Prediction of time-dependent tunnel convergences using a Bayesian updating approach. (December 2019)

• Record Type: Journal Article
• Authors: Feng, Xianda; Jimenez, Rafael; Zeng, Peng; Senent, Salvador
• Is Part Of: Tunnelling and underground space technology. Volume 94(2019)
• Publication Date: 2019-12
• Subjects: Tunnel time-dependent convergences -- Bayesian updating -- Uncertainty -- Measurement error; Tunneling -- Periodicals; Underground construction -- Periodicals; Tunnels -- Periodicals; Underground areas -- Periodicals; 624.193
• DOI: 10.1016/j.tust.2019.103118
• Languages: English
• ISSNs: 0886-7798

Abstract

Highlights: Measured convergences used to update predictions of future tunnel convergences. Various types of uncertainties are considered. Model improves with quantity and quality of measured data. Particularly good early predictions, when their information is more valuable. Abstract: Convergences caused by tunnel excavation may increase with time. The prediction of these time-dependent convergences is important for the safe design and construction of tunnels. This study proposes a Bayesian approach to improve time-dependent convergence predictions, updating them with new information provided by successive convergence measurements. The proposed approach can consider various sources of uncertainties such as model uncertainty, model parameters uncertainty and measurement uncertainty. Three real tunnel projects —the Frejus road tunnel, the Babolak water conveyance tunnel, and the GCS drift of the Underground Research Laboratory (URL) of the French National Radioactive Waste Management Agency (Andra)— are used to demonstrate the applicability and performance of the proposed approach. Results show that the accuracy of predictions is improved, and that their uncertainty is reduced, after the measured convergences are employed to update prior predictions; and results also show that such predictive improvements due to the updating become more significant as the measurement accuracy increases.