This thesis presents the formulation of geo-mechanical design parameters using a combination of laboratory and insitu test results in the heavily overconsolidated cohesive till formation in the city of Edmonton. These data were then used in the design of the recently constructed North LRT twin tunnels.

This study initially focuses on the geology local to the test site as observed in a deep excavation constructed as part of a foundation and cut and cover tunnel section. During excavation of the North LRT twin tunnels, regular face mapping was carried out and several geological structures, not previously identified have been observed. The presence of these formations is discussed in terms of their potential genesis and the potential impact on stability of the unsupported tunnel cutting.

The yield criterion of the major stratigraphic units were then explored using a detailed reassessment of previous laboratory experiments as well as new data obtained from a series of insitu testing. The previous studies indicated the importance of the stress path in terms of the displacements measured around an excavation. What was not determined was the role the stress path had on the yield strength and the strains required to achieve a state of plasticity. This study has ascertained that not only the stress path and strains to yield are critical to the strength of the soil, but also the state of stress prior to testing.

These data indicate that the conventional upper and lower strength bounds of drained and undrained strength might not be the dominating failure state in hard, fissured soils like the Edmonton till. Because the typical construction rates were less than 0.1 m/hour per tunnel, the excavation rate should result in drained conditions around the tunnel cavity. Additional in-situ test methods for heavily overconsolidated soils have been developed in order to shed light on parameters that are conventionally very difficult to obtain. These new methods shed light on limitations of previous test methods and how the results of analysis may be influenced by the unsaturated state of the soil. These data were then supplemented by displacement measurements recorded during the twin tunnel construction. Regular monitoring of both surface and in-tunnel displacements provided a basis for analysis to determine the extent of ground movement into the tunnel cavities as well as the influence of the second, lag tunnel on the constructed lead tunnel. The presence of a variable spacing between the twin tunnels (pillar width) provided a basis for comparison of the stresses and displacements associated with the lag tunnel influence. These displacement data provided the basis for a numerical back analysis as well as determining the efficacy of the tunnel construction methods.

Finally a numerical back analysis of the measured displacements was undertaken and compared to the relevant yield criterion. This demonstrated that the influence of a closely spaced tunnel on a previously constructed tunnel is not as extensive within heavily overconsolidated soils as previously reported. Knowledge of the stress path and the associated yield strains, relevant yield criterion are assigned to the ground in order to illustrate where and how yielding of an unsupported tunnel cutting will occur. This also suggests that non-conventional methods of analysis should be employed as part of the design process to determine the size and shape of translational failure expected within a fissured underground opening.