基于人工智能变形预测隧道坍塌失效概率评估方法

吴波; 丘伟兴; 徐世祥; 蔡俊华; 李贻材; 张耀

doi:10.3799/dqkx.2022.147

基于人工智能变形预测隧道坍塌失效概率评估方法

doi: 10.3799/dqkx.2022.147

吴波^{1, 2, 3,},
丘伟兴^{1, 6},
徐世祥^1, ,,
蔡俊华⁴,
李贻材⁵,
张耀⁵

1.
广西大学土木建筑工程学院, 广西南宁 530004
2.
东华理工大学土木与建筑工程学院, 江西南昌 330013
3.
广州城建职业学院建筑工程学院, 广东广州 510925
4.
三明莆炎高速公路有限责任公司, 福建三明 365000
5.
中铁十一局集团第一工程有限公司, 湖北襄阳 441104
6.
中南大学土木工程学院, 湖南长沙 410075

基金项目:

国家自然科学基金项目 52168055

国家自然科学基金项目 51678164

江西省自然科学基金项目 20212ACB204001

广西自然科学基金项目 2018GXNSFDA138009

详细信息

作者简介:
吴波（1971-），男，教授，主要从事隧道及地下工程安全风险智慧防控与管理研究.E-mail：wubo@gxu.edu.cn

通讯作者:
徐世祥，E-mail：603559081@qq.com

中图分类号: P64
计量
- 文章访问数: 340
- HTML全文浏览量: 139
- PDF下载量: 162
- 被引次数: 0
出版历程
- 收稿日期: 2022-04-18
- 刊出日期: 2024-11-25

A Method for Assessing Probability of Tunnel Collapse Based on Artificial Intelligence Deformation Prediction

Wu Bo^{1, 2, 3
,},
Qiu Weixing^{1, 6},
Xu Shixiang^{1
, ,},
Cai Junhua⁴,
Li Yicai⁵,
Zhang Yao⁵

1.
College of Civil Engineering and Architecture, Guangxi University, Nanning 530004, China
2.
School of Civil and Architectural Engineering, East China University of Technology, Nanchang 330013, China
3.
School of Architectural Engineering, Guangzhou City Construction College, Guangzhou 510925, China
4.
Sanming Traffic Construction Group Co., Ltd., Sanming 365000, China
5.
The 1st Engineering Ltd. of China Railway 11th Bureau Group, Xiangyang 441104, China
6.
School of Civil Engineering, Central South University, Changsha 410075, China

摘要

摘要: 当隧道坍塌事故发生时，决策者往往没有足够的反应时间去采取相应的加固措施.超前预测隧道坍塌失效概率已成为隧道工程建设的关键问题.为了超前评估隧道坍塌失效概率并及时预警，本研究提出了一种多数据融合方法.该方法将云模型（CM）、多输出高斯过程回归（MOGPR）和改进的D-S理论相结合.通过融合多项监测数据（拱顶位移、水平收敛位移），减少数据的不确定性，提高评估结果的准确性和鲁棒性.此外，利用人工智能预测的围岩变形作为信息源，得到超前的坍塌失效概率评估.并将该方法运用于金珠帕隧道，为决策者提供更多的响应时间.最终，围岩支护只产生少量的变形裂缝，避免了隧道坍塌的发生.
- 隧道坍塌 /
- 失效概率评估 /
- 云模型 /
- D-S理论 /
- 多输出高斯过程回归 /
- 安全工程 /
- 工程地质
Abstract: When a tunnel collapse occurs, decision makers often do not have enough reaction time to take appropriate reinforcement measures. Advance prediction of tunnel collapse failure probability has become a key issue in tunnel engineering construction. As for assessing the tunneling collapse failure probability and providing basic risk-controlling strategies, in this study it proposes a novel multi-source information fusion approach that combines the cloud model (CM), the multi-output gaussian process regression (MOGPR), and the improved D-S evidence theory. The fusion of multiple monitoring data (vault displacement, horizontal convergence displacement) reduces data uncertainty and improves the accuracy and robustness of assessment results. In addition, the surrounding rock deformation predicted by artificial intelligence is used as a source of information to obtain an advanced collapse failure probability assessment. As a result, decision makers have a longer response time before the collapse occurs. Applying the method to the Jinzhupa tunnel provides decision makers with more response time. In the end, only a small amount of deformation cracks were generated in the surrounding rock support, avoiding the tunnel collapse.
- tunnel collapse /
- failure probability assessment /
- cloud model /
- D-S evidence theory /
- multi-output gaussian process regression /
- safety engineering /
- engineering geology

HTML全文

图 1 多数据融合评估方法流程

Fig. 1. Flow chart of the proposed hybrid method for multi-source data fusion decision

下载: 全尺寸图片幻灯片

图 2 GA和MOGPR耦合算法流程

Fig. 2. Flow chart of GA and MOGPR coupling algorithm

下载: 全尺寸图片幻灯片

图 3 滚动预测方法原理

Fig. 3. Schematic diagram of the rolling prediction method

下载: 全尺寸图片幻灯片

图 4 金珠帕隧道右线纵截面

Fig. 4. Longitudinal section of the right line of Jinzhupa Tunnel

下载: 全尺寸图片幻灯片

图 5 监测点布置示意

Fig. 5. Monitoring point layout

下载: 全尺寸图片幻灯片

图 6 隧道坍塌

Fig. 6. Tunneling collapse

下载: 全尺寸图片幻灯片

图 7 隧道开挖周期的失效等级评估

Fig. 7. Failure level assessment of tunnel excavation cycle

下载: 全尺寸图片幻灯片

表 1 监控量测数据的分类

Table 1. Classification of monitoring measurement data

隧道坍塌等级	Ⅰ（安全）	Ⅱ（变形）	Ⅲ（小规模坍塌）	Ⅳ（大规模坍塌）
每日变形速率（mm/d）	0≤x < 2	2≤x < 5	5≤x < 10	10≤x≤20
累计变形（mm）	0≤y < 50	50≤y < 100	100≤y < 200	200≤y≤300
注：累积变形量（y）应根据测点与掌子面距离（D）乘以系数（ζ），ζ根据规范《公路隧道监测测量技术规范》（DB 35/T 1067-2010）确定，如表 2所示.

下载: 导出CSV

表 2 累积变形量（y）的系数（ζ）

Table 2. Coefficient (ζ) of cumulative deformation (y)

测点到掌子面的距离（D）	1B	2B	3B	4B~6B
ζ	0.5	0.75	0.85	1
注：B为开挖隧道的跨度.

下载: 导出CSV

表 3 K242+880监测段围岩位移预测结果

Table 3. Prediction results of surrounding rock displacement in K242+880 monitoring section

监测时间（d）	实测拱顶位移（mm）	MOGPR预测值（mm）	SVR预测值（mm）	MOGPR相对误差（%）	SVR相对误差（%）	实测水平收敛值（mm）	MOGPR预测值（mm）	SVR预测值（mm）	MOGPR相对误差（%）	SVR相对误差
10	12.3	12.00	11.40	2.45	7.29	10.7	10.36	10.34	3.14	3.38
11	12.9	12.38	12.00	4.04	6.98	11.3	11.00	10.31	2.65	8.73
12	13.8	12.77	13.00	7.45	5.80	12.2	12.00	9.89	1.64	18.92
13	14.6	14.65	14.43	0.33	1.18	12.9	12.74	12.72	1.25	1.38
14	15.3	15.63	15.10	2.15	1.29	13.7	13.33	13.30	2.70	2.94
15	15.8	16.66	15.68	5.42	0.79	14.4	13.88	13.83	3.60	3.98
16	16.2	16.58	15.92	2.36	1.71	15.1	15.09	14.63	0.05	3.08
17	16.8	17.27	15.81	2.78	5.88	15.9	15.80	14.63	0.62	7.96
18	17.5	17.94	15.42	2.51	11.86	16.5	16.50	14.32	0.02	13.20
19	18.1	17.76	17.76	1.90	1.87	17.0	17.13	16.78	0.76	1.29
20	18.7	18.06	17.73	3.41	5.21	17.5	17.68	16.91	1.05	3.37
21	19.2	18.29	17.30	4.76	9.89	17.9	18.17	16.80	1.53	6.15
22	19.7	19.41	19.30	1.46	2.02	18.1	18.23	18.07	0.73	0.16
23	20.3	19.38	19.19	4.51	5.47	18.4	18.47	18.21	0.38	1.05
24	20.9	19.13	18.82	8.45	9.94	18.7	18.62	18.24	0.40	2.47
25	21.4	21.36	21.23	0.21	0.82	19.1	18.98	18.78	0.62	1.68
26	21.9	21.82	21.54	0.38	1.65	19.3	19.32	18.81	0.08	2.53
27	22.6	22.24	21.77	1.59	3.69	19.5	19.68	18.72	0.91	3.98
28	23.2	23.18	23.00	0.09	0.88	19.7	19.49	19.55	1.07	0.75
29	23.9	23.81	23.36	0.39	2.27	20.1	19.31	19.41	3.93	3.41
30	24.5	24.44	23.56	0.24	3.82	20.4	19.00	19.15	6.85	6.14
31	25.1	25.15	24.84	0.21	1.06	20.6	20.60	20.59	0.00	0.07
32	25.3	25.81	25.03	2.02	1.07	20.8	20.85	20.81	0.22	0.07
33	25.5	26.48	24.99	3.84	1.99	20.9	21.08	21.03	0.88	0.61
34	25.7	25.40	25.42	1.18	1.08	21.1	21.10	20.87	0.01	1.10
35	25.8	25.11	25.11	2.69	2.66	21.4	21.22	20.75	0.85	3.04
36	25.9	24.66	24.58	4.78	5.10	21.6	21.31	20.52	1.36	4.98
37	26.0	26.00	25.94	0.02	0.24	21.7	21.71	21.62	0.04	0.37
38	26.2	26.10	25.98	0.39	0.82	21.9	21.85	21.50	0.25	1.83
39	26.3	26.18	25.97	0.47	1.25	22.1	21.96	21.26	0.61	3.81
40	26.5	26.38	26.21	0.45	1.11	22.4	22.25	22.34	0.65	0.25
41	26.6	26.38	25.98	0.82	2.34	22.6	22.41	22.62	0.83	0.10
42	26.8	26.29	25.70	1.89	4.10	22.99	22.56	22.80	1.86	0.84
位移预测的平均相对误差（%）				2.29	3.43				1.26	2.29

下载: 导出CSV

表 4 两个监控指标的云模型参数值

Table 4. Cloud model parameter values of two monitoring indicators

指标	Ⅰ			Ⅱ			Ⅲ			Ⅳ
指标	Ex	En	He	Ex	En	He	Ex	En	He	Ex	En	He
每日速率	1	0.333	0.002	3.5	0.5	0.002	7.5	0.833	0.002	12.5	0.833	0.002
累计沉降	25	8.333	0.002	75	8.333	0.002	150	16.777	0.002	250	16.777	0.002

下载: 导出CSV

表 5 五个测试样本的融合结果

Table 5. Fusion results of five test samples

隧道断面	评估模型	等级Ⅰ的失效概率	等级Ⅱ的失效概率		等级Ⅲ的失效概率	等级Ⅳ的失效概率	评估值	实际值
No.1	拱顶位移	0.80	0.20	0.00	0.00		Ⅰ	Ⅰ
	水平收敛	0.45	0.55	0.00	0.00		Ⅱ	Ⅰ
Improve D-S		0.77	0.23	0.00	0.00		Ⅰ	Ⅰ
No.2	拱顶位移	0.50	0.50	0.00	0.02		--	Ⅰ
	水平收敛	0.60	0.40	0.00	0.00		Ⅰ	Ⅰ
Improve D-S		0.70	0.30	0.00	0.00		Ⅰ	Ⅰ
No.3	拱顶位移	0.00	0.80	0.20	0.00		Ⅱ	Ⅱ
	水平收敛	0.19	0.81	0.00	0.00		Ⅱ	Ⅱ
Improve D-S		0.00	1.00	0.00	0.00		Ⅱ	Ⅱ
No.4	拱顶位移	0.00	0.30	0.70	0.00		Ⅲ	Ⅲ
	水平收敛	0.00	0.56	0.44	0.00		Ⅱ	Ⅲ
Improve D-S		0.00	0.35	0.65	0.00		Ⅲ	Ⅲ
No.5	拱顶位移	0.00	0.15	0.85	0.00		Ⅲ	Ⅱ
	水平收敛	0.86	0.14	0.00	0.00		Ⅰ	Ⅰ
Improve D-S		0.43	0.15	0.42	0.00		Ⅰ	Ⅰ

下载: 导出CSV

表 6 隧道围岩位移预测

Table 6. The prediction of tunnel surrounding rock displacement

	训练数据(mm)									预测数据(mm)
天数	1	2	3	4	5	6	7	8	9	10	11	12
拱顶位移	3.1	6.6	8.2	9.4	10.2	11.2	13.2	16.8	21.1	26.5	36.5	47.3
水平收敛	2.7	5.1	6.5	7.2	8.1	9.3	11.4	14.4	19.5	25.1	30.6	35.1
实际拱顶位移	3.1	6.6	8.2	9.4	10.2	11.2	13.2	16.8	21.1	25.8	34.4	38.1
实际水平位移	2.7	5.1	6.5	7.2	8.1	9.3	11.4	14.4	19.5	24.6	31.1	33.4

下载: 导出CSV

表 7 预测坍塌失效概率值

Table 7. The prediction of the collapse failure probability

	监测数据(mm)									预测数据(mm)
天数	1	2	3	4	5	6	7	8	9	10	11	12
预测坍塌等级	Ⅰ	Ⅰ	Ⅰ	Ⅰ	Ⅰ	Ⅰ	Ⅰ	Ⅱ	Ⅱ	Ⅲ	Ⅲ	Ⅲ
实际坍塌等级	Ⅰ	Ⅰ	Ⅰ	Ⅰ	Ⅰ	Ⅰ	Ⅰ	Ⅱ	Ⅱ	Ⅲ	Ⅲ	Ⅱ

下载: 导出CSV

参考文献(24)

Adoko, A. C., Jiao, Y. Y., Wu, L., et al., 2013. Predicting Tunnel Convergence Using Multivariate Adaptive Regression Spline and Artificial Neural Network. Tunnelling and Underground Space Technology, 38: 368-376. https://doi.org/10.1016/j.tust.2013.07.023
Cai, B., Liu, Y., Fan, Q., et al., 2014. Multi-Source Information Fusion Based Fault Diagnosis of Ground-Source Heat Pump Using Bayesian Network. Applied Energy, 114: 1-9. https://doi.org/10.1016/j.apenergy.2013.09.043
Chen, F., Zhang, W., 2021. Influence of Spatial Variability on the Uniaxial Compressive Responses of Rock Pillar Based on 3D Random Field. ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering, 7(3): 04021035. https://doi.org/10.1061/AJRUA6.0001162
Chen, F. Y., Zhang, W. G., 2022. Reliability Analysis of Lijiaping Metro Tunnel Based on Conditional Random Field, Chongqing. Journal of Basic Science and Engineering, 30(1): 166-182 (in Chinese with English abstract).
Chen, X., Li, L., Wang, L., et al., 2019. The Current Situation and Prevention and Control Countermeasures for Typical Dynamic Disasters in Kilometer-Deep Mines in China. Safety Science, 115: 229-236. doi: 10.1016/j.ssci.2019.02.010
Duong, P. L. T., Park, H., Raghavan, N., 2018. Application of Multi-Output Gaussian Process Regression for Remaining Useful Life Prediction of Light Emitting Diodes. Microelectronics Reliability, 88-90: 80-84. https://doi.org/10.1016/j.microrel.2018.07.106
Gravina, R., Alinia, P., Ghasemzadeh, H., et al., 2017. Multi-Sensor Fusion in Body Sensor Networks: State-of-the-Art and Research Challenges. Information Fusion, 35: 68-80. https://doi.org/10.1016/j.inffus.2016.09.005
Guo, K., Zhang, L., 2021. Multi-Source Information Fusion for Safety Risk Assessment in Underground Tunnels. Knowledge-Based Systems, 227: 107210. https://doi.org/10.1016/j.knosys.2021.107210
Huang, X., Liu, Q., Liu, H., et al., 2018. Development and In-Situ Application of a Real-Time Monitoring System for the Interaction between TBM and Surrounding Rock. Tunnelling and Underground Space Technology, 81: 187-208. https://doi.org/10.1016/j.tust.2018.07.018
Li, C., Xu, J., Pan, J., et al., 2012. Plastic Zone Distribution Laws and Its Types of Surrounding Rock in Large-Span Roadway. International Journal of Mining Science and Technology, 22: 23-28. https://doi.org/10.1016/j.ijmst.2011.06.002
Li, D., Liu, C., Gan, W., 2009. A New Cognitive Model: Cloud Model. International Journal of Intelligent Systems, 24: 357-375. https://doi.org/10.1002/int.20340
Li, T., De la Prieta Pintado, F., Corchado, et al., 2017. Multi-Source Homogeneous Data Clustering for Multi-Target Detection from Cluttered Background with Misdetection. Applied Soft Computing, 60: 436-446. https://doi.org/10.1016/j.asoc.2017.07.012
Lim, S. L. H., Duong, P. L. T., Park, H., et al., 2020. Assessing Multi-Output Gaussian Process Regression for Modeling of Non-Monotonic Degradation Trends of Light Emitting Diodes in Storage. Microelectronics Reliability, 114: 113794. https://doi.org/10.1016/j.microrel.2020.113794
Liu, K., Liu, B., 2019. Intelligent Information-Based Construction in Tunnel Engineering Based on the GA and CCGPR Coupled Algorithm. Tunnelling and Underground Space Technology, 88: 113-128. https://doi.org/10.1016/j.tust.2019.02.012
Saadi, I., Farooq, B., Mustafa, A., et al., 2018. An Efficient Hierarchical Model for Multi-Source Information Fusion. Expert Systems with Applications, 110: 352-362. doi: 10.1016/j.eswa.2018.06.018
Xue, X., Xiao, M., 2017. Deformation Evaluation on Surrounding Rocks of Underground Caverns Based on PSO-LSSVM. Tunnelling and Underground Space Technology, 69: 171-181. https://doi.org/10.1016/j.tust.2017.06.019
Yan, X. H., Guo, C. B., Liu, Z. B., et al., 2022. Physical Simulation Experiment of Granite Rockburst in a Deep-Buried Tunnel in Kangding County, Sichuan Province, China. Earth Science, 47(6): 2081-2093(in Chinese with English abstract).
Yang, Y., Jing, Z., Gao, T., et al., 2007. Multi-Sources Information Fusion Algorithm in Airborne Detection Systems. Journal of Systems Engineering and Electronics, 18: 171-176. https://doi.org/10.1016/S1004-4132(07)60070-X
Zhang, G. H., Chen, W., Jiao, Y. Y., et al., 2020. A Failure Probability Evaluation Method for Collapse of Drill-and-Blast Tunnels Based on Multistate Fuzzy Bayesian Network. Engineering Geology, 276: 105752. https://doi.org/10.1016/j.enggeo.2020.105752
Zhang, L., Wu, X., Ding, L., et al., 2013a. A Novel Model for Risk Assessment of Adjacent Buildings in Tunneling Environments. Building and Environment, 65: 185-194. https://doi.org/10.1016/j.buildenv.2013.04.008
Zhang, Y., Zhang, H., Nasrabadi, N. M., et al., 2013b. Multi-Metric Learning for Multi-Sensor Fusion Based Classification. Information Fusion, 14: 431-440. https://doi.org/10.1016/j.inffus.2012.05.002
Zhang, L., Wu, X., Zhu, H., et al., 2017. Perceiving Safety Risk of Buildings Adjacent to Tunneling Excavation: An Information Fusion Approach. Automation in Construction, 73: 88-101. https://doi.org/10.1016/j.autcon.2016.09.003
陈福勇, 仉文岗, 2022. 基于条件随机场的重庆李家坪地铁隧道可靠度分析. 应用基础与工程科学学报, 30(1): 166-182.
严孝海, 郭长宝, 刘造保, 等, 2022. 四川康定某深埋隧道花岗岩岩爆物理模拟实验研究. 地球科学, 47(6): 2081-2093. doi: 10.3799/dqkx.2021.153