Global versus Local Simulation of Geotechnical Parameters for Tunneling Projects

• Published in: Journal of geotechnical and geoenvironmental engineering 2020-07, Vol.146 (7)
• Main Authors: Grasmick, Jacob G, Mooney, Michael A, Trainor-Guitton, Whitney J, Walton, Gabriel
• Format: Article
• Language: English
• Subjects: Best practices; Computer simulation; Geology; Geostatistics; Geotechnical engineering; Heterogeneity; Identification methods; Modelling; Parameter uncertainty; Parameters; Probabilistic risk assessment; Probability theory; Risk assessment; Simulation; Soil; Spatial variations; Statistical analysis; Stochastic models; Stochasticity; Technical Papers; Transition probabilities; Tunnel construction; Tunneling; Tunneling shields; Uncertainty; Variability; Variance; Variance analysis
• ISSN: 1090-0241; 1943-5606
• DOI: 10.1061/(ASCE)GT.1943-5606.0002262

AbstractUrban soft-ground tunneling projects involve significant risks related to the spatial variability and uncertainty in geotechnical parameters. However, standard practice typically does not incorporate spatial trends into risk assessment. Geostatistical methods provide a means not only for predicting geotechnical parameter values spatially, but also for modeling the heterogeneity and spatial uncertainty that play a key role in probabilistic risk assessment for tunnel construction. Before incorporating geostatistical analysis into the risk assessment for soft-ground tunneling works, it is necessary to identify best practices with respect to geostatistical methods. In this paper, two approaches were examined and compared for modeling the spatial variability and uncertainty of key geotechnical parameters relevant to shield tunneling in soils. The first approach consisted of the sequential Gaussian simulation of parameters using a single spatial variance model for each respective parameter, which is a common approach adopted in the literature but does not incorporate variability and uncertainty in geological units. The second approach considered the influence of geology by basing the sequential Gaussian simulation of geotechnical parameters on geological unit simulations using a transition probability-based stochastic model. In this approach, a unique spatial variance model of the geotechnical parameter for each geological unit was considered. The results from this analysis revealed that the influence of geology is critical to the spatial modeling of geotechnical parameters and their uncertainty, and, therefore, must be incorporated into the geostatistical analysis for the risk assessment of soft-ground tunneling works.