Data-Driven Approaches in Computational Mechanics for the Aerohydroelastic Analysis of Offshore Wind Turbines

A massive upscaling of offshore wind turbines is necessary to reach the goals of the European Green Deal. However, current methodologies for analysis and design are at their limits. One of the major bottlenecks is that it is so far not possible to directly integrate experimental data into aerohydroelastic simulations of offshore wind turbines. By and large, including these data into the aerohydroelastic analysis is indirectly accomplished through the offline adjustment of those parameters that define instances of existing models. Although such a practice can reasonably improve the short-time predictive capability, the underlying models remain unmodified. Thus, further physics available in the data remains inaccessible. In the numerical simulation context, this represents a main challenge to taking advantage of the experimental data in their entirety. In this context, DATA-DRIVEN OFFSHORE proposes to simultaneously integrate these highly-valuable data into aerohydroelastic simulations through data-driven computational mechanics. Such an approach is one of the most advanced computing frameworks, and relies on the reformulation of classical boundary and initial value problems in solid and fluid mechanics such that constitutive models, boundary conditions and/or applied loads are directly replaced by some form of experimental data. DATA-DRIVEN OFFSHORE will thus enable for the first-time investigation of the aerohydroelastic behavior of an offshore wind turbine relying truly on experimental data, capturing the hidden features that these contain. This will greatly improve the predictive capabilities with respect to existing models, allow the conception of less-conservative designs, and enable upscaling beyond 20 MW of rated power, increasing the efficiency while reducing the cost per unit of power produced. Thus, it will contribute to triggering a change of paradigm for future generations of offshore wind turbines.