基于改进灰狼优化BP网络的城中村火灾预测

doi:10.16265/j.cnki.issn1003-3033.2025.08.1426

摘要/Abstract

摘要： 为了预防城中村火灾,利用改进灰狼优化算法(IGWO)和反向传播(BP)神经网络,对城中村火灾风险进行预测。引入非线性收敛因子和变异算子,改进传统灰狼优化算法(GWO),提高算法的全局搜索能力、收敛速度和稳定性,进而构建基于IGWO优化BP神经网络的城中村火灾风险预测模型(IGWO-BP),结合城中村火灾风险因素的复杂性和特殊性制定指标体系,预测火灾风险,并进行实例验证。结果表明:相较于传统GWO、粒子群算法(PSO)、长城算法(GWCA),IGWO在全局搜索能力、收敛速度和稳定性等方面均有显著提升,IGWO-BP模型可通过处理城中村火灾风险指标,实现对火灾风险的预测。

关键词: 改进灰狼优化算法(IGWO), 反向传播(BP)神经网络, 城中村火灾, 风险预测, 变异算子, 高维函数

Abstract:

In order to prevent fires in urban villages, IGWO and BP neural network were used to predict the risk of fires in urban villages. By introducing nonlinear convergence factors and mutation operators, the traditional grey wolf optimizer (GWO) was improved to enhance its global search capability, convergence speed, and stability. Furthermore, a fire risk prediction model for urban villages based on IGWO optimized BP neural network (IGWO-BP) was constructed. Taking into account the complexity and specificity of urban village fire risk factors, an indicator system was developed to predict fire risk, and an empirical study was conducted for verification. The results show that IGWO has significantly improved global search ability, convergence speed, and stability compared to traditional GWO, particle swarm optimization (PSO), and the Great Wall construction algorithm (GWCA). The IGWO-BP model can predict fire risk in urban villages by processing fire risk indicators.

Key words: improved grey wolf optimizer (IGWO), back propagation (BP) neural network, urban villages fire, risk prediction, mutation operator, high-dimensional function

中图分类号:

X913.4安全系统工程

吕淑然, 田江雪, 党鑫宇. 基于改进灰狼优化BP网络的城中村火灾预测[J]. 中国安全科学学报, 2025, 35(8): 196-204.

LYU Shuran, TIAN Jiangxue, DANG Xinyu. Fire prediction in urban villages based on improved grey wolf optimized BP network[J]. China Safety Science Journal, 2025, 35(8): 196-204.

图/表 11

表1

表2

图1

表3

图2

表4

表5

图3

图4

图5

图6

参考文献 34

[1]	张旭. 城中村火灾的防控难点及防控建议[J]. 今日消防, 2023, 8(12):100-102.
	ZHANG Xu. Difficulties and suggestions of fire prevention and control in urban villages[J]. Fire Protection Today, 2023, 8(12):100-102.
[2]	罗捷. 城中村火灾事故应急救援研究[J]. 建筑安全, 2023, 38(11):43-47.
[3]	黄建明. 加强城中村火灾风险管理的具体对策建议:以珠海为例[J]. 今日消防, 2023, 8(12):75-78.
	HUANG Jianming. Specific countermeasures and suggestions for strengthening fire risk management in urban villages :taking Zhuhai as an example[J]. Fire Protection Today, 2023, 8(12):75-78.
[4]	周培桂, 杨勇彪. 新时期城中村消防安全治理困境与纾解策略:基于潮州市湘桥区凤新街道的调研[J]. 城市与减灾, 2023(5):23-27.
[5]	李进. 当前城中村消防安全与火灾扑救探析[J]. 消防界:电子版, 2023, 9(22):99-101.
[6]	路世昌, 谢天光, 郭伟. 城中村消防安全问题及对策调研分析研究[J]. 中国应急管理, 2024(2):61-65.
[7]	赵文浩. 城中村火灾的特点及防控对策[J]. 今日消防, 2022, 7(12):73-75.
	ZHAO Wenhao. Characteristics and countermeasures of fires in urban villages[J]. Fire Protection Today, 2022, 7(12):73-75.
[8]	雷兰兰. “城中村”火灾的危险因素及其针对性防控对策研究[J]. 今日消防, 2022, 7(9):74-76.
	LEI Lanlan. Research on risk factors and prevention and control countermeasures of urban village fire[J]. Fire Protection Today, 2022, 7(9):74-76.
[9]	李庆渝, 卢荣昆, 杨旭, 等. "城中村"聚集区消防安全治理对策研究:以昆明市为例[J]. 中国应急救援, 2022(4):58-61.
	LI Qingyu, LU Rongkun, YANG Xu, et al. Research on fire safety management in village-in-city area: an empirical study of Kunming city[J]. China Emergency Rescue, 2022(4):58-61.
[10]	YANG Shunfeng, LI He, FAN Yufeng. Research on fire risk level prediction based on XGBoost algorithm[J]. 2022 International Conference on Artificial Intelligence and Autonomous Robot Systems (AIARS), 2022:73-77.
[11]	HERNÁNDEZ-LÓPEZ D, LÓPEZ-REBOLLO J, MORENO M A, et al. Automatic processing for identification of forest fire risk areas along high-voltage power lines using coarse-to-fine LiDAR data[J]. Forests, 2023, 14(4):662-681.
[12]	SALABAR L G F, ROMÃO X, PAUPÉRIO E. Review of vulnerability indicators for fire risk assessment in cultural heritage[J]. International Journal of Disaster Risk Reduction, 2021, 60:DOI:10.1016/j.ijdrr.2021.102286.
[13]	王启睿, 姜学鹏, 薛光桥, 等. 我国公路隧道火灾社会风险可接受标准研究[J]. 中国安全科学学报, 2024, 34(4):121-127. doi: 10.16265/j.cnki.issn1003-3033.2024.04.0687
	WANG Qirui, JIANG Xuepeng, XUE Guangqiao, et al. Study on social risk acceptance criteria of road tunnel fire in China[J]. China Safety Science Journal, 2024, 34(4):121-127. doi: 10.16265/j.cnki.issn1003-3033.2024.04.0687
[14]	张术琳, 张亚楠, 田超, 等. 基于CNN的化工园区火灾火焰图像识别研究[J]. 中国安全科学学报, 2024, 34(1):179-186. doi: 10.16265/j.cnki.issn1003-3033.2024.01.2333
	ZHANG Shulin, ZHANG Ya'nan, TIAN Chao, et al. Study on flame image recognition of chemical industrial park fires based on convolutional neural network[J]. China Safety Science Journal, 2024, 34(1):179-186. doi: 10.16265/j.cnki.issn1003-3033.2024.01.2333
[15]	刘勇, 李肖, 施式亮, 等. 基于多项Logistic的危化品槽罐车火灾类型判别[J]. 中国安全科学学报, 2024, 34(1):231-237. doi: 10.16265/j.cnki.issn1003-3033.2024.01.1871
	LIU Yong, LI Xiao, SHI Shiliang, et al. Fire type discrimination of hazardous chemicals tanker based on multinomial Logistic[J]. China Safety Science Journal, 2024, 34(1):231-237. doi: 10.16265/j.cnki.issn1003-3033.2024.01.1871
[16]	王新迪, 杨夙, 张思源, 等. 基于时空大数据与卫星图像的城市火灾风险预测[J]. 计算机工程, 2023, 49(6):242-249. doi: 10.19678/j.issn.1000-3428.0064765
	WANG Xindi, YANG Su, ZHANG Siyuan, et al. Urban fire risk prediction based on spatial-temporal big data and satellite images[J]. Computer Engineering, 2023, 49(6):242-249. doi: 10.19678/j.issn.1000-3428.0064765
[17]	陈硕, 范恒, 周茂磊. 基于循环神经网络的火灾风险预测及应用[J]. 消防科学与技术, 2022, 41(5):655-657.
	CHEN Shuo, FAN Heng, ZHOU Maolei. Fire risk prediction based on recurrent neural networks and its application[J]. Fire Science and Technology, 2022, 41(5):655-657.
[18]	张克松. 基于事故树和层次分析法的城中村火灾风险评估及防控对策研究[C]. 中国消防协会学术工作委员会消防科技论文集, 2022:456-459.
[19]	丁振瀚, 刘迅, 李晓波, 等. GIS+BIM技术在城中村Gustav法火灾风险预警管理中的应用研究[J]. 土木建筑工程信息技术, 2023, 15(3):65-69.
	DING Zhenhan, LIU Xun, LI Xiaobo, et al. An application of GIS+BIM technology in urban village using Gustav Purt fire assessment of fire risk warning management[J]. Journal of Information Technology in Civil Engineering and Architecture, 2023, 15(3):65-69.
[20]	ZHANG Long, SHI Changjiang, ZHANG Fuquan. Predicting forest fire area growth rate using an ensemble algorithm[J]. Forests, 2024, 15(9):1493-1514.
[21]	周建新, 郑日成, 侯宏瑶. 基于维度学习狩猎搜索策略的改进灰狼算法[J]. 仪表技术与传感器, 2023(12):88-93.
	ZHOU Jianxin, ZHENG Richeng, HOU Hongyao. Improved grey wolf algorithm for hunting and searching based on dimension learning[J]. Instrument Technique and Sensor, 2023(12):88-93.
[22]	NEGI G, KUMAR A, PANT S, et al. GWO: a review and applications[J]. International Journal of System Assurance Engineering and Management, 2021, 12:1-8.
[23]	刘玉超, 段喜萍. 融合自适应权重和变螺旋收敛曲线的鲸鱼优化算法[J]. 长江信息通信, 2023, 36(10):48-50.
[24]	王正通, 尤文, 李双. 改进非线性收敛因子灰狼优化算法[J]. 长春工业大学学报: 自然科学版, 2020, 41(2):121-127.
	WANG Zhengtong, YOU Wen, LI Shuang. A grey wolf optimization algorithm with improved nonlinear convergence factor[J]. Journal of Changchun University of Technology: Natural Science Edition, 2020, 41(2):121-127.
[25]	张文胜, 郝孜奇, 朱冀军, 等. 基于改进灰狼算法优化BP神经网络的短时交通流预测模型[J]. 交通运输系统工程与信息, 2020, 20(2):196-203.
	ZHANG Wensheng, HAO Ziqi, ZHU Jijun, et al. BP neural network model for short-time traffic flow forecasting based on transformed grey wolf optimizer algorithm[J]. Journal of Transportation Systems Engineering and Information Technology, 2020, 20(2):196-203.
[26]	王文丰, 余澜婷, 刘哲, 等. 基于二分法和控制信息素量的改进蚁群算法[J]. 计算机工程与设计, 2023, 44(3):784-790.
	WANG Wenfeng, YU Lanting, LIU Zhe, et al. Improved ant colony optimization based on dichotomy and pheromone control[J]. Computer Engineering and Design, 2023, 44(3):784-790.
[27]	李大海, 李鑫, 王振东. 融合多策略的增强麻雀搜索算法及其应用[J]. 计算机应用研究, 2023, 40(10):3032-3039.
	LI Dahai, LI Xin, WANG Zhendong. Enhanced sparrow search algorithm with multiple strategies and its application[J]. Application Research of Computer, 2023, 40(10):3032-3039.
[28]	GUAN Ziyu, REN Changjiang, NIU Jingtai, et al. Great wall construction algorithm: a novel meta-heuristic algorithm for engineer problems[J]. Expert Systems With Applications, 2023, 233:DOI:10.1016/j.eswa.2023.120905.
[29]	叶倩琳, 王万良, 王铮. 多目标粒子群优化算法及其应用研究综述[J]. 浙江大学学报:工学版, 2024, 58(6):1107-1120.
	YE Qianlin, WANG Wanliang, WANG Zheng. Survey of multi-objective particle swarm optimization algorithms and their application[J]. Journal of Zhejiang University:Engineering Science, 2024, 58(6):1107-1120.
[30]	方烜, 杨帆, 梁家豪, 等. 基于改进灰狼算法-BP神经网络的智能巡检机器人电磁兼容故障诊断[J]. 科学技术与工程, 2022, 22(1):243-249.
	FANG Xuan, YANG Fan, LIANG Jiahao, et al. Electromagnetic compatibility fault diagnosis of power inspection robot based on improved grey wolf optimizer-BP neural network[J]. Science Technology and Engineering, 2022, 22(1):243-249.
[31]	朱庆生, 周冬冬, 黄伟. BP神经网络样本数据预处理应用研究[J]. 世界科技研究与发展, 2012, 34(4):624-626.
	ZHU Qingsheng, ZHOU Dongdong, HUANG Wei. Application research of process in BP neural network sample data[J]. World Sci-Tech R&D, 2012, 34(4):624-626.
[32]	应急管理部消防救援局. 中国消防救援年鉴(2020年卷)[M]. 北京: 应急管理出版社,2020:274-275.
[33]	熊昳. 城中村火灾安全隐患的防控策略探究[J]. 消防界:电子版, 2022, 8(12):124-125.
[34]	陈鸣, 叶升升. 从一起较大亡人火灾事故谈城中村火灾处置[J]. 今日消防, 2023, 8(9):49-51.
	CHEN Ming, YE Shengsheng. From a large fatal fire accident to talk about urban village fire disposal[J]. Fire Protection Today, 2023, 8(9):49-51.

函数	表达式	搜索空间
F₁	${f}_{1}\left(x\right)=\stackrel{\mathrm{n}}{\sum _{i=1}}{x}_{i}^{2}$	[-100,100]
F₂	${f}_{2}\left(x\right)=\left.\stackrel{\mathrm{n}}{\sum _{i=1}}\right\|{x}_{i}\|+\left.\stackrel{\mathrm{n}}{\prod _{i=1}}\right\|{x}_{i}\|$	[-10,10]
F₃	${f}_{3}\left(x\right)=\stackrel{\mathrm{n}}{\sum _{i=1}}{\left(\stackrel{\mathrm{i}}{\sum _{j=1}}{x}_{j}\right)}^{2}$	[-100,100]
F₄	${f}_{4}\left(x\right)=\underset{\mathrm{i}}{\mathrm{m}\mathrm{a}\mathrm{x}}\left\{\right\|{x}_{i}\|,1\le i\le n\}$	[-100,100]
F₅	${f}_{5}\left(x\right)=\stackrel{\mathrm{n}}{\sum _{i=1}}i{x}_{i}^{4}+\mathrm{r}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{o}\mathrm{m}\left[\mathrm{0,1}\right)$	[-1.28,1.28]
F₆	${f}_{6}\left(x\right)=\stackrel{\mathrm{n}}{\sum _{i=1}}[{x}_{i}^{2}-10\mathrm{c}\mathrm{o}\mathrm{s}(2\mathrm{\pi }{x}_{i}\left)\right]+10n$	[-5.12,5.12]
F₇	${f}_{7}\left(x\right)=-20\mathrm{e}\mathrm{x}\mathrm{p}\left(-0.2\sqrt{\frac{1}{\mathrm{n}}\stackrel{\mathrm{n}}{\sum _{\mathrm{i}=1}}{\mathrm{x}}_{\mathrm{i}}^{2}}\right)-\mathrm{e}\mathrm{x}\mathrm{p}\left(\frac{1}{\mathrm{n}}\stackrel{\mathrm{n}}{\sum _{\mathrm{i}=1}}\mathrm{c}\mathrm{o}\mathrm{s}2\mathrm{\pi }{\mathrm{x}}_{\mathrm{i}}\right)+20+\mathrm{e}$	[-32,32]
F₈	${f}_{8}\left(x\right)=\frac{1}{4\mathrm{ }000}\stackrel{\mathrm{n}}{\sum _{i=1}}{x}_{i}^{2}-\stackrel{\mathrm{n}}{\prod _{i=1}}\mathrm{c}\mathrm{o}\mathrm{s}\left(\frac{{\mathrm{x}}_{\mathrm{i}}}{\sqrt{\mathrm{i}}}\right)+1$	[-600,600]

函数	GWO		IGWO		PSO		GWCA
函数	均值	标准差	均值	标准差	均值	标准差	均值	标准差
F₁	2.50×10^-27	7.79×10^-27	3.39×10^-205	1.04×10^-107	2.56×10⁰	1.06×10⁰	1.35×10^-1	7.00×10^-1
F₂	1.08×10^-16	7.31×10^-17	4.93×10^-107	4.15×10^-107	4.31×10⁰	1.18×10⁰	4.01×10^-1	7.72×10^-1
F₃	1.15×10^-5	3.08×10^-5	1.76×10^-167	5.98×10^-95	1.82×10²	5.41×10¹	2.77×10³	1.91×10³
F₄	8.77×10^-7	8.92×10^-7	1.12×10^-93	2.21×10^-93	2.02×10⁰	2.39×10^-1	2.83×10¹	4.20
F₅	2.49×10^-3	1.12×10^-3	1.00×10^-4	9.74×10^-5	1.47×10¹	9.67×10⁰	1.59×10⁰	6.55×10^-1
F₆	2.79×10⁰	4.22×10⁰	0	0	1.72×10²	2.76×10¹	5.75×10¹	1.24×10¹
F₇	1.03×10^-13	2.16×10^-14	4.00×10^-15	0	2.73×10⁰	4.67×10^-1	1.10×10¹	1.26×10⁰
F₈	4.67×10^-3	7.67×10^-3	0	0	1.35×10^-1	4.85×10^-2	1.30×10⁰	1.48×10⁰

函数	维数	GWO		IGWO		PSO		GWCA
函数	维数	均值	标准差	均值	标准差	均值	标准差	均值	标准差
F₁	500	1.88×10^-3	5.10×10^-4	1.58×10^-164	0	1.25×10⁴	8.12×10²	1.66×10⁵	2.78×10⁴
F₁	1 000	2.81×10^-1	7.71×10^-2	3.08×10^-161	7.20×10^-161	7.32×10⁴	2.33×10³	5.04×10⁵	7.18×10⁴
F₂	500	1.03×10^-2	1.42×10^-3	5.18×10^-85	1.50×10^-85	2.15×10²	2.40×10¹⁵	8.23×10²	1.17×10²
F₂	1 000	6.23×10^-1	3.58×10^-1	1.00×10^-83	9.81×10^-84	1.69×10³	1.31×10²⁵	1.50×10³	2.61×10²
F₃	500	3.62×10⁵	8.23×10⁴	2.61×10^-141	3.48×10^-139	5.60×10⁵	1.30×10⁵	1.46×10⁶	6.43×10⁵
F₃	1 000	1.16×10⁶	2.26×10⁵	8.63×10^-139	7.73×10^-138	3.07×10⁶	6.81×10⁵	7.55×10⁶	2.64×10⁶
F₄	500	6.73×10¹	5.37×10⁰	2.12×10^-60	2.88×10^-54	3.02×10¹	1.62×10⁰	5.61×10¹	1.70×10⁰
F₄	1 000	7.63×10¹	4.08×10⁰	4.10×10^-33	3.87×10^-10	3.52×10¹	1.55×10⁰	6.04×10¹	1.94×10⁰
F₅	500	2.75×10^-2	1.27×10^-2	3.59×10^-5	9.87×10^-5	5.54×10⁴	1.92×10³	1.75×10³	5.15×10²
F₅	1 000	2.22×10^-1	3.77×10^-2	1.62×10^-4	1.15×10^-4	2.41×10⁵	6.93×10³	1.33×10⁴	2.25×10³
F₆	500	6.99×10¹	2.16×10¹	0	0	8.13×10³	2.01×10²	3.95×10³	1.50×10²
F₆	1 000	1.83×10²	3.86×10¹	0	0	1.67×10⁴	2.35×10²	8.91×10³	2.97×10²
F₇	500	1.79×10^-3	3.60×10^-4	1.02×10^-16	0	1.41×10¹	2.83×10^-1	1.66×10¹	4.10×10^-1
F₇	1 000	1.55×10^-2	2.42×10^-3	4.00×10^-15	0	1.71×10¹	1.91×10^-1	1.70×10¹	3.46×10^-1
F₈	500	1.40×10^-4	2.88×10^-2	0	0	1.92×10¹	2.12×10⁰	1.51×10³	2.15×10²
F₈	1 000	2.19×10^-1	6.74×10^-2	0	0	1.24×10²	7.71×10⁰	5.15×10³	5.42×10²

一级指标	二级指标
环境因素	建筑老旧程度E₁
	建筑材料易燃程度E₂
	电线老化程度E₃
	电器使用情况E₄
	建筑用途复杂程度E₅
	燃气使用率E₆
	易燃物储存情况E₇
	煤气-燃气管道及电气线路铺设状况E₈
	管道老化情况E₉
人口因素	人口密度E₁₀
	外来人口比例E₁₁
	老年人儿童比例E₁₂
	人员安全意识E₁₃
管理因素	消防安全教育及培训E₁₄
管理因素	火灾风险评估及隐患排查E₁₅
其他因素	区域气候特征(温度、湿度)E₁₆
其他因素	雷击、纵火、燃放爆竹等因素E₁₇

编号	火灾起数	面积/km²	概率/(起·km^-2)
1	2	2.84	0.704
2	2	3.07	0.651
3	1	1.41	0.711
4	2	3.23	0.620
5	3	4.06	0.740
6	2	3.04	0.657
7	1	1.38	0.725
8	2	2.77	0.723
9	2	3.27	0.612
10	2	2.86	0.699
11	2	2.90	0.689
12	2	3.20	0.625
13	1	1.49	0.669
14	1	1.26	0.795
15	1	1.65	0.606
16	2	2.54	0.789
17	1	1.35	0.739
18	1	1.54	0.648
19	1	1.30	0.767
20	2	2.68	0.748
21	2	2.94	0.680
22	1	1.60	0.627
23	2	2.50	0.800
24	3	4.06	0.738
25	2	2.57	0.778
26	1	1.62	0.619
27	1	1.53	0.653
︙	︙	︙	︙
360	2	3.35	0.597