Text mining of causes of hot working accidents based on 24Model

doi:10.16265/j.cnki.issn1003-3033.2025.03.0757

Abstract

Abstract:

In order to systematically explore the root causes of industrial hot work accidents through a large amount of text data, a text mining method based on 24Model was proposed. Firstly, 220 hot work accident reports were collected and sorted as datasets, and a 24Model classifier based on Bidirectional Encoder Representations from Transformers (BERT) was constructed. The pre-trained model was used to train and evaluate the accident report dataset to construct a classification model. Then, through the combination weight of the Keyword extraction algorithm based on BERT (KeyBERT) and Term Frequency-Inverse Document Frequency (TF-IDF) algorithms, combined with the 24Model framework, a keyword index system for hot work accident text was established. Finally, the interrelationships between accident causes were obtained through the analysis of the network co-occurrence relationship between text-mining keywords. The results show that the BERT-based 24Model classifier model can systematically and accurately determine the causative categories of hot work accidents. The weight of the safety management system was the largest among the 4-level keyword index systems obtained through the combination of weights. Furthermore, 7 key causative factors of hot work accidents were obtained by combining them with the co-occurrence network analysis. This shows that 24Model can strengthen the interpretability of text mining results, which provides an important reference for the prevention and management of hot work accidents.

Key words: "2-4" model (24Model), hot work, accident causes, text mining, index system

CLC Number:

X928.02事故处理

NIU Maohui, LI Weijun, LIU Yin, WANG Lu. Text mining of causes of hot working accidents based on 24Model[J]. China Safety Science Journal, 2025, 35(3): 151-158.

Figures/Tables 11

Fig.1

Table 1

Fig.2

Table 2

Table 3

Fig.3

Table 4

Table 5

Fig.4

Table 6

Table 7

References 18

[1]	WANG Xin. Analysis and improvement of hot work management in China[J]. Process Safety Progress, 2023, 42(1):72-78.
[2]	山东省人民政府. 山东五彩龙投资有限公司栖霞市笏山金矿“1·10”重大爆炸事故调查报告[EB/OL]. (2021-02-23). http://www.shandong.gov.cn/art/2021/2/23/art_305296_10334792.html.
[3]	中华人民共和国应急管理部. 佳乐苑小区特别重大火灾事故调查报告[EB/OL]. (2024-09-21). https://www.mem.gov.cn/xw/bndt/202409/t20240921_501982.shtml.
[4]	李显, 焦宇, 陈文涛, 等. 船舶修造企业火灾事故特征挖掘与致因分析[J]. 中国安全科学学报, 2024, 34(3):29-38. doi: 10.16265/j.cnki.issn1003-3033.2024.03.0368
	LI Xian, JIAO Yu, CHEN Wentao, et al. Feature mining and causative analysis of fire accidents in ship repair and building enterprises[J]. China Safety Science Journal, 2024, 34(3):29-38. doi: 10.16265/j.cnki.issn1003-3033.2024.03.0368
[5]	董秉聿, 王宇昊, 孙鹏依, 等. 基于AcciMap模型的动火作业火灾事故案例分析[J]. 安全, 2023, 44(6):94-99.
	DONG Bingyu, WANG Yuhao, SUN Pengyi, et al. Case analysis on fire accident caused by fire operation based on AcciMap model[J]. Safety and Security, 2023, 44(6):94-99.
[6]	LI Weijun, SUN Qiqi, ZHANG Jiwang, et al. Quantitative risk assessment of industrial hot work using adaptive bow tie and Petri nets[J]. Reliability Engineering and System Safety, 2024,242: DOI: 10.1016/j.ress.2023.109784.
[7]	田水承, 黄权. 基于文本挖掘的煤矿瓦斯事故致因分析[J]. 煤矿安全, 2022, 53(5):241-245.
	TIAN Shuicheng, HUANG Quan. Analysis of causes for coal mine gas accidents based on text mining method[J]. Safety in Coal Mines, 2022, 53(5): 241-245.
[8]	XU Hui, LIU Yi, SHU Chimin, et al. Cause analysis of hot work accidents based on text mining and deep learning[J]. Journal of Loss Prevention in the Process Industries, 2022,76: DOI: 10.1016/j.jlp.2022.104747.
[9]	FU Gui, XIE Xuecai, JIA Qingsong, et al. The development history of accident causation models in the past 100 years: 24Model, a more modern accident causation model[J]. Process Safety and Environmental Protection, 2020,134:47-82.
[10]	傅贵, 陈奕燃, 许素睿, 等. 事故致因“2-4”模型的内涵解析及第6版的研究[J]. 中国安全科学学报, 2022, 32(1):12-19. doi: 10.16265/j.cnki.issn1003-3033.2022.01.002
	FU Gui, CHEN Yiran, XU Surui, et al. Detailed explanations of 24Model and development of its 6^th version[J]. China Safety Science Journal, 2022, 32(1):12-19. doi: 10.16265/j.cnki.issn1003-3033.2022.01.002
[11]	DEVLIN J, CHANG Mingwei, LEE K, et al. BERT: pre-training of deep bidirectional transformers for language understanding[C]. Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics:Human Language Technologies, 2019:4171-4186.
[12]	GARRIDO-MERCHAN E C, GOZALO-BRIZUELA R, GONZALEZ-CARVAJAl S. Comparing BERT against traditional machine learning models in text classification[J]. Journal of Computational and Cognitive Engineering, 2023, 2(4):352-356.
[13]	李思锐. 基于BERT的子词级中文文本分类方法[J]. 计算机科学与应用, 2020, 10(6):1075-1086.
	LI Sirui. Subword-level Chinese text classification method based on BERT[J]. Computer Science and Application, 2020, 10(6):1075-1086.
[14]	兰晓芳, 刘卓, 许志豪, 等. 基于TF-IDF和TextRank结合的中文文本关键词提取方法:以体育新闻为例[J]. 软件工程, 2023, 26(8):6-10.
	LAN Xiaofang, LIU Zhuo, XU Zhihao, et al. A Chinese text keyword extraction method based on the combination of TF-IDF and TextRank:a case study of sports news[J]. Software Engineering, 2023, 26(8):6-10.
[15]	郑彬彬, 冯婷婷, 王佳贺, 等. 基于文本挖掘的城镇燃气事故致因及关联分析[J]. 中国安全科学学报, 2023, 33(7):190-195. doi: 10.16265/j.cnki.issn1003-3033.2023.07.1644
	ZHENG Binbin, FENG Tingting, WANG Jiahe, et al. Causes and correlation analysis of urban gas accidents based on text mining[J]. China Safety Science Journal, 2023, 33(7):190-195. doi: 10.16265/j.cnki.issn1003-3033.2023.07.1644
[16]	BASTIAN M, HEYMANN S, JACOMY M. Gephi: an open source software for exploring and manipulating networks[C]. International AAAI Conference on Weblogs and Social Media, 2009: 361-362.
[17]	LIU Muyang, LUO Xiaowei, WANG Guangbin, et al. Intelligent information extraction from government on-site inspection reports of construction projects: a graph-based text mining approach[J]. Advanced Engineering Informatics, 2023,58: DOI: 10.1016/j.aei.2023.102163.
[18]	王亮, 赖佳燕, 张自欣, 等. 基于文本挖掘和改进DEMATEL法的化工事故关键因素识别[J]. 中国安全科学学报, 2024, 34(3):20-28. doi: 10.16265/j.cnki.issn1003-3033.2024.03.0230
	WANG Liang, LAI Jiayan, ZHANG Zixin, et al. Identification of critical factors in chemical accidents based on text mining and improved DEMATEL method[J]. China Safety Science Journal, 2024, 34(3):20-28. doi: 10.16265/j.cnki.issn1003-3033.2024.03.0230

层级要素	解释词
安全文化	文化建设、思想、责任、价值观、投入、态度、理念、方针、指导思想
安全管理体系	许可证、培训、监管、规程、风险管控、隐患排查、管理体系、标准、资质、规章、制度、岗位责任
安全知识	教育、技能、知识、能力、资格、经验
安全意识	不重视、疏忽、逃避、动机
安全习惯	多次违章、多次违反
安全心理	冒险、压力、紧张、情绪
安全生理	疲劳、迟钝、年龄、性别、体重
不安全动作	未正确使用、违法采用、违规操作、违反操作规程、整改不力、违章作业、违章指挥
不安全状态	条件、环境、防护用品、工具、设备、附件、安全防护措施、标志

层级要素	P	R	F₁
安全文化	0.750 0	0.750 0	0.750 0
安全管理体系	0.918 4	0.882 4	0.900 0
安全知识	1.000 0	0.800 0	0.888 9
安全意识	0.750 0	0.857 1	0.800 0
安全习惯	1.000 0	0.500 0	0.667 0
安全心理	0.833 0	1.000 0	0.909 1
安全生理	1.000 0	1.000 0	1.000 0
不安全动作	0.666 7	0.909 1	0.769 2
不安全状态	0.727 3	0.761 9	0.744 2
噪声文本	0.761 9	0.695 7	0.727 3
A	0.820 1

关键词	W_K	W_T	W
安全防护措施	0.074 9	0.028 1	0.051 5
安全管理	0.05 94	0.026 5	0.042 9
安全制度不完善	0.054 6	0.022 5	0.038 6
教育和培训	0.039 9	0.028 3	0.034 1
安全意识淡薄	0.051 0	0.016 3	0.033 6
动火作业审批	0.041 7	0.013 3	0.027 5
动火作业许可证	0.038 2	0.016 1	0.027 2
安全生产主体责任	0.036 0	0.012 1	0.024 1
现场监护不力	0.027 3	0.020 0	0.023 7
违规动火作业	0.032 2	0.011 4	0.021 8

层级要素	分层权重	关键词	归一化权重	层级要素	分层权重	关键词	归一化权重
A	0.075 3	A₁	0.070 0	—	—	D₅	0.002 4
		A₂	0.002 0	E	0.0142	E₁	0.008 4
		A₃	0.002 3			E₂	0.002 8
		A₄	0.001 0			E₃	0.003 0
B	0.480 0	B₁	0.088 2	F	0.001 8	F₁	0.001 8
		B₂	0.048 6	G	0.005 8	G₁	0.005 8
		B₃	0.055 9	H	0.128 2	H₁	0.044 8
		B₄	0.056 6			H₂	0.033 1
		B₅	0.049 5			H₃	0.010 6
		B₆	0.040 5			H₄	0.018 6
		B₇	0.079 3			H₅	0.008 9
		B₈	0.039 3			H₆	0.012 2
		B₉	0.022 1	I	0.186 0	I₁	0.105 9
C	0.017 6	C₁	0.010 2			I₂	0.024 0
		C₂	0.003 2			I₃	0.018 5
		C₃	0.002 1			I₄	0.014 4
		C₄	0.002 1			I₅	0.010 2
D	0.091 1	D₁	0.069 2			I₆	0.007 9
		D₂	0.007 3			I₇	0.003 6
		D₃	0.006 2			I₈	0.001 5
		D₄	0.006 0

关键词	C_D	C_C	C_B	关键词	C_D	C_C	C_B
B₈	40	1.00	22.25	H₂	36	0.91	9.91
A₁	39	0.98	18.20	H₃	32	0.83	3.99
I₁	39	0.98	18.73	H₆	32	0.83	5.00
B₁	40	1.00	22.25	H₅	32	0.83	5.11
B₇	40	1.00	22.25	H₄	34	0.87	6.54
B₉	38	0.95	18.03	I₅	28	0.77	2.44
B₆	34	0.87	8.15	E₁	32	0.83	4.62
B₂	36	0.91	12.54	D₃	25	0.73	0.61
B₄	35	0.89	9.06	I₄	24	0.71	1.21
B₃	34	0.87	8.01	C₂	28	0.77	1.84
B₅	32	0.83	3.62	D₄	20	0.67	1.31
I₂	35	0.89	11.46	C₁	25	0.73	0.74
D₁	38	0.95	16.99	D₅	22	0.69	0.28
I₃	34	0.87	9.87	E₃	25	0.73	0.96
H₁	32	0.83	6.96	D₂	28	0.77	2.18