The nature of environmental interactions, as well as large dimensions and complex structure of marine offshore objects, make designing, building and operation of these objects a great challenge. This is the reason why a vast majority of investment cases of this type include structural analysis, performed using scaled laboratory models and complemented by extended computer simulations. The present paper focuses on FEM modelling of the offshore wind turbine supporting structure. Then problem is studied using the modal analysis, sensitivity analysis, as well as the design of experiment (DOE) and response surface model (RSM) methods. The results of modal analysis based simulations were used for assessing the quality of the FEM model against the data measured during the experimental modal analysis of the scaled laboratory model for different support conditions. The sensitivity analysis, in turn, has provided opportunities for assessing the effect of individual FEM model parameters on the dynamic response of the examined supporting structure. The DOE and RSM methods allowed to determine the effect of model parameter changes on the supporting structure response.
The analyzes were aimed at demonstrating the influence of parameters describing the deformation of the structure on the uncertainty of critical force, and the impact of technological imperfections on stress uncertainty in compression conditions. In a linear buckling analysis, the problem is considered only for the initial, permanent state of the stiffness matrix. In the case of demonstrating the influence of initial deformations on the behavior of the structure under load, it is necessary to visualize changes in stiffness over time. To this end, a non-linear MES analysis was carried out, which will take into account local changes in the stiffness of the model through a gradual increase in the load. Thus, the difference in stiffness is taken into account, which in the linear problem is infinite. The analysis was used to examine the local and global sensitivity of the parameters describing: plating thickness as well as deformation caused by the technological process on the stress value reduced by Huber hypothesis, and the value of normal stress. To take into account the influence of non-specified values of the magnitude of geometric deviations, and their simultaneous influence on the range of obtained results, the Experimental Planning Method and the Surface Method of Answers were used.