A probabilistic assessment of the effect of tunnelling on groundwater system considering the uncertainty of hydraulic conductivity

• Published in: Tunnelling and underground space technology Vol. 21; no. 3; p. 445
• Main Authors: Ryu, Dong-Woo, Son, Bong-Ki, Song, Won-Kyoung, Lee, Chung-In
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.07.2006
• Subjects: Groundwater; Hydraulic conductivity; Simulated annealing; Uncertainty; Uncertainty; Hydraulic conductivity; Groundwater; Simulated annealing
• ISSN: 0886-7798; 1878-4364
• DOI: 10.1016/j.tust.2005.12.085

MODFLOW, a three-dimensional finite difference code, is widely used to model groundwater flow and assess the effect of excavations on a groundwater system due to construction of subways and mountain tunnels. The accuracy of the results of such numerical analysis largely depends on boundary conditions, initial conditions, conceptual models and hydrogeo-logic properties. Therefore, its accuracy can only be enhanced using more realistic and field oriented input parameters. Accurate modelling of the subsurface parameter distributions becomes more difficult with increasing heterogeneity and uncertainty with respect to spatial variability of the available data. To infer how the uncertainty of the input parameters, especially hydraulic conductivity, influences the behavior of groundwater flow, the hydraulic conductivity in a groundwater flow system is assumed to be spatially distributed as a stochastic process. The uncertainty depends both on the quantity and the quality of the available data. The number of field tests such as the injection test is very much deficient to cover a study domain. In this study, SA technique was used to consider the uncertainty of hydraulic conductivity and heterogeneity and anisotropy of subsurface condition. SA technique was applied for the integration and optimization of hard data like hydraulic conductivity and soft data like P-wave velocity image with satisfying the predefined statistics and spatial variability. It is found that SA technique quantified the uncertainty of the hydraulic conductivity in the hydrogeologic medium and identified how an uncertainty of hydraulic conductivity propagated into results of groundwater modelling. This approach provided an efficient tool for the probabilistic assessment of the effect of tunnelling on a groundwater system through a number of realizations of hydraulic conductivity random field by SA technique. As an environmental aspect, the results of probabilistic assessment are found to be useful information for a reasonable construction strategy.