Institute for Geotechnical Engineering Research Research Projects
Probabilistic Design of Offshore Foundation Piles – IRP-WIND

Probabilistic Design of Offshore Foundation Piles – IRP-WIND

Led by:  Prof. Dr.-Ing. Martin Achmus
Team:  Dipl.-Ing. Kirill Schmoor
Year:  2018
Funding:  13 EU organisations
Duration:  03/2014 – 02/2018
Is Finished:  yes


In contrast to the conventional deterministic design of offshore foundation piles, the probabilistic design describes uncertainties from the soil properties and their determination not by characteristic values but by specifying statistical values such as mean value, standard deviation and distribution form. In addition, other sources of scatter, such as the uncertainty for the design method used (model error), can also be taken into account in the design.

Taking the effect into account, a limit state equation for a failure of the system can be established. By integrating the limit state equation, e.g. using the Monte Carlo method, the probability of failure of the system under consideration can finally be determined. By comparing this with the desired target safety, a statement can be made as to whether the system fulfils the required reliability.

The focus of the IRP-WIND research project is on the effect of the model error of the tensile load-bearing capacity on the stochastic resistance for axially loaded piles. In this context, the model error is usually expressed as the quotient of the actual measured resistance value by the calculated resistance. Numerous studies confirm that the model error generally has the greatest influence on the stochastic resistance and should therefore be taken into account with great care. The model error is strongly dependent on the data basis under consideration. Only a few tests are currently available for offshore boundary conditions, meaning that a statistical evaluation of the model error is not very meaningful.

As part of the IRP-WIND project, five static tensile tests are being carried out on piles. By analysing the measured and calculated load-bearing capacities, the data basis for the model error can be expanded in order to obtain a reliable statement regarding the probability of failure.