Methods to Reduce Geotechnical Uncertainty and Risk Using Big Data Collected during Construction

The very high level of geotechnical uncertainty and geotechnical risk during the planning and design phase of a tunnel or underground construction project is attributed to current site investigation techniques that result in a very low fraction (e.g., less than 0.01%) of the ground being characterized. This paper demonstrates how the extensive collection of real-time geotechnical monitoring and tunnel construction equipment data can be used to update local estimation of key geotechnical parameters that in turn enables the reduction of geotechnical risk. The paper first demonstrates how site investigation data can be used to develop pre-construction predictions of important subsurface project aspects, including tunneling-induced ground deformation and TBM performance prediction, e.g., tool wear, clogging risk. These apriori probabilistic predictions carry a high degree of uncertainty due to the very low percentage of ground that is characterized. The value of such predictions is therefore tempered by such high uncertainty. The uncertainty can be dramatically reduced through Bayesian updating, herein applied to tunneling-induced ground deformation.