Presenters

 

Julien Chauchat

Title: SedFoam: an open-source multi-dimensional two-phase flow model for sediment transport applications"

Julia Steiner

Title: FSI of membrane wings at high Reynolds numbers

Abstract: The field of Airborne Wind Energy (AWE) is concerned with harvesting high-altitude wind power. While there exist a multitude of approaches towards harvesting the wind potential at high altitude, in the AWE group in Delft substantial research efforts have been put into accomplishing this using a kite connected to a fixed ground station equipped with a generator through its tether. Due to the high complexity that such a system ensues, it is of vital interest to have a fundamental understanding of the critical aero-elastic modes of the structure in flight for design and control of the device.

The aim of this project was to further extend on the existing research efforts in this area by coupling a high fidelity Reynolds-Averaged Navier-Stokes solver with an improved version of the existing structural solver framework and validate the model on simplified test cases using a partitioned approach. The devel- oped methodology uses the open-source Computational Fluid Dynamics (CFD) solver foam-extend with an adapter to the coupling library preCICE as implemented in the FOAM-FSI toolbox for the fluid model. For the structure model, an in-house Python code based on a nonlinear shell element formulation is utilised.
A literature survey revealed that while there exists an abundance of publications on strongly coupled Fluid-Structure Interaction problems, the majority of them are targeted at applications set in completely different flow regimes. Thus, the most difficult part of the project was to find appropriate validation data for membrane wings at high Reynolds numbers. Nevertheless, the capabilities of the approach were demonstrated on three test cases.

First, the solver was applied to the classic FSI benchmark from Turek & Hron set in the laminar flow regime [1]. The relative error in the results for the relevant properties was at worst 10 %. The main source of inaccuracy is most likely the application of the thin shell element formulation to a comparatively thick beam. Thus, the error is expected to decrease when applied to thinner structures for which the shell model was developed for.

Finally, the solver was applied to two test cases with more realistic Reynolds numbers, where a thin, flexible material was wrapped around rigid leading and trailing edge supports and the force on the wing was measured for different membrane slack and angle of attack configurations [2, 3]. Comparison of the results from numerical models and experiments showed that the general trends for different slack lengths of the wing and angle of attack sensitivity were well captured in the numerical outcomes. However, the qualitative validation was less promising with a relative error in the mean lift and drag values of up to 15 % and 50 %, respectively. While some of the deviations can be traced back to ambiguities in the experiment description, the actual accuracy of the aeroelastic solver for membrane wings at high Reynolds numbers is yet to be determined.

Concluding, this project developed a methodology to model membrane wings at high Reynolds numbers. While the capabilities of the method have been successfully showcased on a classic FSI benchmark case, only a partial validation on benchmark cases at more realistic Reynolds numbers has been carried out. Hence going forward a full quantification of the accuracy of the method for its target application range is recommended.

 Kelvin Loh

Title: CFD for the built environment

Dave Weij

Title: Modeling the unstable breaching process with OpenFOAM