Validation of earthquake analysis methodology of a suction-caisson foundation-structure through model testing

Suction caissons are one of the most widely used foundation solutions for subsea structures and wind farms. Seismic response of subsea structures is however seldom documented properly, often just treated as a foundation capacity issue applying a quasi-static acceleration and not considering the inertial interaction between the structure and the soil. The more relevant tasks to document are the motions of the unit and the response of the externally connected flowlines and equipment/systems on the unit.Based on a case study located in the Shah Deniz field in the Caspian Sea, model centrifuge tests and numerical modelling were carried out to validate the global response of a 4-caisson supported manifold structure subject to seismic motions in soft clay. The centrifuge tests were carried out at 58 g at the centre for geotechnical modelling at UC Davis. To simulate the soil-structure interaction, a series of non-linear springs defined by kinematic hardening models were used in analyses with the ABAQUS software. This development includes the algorithms for determining the required model parameters. A very good agreement between recorded response from the centrifuge test and calculated response from the FE-analyses was achieved.The development and validation of the soil model presented in this paper is an improvement in design methodology for caisson foundations subjected to earthquake loading. The non-linear soil springs are well suited to incorporate in more detailed structural analyses where an accurate representation of the foundation response is required. The paper also briefly describes how the subsequent earthquake design analyses were performed for the Shah Deniz manifold structures making use of the validated soil spring model and the added value it gave to the project.