Designing a Glider
Recently, some of my friends have been working on designing a small DIY glider. For those who don’t know what glider is, it’s an unpowered aeroplane. How does it fly then? Well, in most cases, a small aircraft such as Cessna gives the glider a tow when they’re taking off, and once they’re up in the air, the glider is released and glides down gradually to the ground.
One of the guys is in charge of designing the overall design including the main wings. He’s calculated the approximate airspeed, weight, length, etc. and has chosen an airfoil which in theory is suitable for the glider.
So now, here, me having been using ANSYS Fluent and Polyflow for my graduation thesis for well over half a year, I wanted to see the actual flow around the glider (although it’s not 100% accurate).
CFD stands for Computational Fluid Dynamics. It’s like having a wind-tunnel inside a computer.
For the simulation, I’ve used Simscale which is a cloud based FEM, CFD and Thermal Simulation service. Although it utilises openFOAM for all simulations, it doesn’t even require any software installation.
Here is the CAD model of the glider that my friends have created.
The construction is fairly simple and it consists of four main parts: wings, body, vertical and horizontal stabilisers. The actual airfoil shape has only been applied to its main wings since the stabilisers are not there to generate the lift. (well, at least for this project)
Especially for the external flow simulations, a rather large boundary box is required as well as some additional cartesian boxes in which the meshing size is significantly smaller thus the calculation is more accurate whilst saving a lot of time and CPU load. Plus, as the glider is symmetry. only the half of the original model is actually inside the boundary box to achieve faster process.
I originally tried to generate the mesh with Autosizing. However, as I’d imagined, the mesh quality around the glider’s surface was pretty rocky. So I decided to create another one manually myself. I looked into quite a few public projects as references, and finally here is what I’ve got.
Inside the general mesh box, I’ve inserted three cartesian boxes, the largest one is around the glider itself, extra box around the wings, and the smallest one around the stabilisers at the back. The surface refinements have also been applied along with the surface boundary creation.
However, mainly due to the model being too sharp on some edges, the mesh refinement on the surfaces has failed.
In the end, unless I have a look into the CAD model and revise the overall design slightly to make it analysable, an accurate result won’t be achieved. I’ve also asked the Simscale community for help on this project and one of the members has managed to run the simulation successfully with the automatic meshing. But one of the objectives of this project was actually to learn Simscale, so I had opened Solid Works to reconstruct the model before my mate informed me he had changed the overall design, into a flying boat….