This report describes aspects of probabilistic analysis for offshore substructures and the results obtained in various illustrative investigations. This includes probabilistic modeling of model uncertainties using results from WP7.2 and WP7.3 on experimental data and conventional design approaches.Since no failures / collapses of wind turbine support structures are reported in the literature / available databases statistical analyses have not been performed to assess the reliability level. However, failures of the grouting in a large number of grouted monopiles have been observed. These failures which can be considered as design errors / lack of knowledge seem only to have resulted in local failures without collapse of the substructures, but haverequired various mitigation efforts incl. extensive inspection and monitoring programs. For new designs the problem with groutings has resulted in changed designs with e.g. shear keys or conical joints. The target reliability level in the design standardsfor substructures corresponds to an annual probability of failure equal to approximately 10-4–5 10-5.Reliability analysis of offshore support structures can be performed using structural reliability methods implying formulation of limit state equations for the critical failure modes due to extreme loads and fatigue, and establishment of stochastic models for the uncertain parameters in the limit state equations incl. physcial, statistical and model uncertainties. The probability of failure can be estimated by simulation techniques or First/SecondOrder Reliability Methods. For the grouted joints a number of limit state equations are formulated related to critical failure modes, including failure of the grout concrete in extreme loading and fatigue, and fatigue failure for the welded steel details. Additionally,stochastic models are established. Reliability analyses are performed using the limit states and illustrated for the 10 MW reference wind turbine substructure for which load effects areobtained for extreme loads, and combined stress ranges and mean stresses are obtained for fatigue analyses using a detailed finite element modelling of the grouted connection. For extreme loading failure of the concrete in shear and compression is investigated as well as critical vertical settlement. Further, failure of the concrete in fatigue is considered. Among others, the results show that the reliability level is sufficient related to settlement of transition piece keeping the gap between the pile top edge and the jacking brackets below 72 mm.Fatigue reliability of concrete grout and steel components of grouted connection with shear keys was assessed and found to be satisfactory at this stage of preliminary design. The results show that the concrete grout reliability models are highly sensitive to model errors related to the estimation of the SN curve, especially for the ‘compression to tension’ case where more test data are required. The reliability analyses of fatigue failure of the welded steel details show a satisfactory reliability level, assuming cathodic protection during the whole lifetime but also high sensitivity to assessment of the load stress ranges and stress concentration/magnification factor calculations. Additionally, the results show that for welds in both the monopile and transition piece larger initial cracks could be allowed potentially reducing the manufacturing costs of monopiles and transition pieces.For the geotechnical reliability assessment of foundation piles typical offshore soil and foundation pile properties for offshore wind converters in the North Sea were assumed. Five different design methods for the determination of the axial tensile resistance were considered, and 60 deterministic designs were evaluated within a reliability based design. Typical variability for the assumed soil condition and estimated model uncertainties were applied for a reliability based estimation of failure probabilities in terms of reliability indices. A new calibration approach was developed to determine the global safety factor for a prescribed acceptable failure probability. Further, quality factors were derived for each design method as function of the load, pile diameter, and soil density. Nest, probabilistic calculations of fully coupled offshore wind turbines were considered in order to assess the reliability of substructures, based on sophisticated, non-linear aero-elastic simulations, and furthermore, to create the basis for safety factor (SF) calibrationIt can be concluded that SFs can be reduced if probabilistic calculations are performed implying possible weight saving potentials. The results also include global sensitivity analyses whereby the most influential probabilistic parameters were identified showing that wind and wave, and some soil parameters, have to be modelled by probabilistic models.
|Number of pages||109|
|Publication status||Published - 2018|