Uplift resistance of soils

Abstract

A scheme is described for the water coolant system at Sizewell A nuclear power station, where eight shafts were jacked up from tunnels to the sea-bed through 6 m of cohesionless soil. This represents the design for failure case. Existing theories based on the design for safety case were inadequate and the required jacking pressures, to overcome the uplift resistance of the soil, were successfully determined from laboratory model tests where dimensional similarity was satisfied between model and prototype. The Sizewell and field test results were compared with predictions from centrifuge tests and finite element predictions which were carried out later for the design for safety case. For cohesionless soils the centrifuge tests gave similar results to the model and field tests, whereas the finite element predictions grossly underestimated the ultimate uplift resistance for other than loose sands. A comparison is also made between predictions of the uplift resistance of cohesive soils from model tests and from theoretical studies, including finite element methods. Recent field tests on belled-out piers in cohesionless soils are compared with the results from the original Sizewell model tests. Attention is drawn to the increasing number of specializations within geotechnical engineering, the fragmentation that could develop between geotechnical engineers because of these specializations, and the necessity of keeping an open mind on, and a broad approach to, the solution of problems in geotechnical engineering.