Applicability of predicted cone penetration test profiles from geostatistical co-simulation on deterministic and probabilistic monopile foundation design for offshore wind turbines

To reduce site investigation costs for site characterization as well as timelines in offshore windfarm development, methods to predict the soil profile and essential geotechnical parameters at unexplored locations gain increasing interest from the industry. Hereby, the more abundant geophysical data can be used as guiding information in addition to the scarcely spaced direct investigations to improve the prediction of ground conditions and reduce its uncertainty at unexplored locations. However, since it is a rather new tool in offshore wind site characterization, little is known about the applicability of those predicted soil profiles and the associated geotechnical parameters for the design process of offshore wind infrastructures.This study aims to highlight the generation of predicted geotechnical profiles and parameters in the form of predicted cone penetration tests (pCPTs) and its use for the design of monopile foundations. Using real data from a German North Sea site the pCPTs are generated by applying geostatistical co-simulation, resulting in a best estimate with a location and depth specific uncertainty. These probabilistic results are utilized for the derivation of a design pile length. This pile length is either computed based on a deterministic estimate of the pCPT, taken as a quantile values, or with a reliability-based approach with respect to a target reliability level. It could be shown that a safe design length in comparison to the measured values is achieved by using a quantile value of 5% or less within a deterministic design or the reliability-based design approach.