Probabilistic Analysis of Tunnel Roof Deflection under Sequential Excavation Using ANN-Based Monte Carlo Simulation and Simplified Reliability Approach
Tunnel roof deflection is an important measure to control the safety of excavation activity, especially for large-section tunnels using the sequential construction method. A three-dimensional (3D) model of a sequential excavation project in the Lijiaping metro tunnel, Chongqing, China, was developed and the maximum tunnel roof deflection was calculated deterministically at the end of each excavation stage. By defining the characteristics of the soft rock layers surrounding the tunnel section as the random variables, a probabilistic analysis was conducted and the problem was formulated as a reliability model. Two different approaches were used to solve the established reliability problem. One entails an artificial neural network (ANN) trained by a large data set obtained from numerical simulations and then accompanied by the Monte Carlo (MC) sampling method to calculate the probability. The other is a simplified reliability approach using a small data set to approximate the exceedance probability via a regression-based algorithm. The proposed ANN-based metamodel used in the first method could accurately predict tunnel roof deflection and replace the software simulator. Subsequently, the probabilistic results obtained from this method following MC sampling could well converge and provide the probability with enough accuracy. Interestingly, the simplified approach with more than 40 random samples can also provide acceptable results, which provides an economical approach to estimate the exceedance probability.
- Record URL:
-
Availability:
- Find a library where document is available. Order URL: http://worldcat.org/issn/23767642
-
Supplemental Notes:
- © 2021 American Society of Civil Engineers.
-
Authors:
- Zhou, Mingliang
-
0000-0002-2640-042X
- Shadabfar, Mahdi
-
0000-0002-3885-5050
- Xue, Yadong
-
0000-0002-0363-9702
- Zhang, Yi
- Huang, Hongwei
- Publication Date: 2021-12
Language
- English
Media Info
- Media Type: Web
- Features: References;
- Pagination: 04021043
-
Serial:
- ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering
- Volume: 7
- Issue Number: 4
- Publisher: American Society of Civil Engineers
- ISSN: 2376-7642
- Serial URL: https://ascelibrary.org/journal/ajrua6
Subject/Index Terms
- TRT Terms: Deflection; Excavation; Monte Carlo method; Railroad tunnels; Rapid transit; Reliability
- Geographic Terms: Chongqing (China)
- Subject Areas: Bridges and other structures; Construction; Public Transportation; Railroads;
Filing Info
- Accession Number: 01783956
- Record Type: Publication
- Files: TRIS, ASCE
- Created Date: Sep 30 2021 9:33AM