Decision-support model for safety evaluation of existing civil buildings and its application

doi:10.16265/j.cnki.issn1003-3033.2025.09.1527

Abstract

Abstract:

To improve the efficiency of safety inspections for existing civil buildings, this study develops a decision-support model for safety evaluation of existing civil buildings based on ML techniques. Building safety feature data were first collected from building evaluation reports. Then, a multi-dimensional indicator system integrating "design-evolution-status" features was established through correlation analysis and recursive feature elimination with cross-validation. Subsequently, five ML models were built and evaluated using performance metrics such as accuracy, precision, and recall. Furthermore, a decision-support platform for safety evaluation of existing civil buildings was developed and validated through a real engineering project to examine its practical operability. The results demonstrate that, compared to design features, evolution and status features more effectively reflect the actual safety conditions of civil buildings. In particular, building age, renovation or extension history, and concrete beam load capacity are identified as key features. Among the tested models, the Decision Tree algorithm shows the best performance in evaluating the safety of enclosure system, superstructure, foundation, and individual evaluation unit.

Key words: machine learning (ML), existing civil buildings, safety evaluation, decision-support model, indicator system

CLC Number:

X928.9其他

DAI Mengfan, LI Lingzhi, QIAN Yuxin, YUAN Jingfeng, HAN Xiaojian, ZHAO Changhao. Decision-support model for safety evaluation of existing civil buildings and its application[J]. China Safety Science Journal, 2025, 35(9): 193-201.

Figures/Tables 13

Fig.1

Table 1

Table 2

Fig.2

Table 3

Table 4

Table 5

Fig.3

Table 6

Table 7

Fig.4

Table 8

Confusion matrices for model test results

预测目标	KNN	NB	DT	RF	SVM
围护系统安全性(E₁)	$11111221011011201$	$10031112111101111$	$13100140000120013$	$14000320020100013$	$12200221011010013$
上部承重结构安全性(E₂)	$5450111210021300$	$5720301120100121$	$14000050011100112$	$13010041002010004$	$13100131011100013$
地基基础安全性(E₃)	$8640121002102000$	$52110112001202000$	$17100220000300101$	$18000121010200011$	$17100112010201001$
鉴定单元安全性(G₁)	$9302040121000202$	$31001230011010013$	$14000220000300004$	$14000220102010103$	$13100112101110031$

Table 8

Fig.5

References 17

[1]	中华人民共和国住房和城乡建设部. 住房城乡建设部关于全面开展城市体检工作的指导意见[EB/OL]. (2023-11-29). https://www.gov.cn/lianbo/bumen/202312/content_6918797.htm.
[2]	张鑫, 李安起, 赵考重. 建筑结构鉴定与加固改造技术的进展[J]. 工程力学, 2011, 28(1):1-11,25.
	ZHANG Xin, LI Anqi, ZHAO Kaozhong. Advances in assessment and retrofitting of building structures[J]. Engineering Mechanics, 2011, 28(1):1-11,25.
[3]	李俊杰, 葛德, 孙文. 基于结构健康监测系统的某办公建筑运营期结构响应分析[J]. 建筑结构, 2023, 53(21):107-111.
	LI Junjie, GE De, SUN Wen. Structural response analysis of an office building during operation based on structural health monitoring system[J]. Building Structure, 2023, 53(21):107-111.
[4]	蒋维乐, 姚坤, 林启敬, 等. 结构健康监测系统在木结构古建筑中的应用[J]. 传感技术学报, 2024, 37(8):1295-1306.
	JIANG Weile, YAO Kun, LIN Qijing, et al. Application of structural health monitoring system in wooden-structure ancient buildings[J]. Chinese Journal of Sensors and Actuators, 2024, 37(8):1295-1306.
[5]	黄永, 鲍跃全, 李惠. 结构状态识别与评估的机器学习方法研究进展[J]. 力学进展, 2023, 53(4):774-792.
	HUANG Yong, BAO Yuequan, LI Hui. Research advances in machine learning for structural state identification and condition assessment[J]. Advances in Mechanics, 2023, 53(4):774-792.
[6]	周艳兵, 施钟淇, 金典琦, 等. 天空地一体化监测技术在既有建筑安全管控中的应用和展望[J]. 建筑结构, 2022, 52(增2):1893-1897.
	ZHOU Yanbing, SHI Zhongqi, JIN Dianqi, et al. Application and prospect of integrated spaceborne, airborne and terrestrial monitoring technology in safety management and control of existing buildings[J]. Building Structure, 2022, 52(S2):1893-1897.
[7]	VENKATESH P, ALAPATI M. Condition assessment of existing concrete building using non-destructive testing methods for effective repair and restoration: a case study[J]. Civil Engineering Journal, 2017, 3(10):841-855.
[8]	ALAVI H, BORTOLINI R, FORCADA N. BIM-based decision support for building condition assessment[J]. Automation in Construction, 2022,135:DOI:10.1016/j.autcon.2021.104117.
[9]	DING Zhikun, NIU Jindi, LIU Shan, et al. An approach integrating geographic information system and building information modelling to assess the building health of commercial buildings[J]. Journal of Cleaner Production, 2020,257:DOI:10.1016/j.jclepro.2020.120532.
[10]	GB50292—2015, 民用建筑可靠性鉴定标准[S].
	GB50292-2015, Standard for appraisal of reliability of civil buildings[S].
[11]	汪明元, 陈松庭, 陈彪, 等. 考虑数据预处理和特征选择的静力触探参数研究[J]. 浙江工业大学学报, 2025, 53(1):15-21.
	WANG Mingyuan, CHEN Songting, CHEN Biao, et al. Study on parameters of cone penetration test considering data pre-processing and feature selection[J]. Journal of Zhejiang University of Technology, 2025, 53(1):15-21.
[12]	段在鹏, 李帆, 郭进, 等. 城市易涝区房屋结构安全集成预警模型[J]. 中国安全科学学报, 2023, 33(7):173-180. doi: 10.16265/j.cnki.issn1003-3033.2023.07.0759
	DUAN Zaipeng, LI Fan, GUO Jin, et al. Integrated warning model for structural safety of buildings in urban waterlogged area[J]. China Safety Science Journal, 2023, 33(7):173-180. doi: 10.16265/j.cnki.issn1003-3033.2023.07.0759
[13]	刘立渠, 彭立新, 赵振红, 等. 既有建筑结构安全性鉴定中分项系数的调整研究[J]. 建筑结构学报, 2024, 45(4):247-253.
	LIU Liqu, PENG Lixin, ZHAO Zhenhong, et al. Research on adjusting partial factors in safety appraisal of existing building structure[J]. Journal of Building Structures, 2024, 45(4):247-253.
[14]	SEDGWICK P. Spearman's rank correlation coefficient[J]. British Medical Journal, 2014,349:DOI:10.1136/bmj.g7327.
[15]	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
[16]	HAIRANI H, WIDIYANINGTYAS T, PRASETYA D D. Addressing class imbalance of health data: a systematic literature review on modified synthetic minority oversampling technique (SMOTE) strategies[J]. JOIV: International Journal on Informatics Visualization, 2024, 8(3):1310-1318.
[17]	LI Huixia, NYIRANDAYISABYE R, DONG Qiming, et al. Crack damage prediction of asphalt pavement based on tire noise: a comparison of machine learning algorithms[J]. Construction and Building Materials, 2024,414:DOI: 10.1016/j.conbuildmat.2024.134867.

等级	分级标准(是否影响整体承载)	处理要求
A_su	不影响	极少数一般构件应采取措施
B_su	尚不显著影响	极少数构件应采取措施
C_su	显著影响	极少数构件须及时采取措施
D_su	严重影响	必须立即采取措施

变量类型	变量名称	取值说明	示例	数据类型
输入变量	房屋层数X₁/层	1~12	2	连续型
	建筑面积X₂/m²	1~22 000	1 817
	房屋长度X₃/m	1~190	41.6
	房屋宽度X₄/m	1~60	24
	底层层高X₅/m	1~6.7	3.4
	标准层层高X₆/m	1~6.5	3.45
	房屋总高X₇/m	1~50	10.3
	建筑年龄X₈/a	1~45	17
	最大倾斜率X₉/‰	0~9	0.96
	建筑使用用途X₁₀	商业建筑=0;教学建筑=1;住宅=2; 敬老院=3;医院=4;其他=5	1	无序离散型
	楼屋盖形式X₁₁	平屋面=0;坡屋面=1	1
	平面形状X₁₂	矩形=0;多边形=1;L形=2;环形=3;其他=4	1
	有无设计修改X₁₃	否=0;是=1	1
	有无改扩建X₁₄		1
	围护系统渗漏/开裂/脱落X₁₅—X₁₇		0/1/0
	混凝土柱/梁/板承载能力X₁₈—X₂₀	a_u=1;b_u=2;c_u=3;d_u=4	1/1/2	有序离散型
	混凝土柱/梁/板构造合理性X₂₁—X₂₃		1/1/1
	混凝土柱/梁/板位移或变形X₂₄—X₂₆		1/1/1
	混凝土柱/梁/板开裂损伤X₂₇—X₂₉		1/1/1
输出变量	围护系统承重安全性E₁	A_u=1;B_u=2;C_u=3;D_u=4	2
	上部承重结构安全性E₂		2
	地基基础安全性E₃		2
	鉴定单元安全性G₁	A_su=1;B_su=2;C_su=3;D_su=4	2

指标	E₁	E₂	E₃	G₁
X₁₀	0.435	0.405	0.351	0.412
X₁₁	不显著	不显著	不显著	0.301
X₁₂	不显著	不显著	不显著	不显著
X₁₃	0.476	0.461	0.412	0.483
X₁₄	0.612	0.579	0.618	0.599
X₁₅	0.520	0.480	0.372	0.466
X₁₆	0.461	0.486	0.417	0.479
X₁₇	0.492	0.536	0.438	0.528

预测目标	设计特征			演变特征							状态特征
预测目标	X₁	X₂	X₁₀	X₈	X₉	X₁₃	X₁₄	X₁₅	X₁₆	X₁₇	X₁₈	X₁₉	X₂₀	X₂₁	X₂₂	X₂₃	X₂₄	X₂₅	X₂₆	X₂₇	X₂₈	X₂₉
E₁	√	√	√	√	√	√	—	√	√	√	√	√	√	√	√	√	—	√	—	√	√	√
E₂	√	√	—	√	√	√	√	—	√	√	√	√	√	√	√	√	√	—	√	√	√	√
E₃	—	√	—	√	√	√	√	—	√	√	√	√	√	√	√	√	√	√	√	√	√	√
G₁	√	√	—	√	√	√	√	—	√	√	√	√	√	√	√	√	—	—	√	√	√	√

预测目标	样本数量(总样本/初始训练集样本)
预测目标	A_(s)u	B_(s)u	C_(s)u	D_(s)u	合计
E₁	47/33	18/13	10/7	12/8	87/61
E₂	46/32	17/12	11/8	13/9	87/61
E₃	60/42	12/8	9/6	6/4	87/60
G₁	46/32	16/11	11/8	14/10	87/61