Finite-element analysis of the failure and reconstruction of the main dam embankment at Abberton Reservoir, Essex, UK

摘要

An existing 15·5 m high main dam embankment at Abberton Reservoir in Essex was completed in August 1938, since when its performance has been satisfactory. However, the upstream embankment shoulder of the original dam suffered a deep-seated failure through its foundation towards the end of construction in July 1937, 9 days before a similar and well-known failure occurred at Chingford Reservoir in close proximity to Abberton. Whereas the failure at Chingford became an important case in the history of soil mechanics through the involvement of Karl Terzaghi and marked one of the first applications of modern soil mechanics principles, the failure at Abberton has remained largely unknown, until recently when raising of the existing dam started to be considered. This paper describes advanced finite-element analyses which were carried out to investigate the failure of the original dam at Abberton and the stability of the existing main dam. The parameters used in the constitutive models were derived on the basis of the available site investigations and laboratory testing and on experience in the back-analysis of other failures in London Clay. The analyses demonstrated that the upstream shoulder of the original embankment failed through the mechanism of progressive failure, which involved the top of the stiff plastic London Clay rather than the overlying alluvium in the foundation. The relatively rapid rate of embankment filling, achieved by using modern earth-moving equipment, contributed significantly to the original dam failure. The analyses also demonstrated satisfactory behaviour of the existing dam during reconstruction, the first impounding and in the long term, with its response being similar to that observed. Thus the constitutive models used and parameters derived were successfully calibrated against the observed behaviour of both the original and existing main dams at Abberton, and could be used in predicting the behaviour of the dam during and after its proposed raising.