Efficient modelling of wind turbines in time-domain considering uncertain parameters

• Applicant: Professor Dr.-Ing. Raimund Rolfes, since 4/2021
• Subject Area: Applied Mechanics, Statics and Dynamics
• Term: since 2019
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 436547100

Wind energy is a promising technology to achieve the objectives set for the development of renewable energy. To increase competitiveness, costs have to be reduced and the structural reliability has to be improved. A promising approach are more realistic simulations of wind turbines by considering polymorphic uncertainty. In this context, uncertainty is, for example, variability, incompleteness, and inaccuracy of data. Polymorphic uncertainty can be modelled by using imprecise probability. In research, for classical applications of civil engineering, imprecise probability becomes increasingly popular in recent years. However, for wind turbine applications, there are no approaches that use imprecise probability. The main reason is the complexity of wind turbines that combine challenges regarding uncertain and scattering inputs (typical for civil engineering) and complex controller actions (typical for mechanical and electrical engineering). This complexity leads to high computing times and hinders accurate meta-modelling, that is normally used, if computing times are not manageable. That is why in this project, at first, adequate imprecise probability methods are applied to wind turbine models. Subsequently, the efficiency of the uncertain analysis is increased by reducing the required number of model evaluations. This is the core of this project. First, the increase in efficiency is achieved by using enhanced sensitivity analyses, which can be applied when imprecise probability is utilised. By means of sensitivity analyses, the number of uncertain parameters can be reduced. Second, sampling techniques are developed, which can be combined with imprecise probabilities and load extrapolations for wind turbine fatigue loads.This enables an efficient modelling of complex wind turbines using polymorphic uncertain data. At the same time, computing times are kept manageable. Hence, more realistic simulations are possible.

• DFG Programme: Research Grants
• Ehemaliger Antragsteller: Professor Dr.-Ing. Cristian Guillermo Gebhardt, until 3/2021