At Airbus Defense and Space we design and build solar array wings to supply large scientific satellites with power. Current development in space industry show that design by simulations without system tests is becoming the holy grail. Understanding the behavior of a wing and being able to quantify the credibility and the uncertainty of the simulation results is than essential.
Unfortunately, presently, we cannot predict the dynamic amplification of a wing with a FEM model only. Dynamic tests at system level are still always required for deriving the actual dynamic amplification. It is possible to predict the eigenfrequencies rather accurately (at least the main important frequencies), but especially damping which is driving the amplification is a guess. We assume usually 2.5% viscous damping, but that is uncertain and also depends on the frequency and the level of input. Apparently, the resulting dynamic amplification is a non-linear effect that is still not understood and thus is not correctly modelled.
So for a better understanding of the dynamics and more accurate predictions better models are required. However, we do not have good models to introduce the damping. We think it is mainly due to local behavior at the interfaces between the components the system is composed of and which is not modelled. However, this has to be established. That means, we are searching for models that describe the damping (dynamic behavior) at the interfaces and use these models in non-linear FEM models to predict the behavior. Subsequently we need to do test to verify and validate these models. The final goal is of course to use these models in a wing model and verify the predictions with the dynamic behavior of a solar array when it is excited at a base.
To summarize
- Dynamic models are not sufficient accurate and therefore still expensive system tests are required
- Probably damping at the interfaces is not correctly modelled
Therefore, the goal is to
- Improve the dynamic models that enable the prediction of dynamic behavior sufficient accurately
- Enable this with help of local models that still can be applied combined with linear FEM predictions
The thesis both requires knowledge of FEM and requires tests to be carried out. Therefore, dr. Dario di Maio will also be heavily involved.
- Company supervisor: dr. ir. Marcel Ellenbroek
- UT supervisor: dr. Dario di Maio
For more information on this thesis please contact:
Contact Person
