当前位置:
X-MOL 学术
›
Tunn. Undergr. Space Technol.
›
论文详情
Our official English website, www.x-mol.net, welcomes your
feedback! (Note: you will need to create a separate account there.)
Prediction method of rock spalling risk in large underground cavern excavation based on Bayesian network: A case study from the Baihetan hydropower station, China
Tunnelling and Underground Space Technology ( IF 6.7 ) Pub Date : 2025-01-13 , DOI: 10.1016/j.tust.2025.106381
Guo-Feng Liu, Zhi-Qiang Liu, Ding-Ping Xu, Quan Jiang
Tunnelling and Underground Space Technology ( IF 6.7 ) Pub Date : 2025-01-13 , DOI: 10.1016/j.tust.2025.106381
Guo-Feng Liu, Zhi-Qiang Liu, Ding-Ping Xu, Quan Jiang
Progressive cracking and the resulting rock spalling failure is a common type of disaster in the construction of deep large underground cavern in hard rock, which seriously affects the stability of surrounding rock and construction safety. Relying on the largest hydroelectric underground cavern engineering in the world, i.e. the underground powerhouse at the Baihetan hydropower station located in China, a prediction method of rock spalling risk in large cavern excavation process based on Bayesian network was proposed and applied in this study. Through investigations and statistics on the excavation of the first two layers of the Baihetan caverns, a database containing 101 sets of rock spalling samples was constructed, and the distribution characteristics and regularities of the indicators causing rock spalling were revealed. A comprehensive prediction index system that considers both internal and external indicators influencing rock spalling failures, including stress-strength ratio, geological structure, rock type and the geological strength index (GSI ), local energy release rate (LERR ), excavation rate and supporting time, was further constructed. The Bayesian prediction model of rock spalling based on multi-indicators was then constructed by a learning and testing process. Case back-testing analysis results showed that the model accuracy can reach 85 %, and inferential analysis of the BN model illustrates the influence extent of different indicators on the rock spalling risk. The proposed model was then applied to predict the rock spalling risk in the excavation layer III of Baihetan underground caverns, with an accuracy ratio of 82.93 %, proving the good applicability of the model in practical engineering, and the comparative analysis indicates that the accuracy of this proposed model is higher than that of prediction models constructed without considering external construction factors. This study can provide important support for the prediction and control of rock disasters in deep underground engineering.
更新日期:2025-01-13