Probabilistic Rock Mass Classification from Prior Construction for the Future Extension of a Tunnel

This paper presents a procedure to re-analyze the original design of the Eisenhower-Johnson Memorial Tunnel (EJMT), which opened in 1973 and 1979 when rock mass classification systems were being introduced.The re-analysis used a new probabilistic approach to rock mass classification and stochastic representation of ground and support parameters.The results will be used to select the alignment of future extensions of the EJMT.Rock mass classification systems using Q and RMR were applied probabilistically in the back analysis of the tunnel design.Distributions of Q and RMR were developed for discrete points along the tunnel for prior and future tunnel alignments.Based on the probabilistic rock mass classifications, installed support systems were compared to actual support systems.Important differences between deterministic and probabilistic results were observed, including bias of deterministic results away from peak probabilistic results.Numerical models using rock mass properties based on probabilistic rock mass classifications were found to provide a good match to rock loads recorded during tunnel construction.This result suggests that forward modeling of tunnels using probabilistic rock mass classifications may provide a more realistic estimate of tunnel support requirements, costs, and risks.Current deterministic geotechnical baseline reports may be inadequate to provide a complete picture of a proposed project and may bias design towards a sub-optimal solution.