Undrained behaviour of brecciated Upper Lias Clay at Empingham Dam

Empingham Dam was built in the early 1970s on a brecciated Upper Lias Clay foundation of fill derived from it. When the dam was designed, the undrained behaviour of intact stiff plastic clay in the field was poorly understood. Thus, a conservative bound to conventional strength tests was adopted to provide an undrained strength profile in the foundation. The shearing resistance of the foundation was also investigated by incorporating an instrumented 24 m high trial bank in the upstream berm of the main dam. As the trial bank was nearing completion, two slips occurred in temporary borrow pit slopes. It was considered that this field experience, together with undrained triaxial tests with pore pressure measurement on large diameter samples carried out in the laboratory, justified the original design assumptions. However, the end lengths of the main embankment on the undrained foundation started to spread laterally when the fill was some 2 m below the final level. This was in spite of the fact that the conventional factor of safety, calculated using the original design assumptions, was in excess of two. This paper describes the finite element (FE) back analysis of the observed behaviour. For this purpose two different constitutive models have been employed: a kinematically hardening ‘bubble’ model accounting for pre-peak plasticity, and a non-linear elastic Mohr-Coulomb plastic model accounting for strain-softening. It is demonstrated that the observed undrained behaviour of the Empingham foundation can be reproduced by FE analyses, but only if the undrained shearing resistance varies from section to section.