Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Bayesian-Network-Based Predictions of Water Inrush Incidents in Soft Rock Tunnels
https://orcid.org/https://orcid.org/0000-0002-2737-6046
This study proposes a Bayesian network-based model for predicting the probability of water inrush incidents in soft rock tunnels. The risk decomposition structure method was used to statistically analyze 70 groups of water inrush incidents in typical soft rock tunnels; the nine primary factors affecting these incidents were identified across three categories: hydrological characteristics, stratigraphic characteristics, and construction factors. Correlation coefficients and expert experience methods were used to analyze the cause–effect relationship between the factors and establish the Bayesian network structure for predicting these water inrush incidents. The non-water inrush cases were identified using the hierarchical analysis method and the generative adversarial network, thereby effectively addressing the imbalance of sample classification in the database. The maximum expectation algorithm was used to obtain 140 sets of data (including 70 sets generated) from the Bayesian network. The overall accuracy of the model reached 87.85%. The model was applied to the No. 1 slant shaft of the Lanzhou–Chongqing railway tunnel, and the prediction results were consistent with the observations in the actual project. The model can effectively predict the probability of a water inrush incident during the construction of soft rock tunnels.
Bayesian-Network-Based Predictions of Water Inrush Incidents in Soft Rock Tunnels
https://orcid.org/https://orcid.org/0000-0002-2737-6046
This study proposes a Bayesian network-based model for predicting the probability of water inrush incidents in soft rock tunnels. The risk decomposition structure method was used to statistically analyze 70 groups of water inrush incidents in typical soft rock tunnels; the nine primary factors affecting these incidents were identified across three categories: hydrological characteristics, stratigraphic characteristics, and construction factors. Correlation coefficients and expert experience methods were used to analyze the cause–effect relationship between the factors and establish the Bayesian network structure for predicting these water inrush incidents. The non-water inrush cases were identified using the hierarchical analysis method and the generative adversarial network, thereby effectively addressing the imbalance of sample classification in the database. The maximum expectation algorithm was used to obtain 140 sets of data (including 70 sets generated) from the Bayesian network. The overall accuracy of the model reached 87.85%. The model was applied to the No. 1 slant shaft of the Lanzhou–Chongqing railway tunnel, and the prediction results were consistent with the observations in the actual project. The model can effectively predict the probability of a water inrush incident during the construction of soft rock tunnels.
Bayesian-Network-Based Predictions of Water Inrush Incidents in Soft Rock Tunnels
https://orcid.org/https://orcid.org/0000-0002-2737-6046
This study proposes a Bayesian network-based model for predicting the probability of water inrush incidents in soft rock tunnels. The risk decomposition structure method was used to statistically analyze 70 groups of water inrush incidents in typical soft rock tunnels; the nine primary factors affecting these incidents were identified across three categories: hydrological characteristics, stratigraphic characteristics, and construction factors. Correlation coefficients and expert experience methods were used to analyze the cause–effect relationship between the factors and establish the Bayesian network structure for predicting these water inrush incidents. The non-water inrush cases were identified using the hierarchical analysis method and the generative adversarial network, thereby effectively addressing the imbalance of sample classification in the database. The maximum expectation algorithm was used to obtain 140 sets of data (including 70 sets generated) from the Bayesian network. The overall accuracy of the model reached 87.85%. The model was applied to the No. 1 slant shaft of the Lanzhou–Chongqing railway tunnel, and the prediction results were consistent with the observations in the actual project. The model can effectively predict the probability of a water inrush incident during the construction of soft rock tunnels.
Bayesian-Network-Based Predictions of Water Inrush Incidents in Soft Rock Tunnels
KSCE J Civ Eng
Feng, Xianda (Autor:in) / Lu, Yingrui (Autor:in) / He, Jiazhi (Autor:in) / Lu, Bin (Autor:in) / Wang, Kaiping (Autor:in)
KSCE Journal of Civil Engineering ; 28 ; 5934-5945
01.12.2024
12 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Bayesian-Network-Based Predictions of Water Inrush Incidents in Soft Rock Tunnels
| Springer Verlag | 2024
A New Bayesian Network Model for the Risk Assessment of Water Inrush in Karst Tunnels
| DOAJ | 2022
Analysis of Mechanism of Water Inrush in Karst Tunnels
| ASCE | 2006
| Europäisches Patentamt | 2022
Analysis of the Mechanism of Water Inrush in Karst Tunnels
| British Library Conference Proceedings | 2006