
author_facet Wei Li
Fuqi Ma
Zhiyuan Zuo
Rong Jia
Bo Wang
Abdullah M Alharbi
author_sort Wei Li
collection DOAJ
description Workers’ unsafe behavior is one of the major causes of accidents in electric power production. Intelligent monitoring of workers’ unsafe behaviors can effectively prevent the expansion of safety risks, thereby blocking the development process of risks to accidents. Electric power production processes are diverse in nature and require the frequent switching of operating scenarios. This makes it difficult to identify what is “unsafe” since worker behaviors within the given electrical context also exhibit variability and diversity. Existing methods have insufficient generalization and adaptability, which makes them inadequate for the case of electric power production. Therefore, this paper proposes Safety Generative Pre-trained Transformers (SafetyGPT), an autonomous agent of safety risk based on a multi-modal large language model, which incorporates a human–machine collaborative monitoring mode for unsafe behaviors of workers. SafetyGPT loads the electric power production video, and the backend supervisors set instructions for SafetyGPT based on task requirements. The model encodes visual and textual features into corresponding tokens, realizes multi-modal feature alignment and fusion through the cross-attention mechanism, and then generates targeted responses through the large language model. Next, the proposed method is applied to real production site data to confirm the effectiveness and superiority through comparison with other methods designed to identify unsafe behaviors. Experimental results show that the accuracy of the proposed method for the identification of unsafe behaviors in complex environments is 96.5%, and that it can generate reasonable recommended plan based on the identification results, assist backend supervisors in making decisions, and effectively improve the safety level of power production.
format Article
id doaj-art-6a39e45e8df54656b3f8c18ee02fc283
institution Kabale University
issn 0142-0615
language English
publishDate 2025-07-01
publisher Elsevier
record_format Article
series International Journal of Electrical Power & Energy Systems
spelling doaj-art-6a39e45e8df54656b3f8c18ee02fc2832025-08-20T03:49:41ZengElsevierInternational Journal of Electrical Power & Energy Systems0142-06152025-07-0116811067210.1016/j.ijepes.2025.110672SafetyGPT: An autonomous agent of electrical safety risks for monitoring workers’ unsafe behaviorsWei Li0Fuqi Ma1Zhiyuan Zuo2Rong Jia3Bo Wang4Abdullah M Alharbi5School of Electrical Engineering, Xi’an University of Technology, Xi’an, Shaanxi 710048, ChinaSchool of Electrical Engineering, Xi’an University of Technology, Xi’an, Shaanxi 710048, China; School of Electrical Engineering and Automation, Wuhan University, Wuhan, Hubei 430072, China; Corresponding author at: Xi’an University of Technology, Yanxiang Road No. 58, Xi’an, Shaanxi Province 710054, China.School of Electrical and Electronic Engineering, The University of Manchester, Manchester M13 9PL, United KingdomSchool of Electrical Engineering, Xi’an University of Technology, Xi’an, Shaanxi 710048, ChinaSchool of Electrical Engineering and Automation, Wuhan University, Wuhan, Hubei 430072, ChinaDepartment of Electrical Engineering, College of Engineering in Wadi Addawasir, Prince Sattam bin Abdulaziz University, Wadi Addawasir 11991, Saudi ArabiaWorkers’ unsafe behavior is one of the major causes of accidents in electric power production. Intelligent monitoring of workers’ unsafe behaviors can effectively prevent the expansion of safety risks, thereby blocking the development process of risks to accidents. Electric power production processes are diverse in nature and require the frequent switching of operating scenarios. This makes it difficult to identify what is “unsafe” since worker behaviors within the given electrical context also exhibit variability and diversity. Existing methods have insufficient generalization and adaptability, which makes them inadequate for the case of electric power production. Therefore, this paper proposes Safety Generative Pre-trained Transformers (SafetyGPT), an autonomous agent of safety risk based on a multi-modal large language model, which incorporates a human–machine collaborative monitoring mode for unsafe behaviors of workers. SafetyGPT loads the electric power production video, and the backend supervisors set instructions for SafetyGPT based on task requirements. The model encodes visual and textual features into corresponding tokens, realizes multi-modal feature alignment and fusion through the cross-attention mechanism, and then generates targeted responses through the large language model. Next, the proposed method is applied to real production site data to confirm the effectiveness and superiority through comparison with other methods designed to identify unsafe behaviors. Experimental results show that the accuracy of the proposed method for the identification of unsafe behaviors in complex environments is 96.5%, and that it can generate reasonable recommended plan based on the identification results, assist backend supervisors in making decisions, and effectively improve the safety level of power production.http://www.sciencedirect.com/science/article/pii/S0142061525002236Generative artificial intelligenceRisk identificationElectric power productionMulti-modal large language modelUnsafe behaviors
spellingShingle Wei Li
Fuqi Ma
Zhiyuan Zuo
Rong Jia
Bo Wang
Abdullah M Alharbi
SafetyGPT: An autonomous agent of electrical safety risks for monitoring workers’ unsafe behaviors
International Journal of Electrical Power & Energy Systems
Generative artificial intelligence
Risk identification
Electric power production
Multi-modal large language model
Unsafe behaviors
title SafetyGPT: An autonomous agent of electrical safety risks for monitoring workers’ unsafe behaviors
title_full SafetyGPT: An autonomous agent of electrical safety risks for monitoring workers’ unsafe behaviors
title_fullStr SafetyGPT: An autonomous agent of electrical safety risks for monitoring workers’ unsafe behaviors
title_full_unstemmed SafetyGPT: An autonomous agent of electrical safety risks for monitoring workers’ unsafe behaviors
title_short SafetyGPT: An autonomous agent of electrical safety risks for monitoring workers’ unsafe behaviors
title_sort safetygpt an autonomous agent of electrical safety risks for monitoring workers unsafe behaviors
topic Generative artificial intelligence
Risk identification
Electric power production
Multi-modal large language model
Unsafe behaviors
url http://www.sciencedirect.com/science/article/pii/S0142061525002236
work_keys_str_mv AT weili safetygptanautonomousagentofelectricalsafetyrisksformonitoringworkersunsafebehaviors
AT fuqima safetygptanautonomousagentofelectricalsafetyrisksformonitoringworkersunsafebehaviors
AT zhiyuanzuo safetygptanautonomousagentofelectricalsafetyrisksformonitoringworkersunsafebehaviors
AT rongjia safetygptanautonomousagentofelectricalsafetyrisksformonitoringworkersunsafebehaviors
AT bowang safetygptanautonomousagentofelectricalsafetyrisksformonitoringworkersunsafebehaviors
AT abdullahmalharbi safetygptanautonomousagentofelectricalsafetyrisksformonitoringworkersunsafebehaviors