How to model a wind tunnel for aerodynamics testing in Solidworks?
A tutorial would be great. All flow simulation tutorials cover liquid flow mainly, I haven't seen a tutorial about solving aerodynamic problems in Solidworks yet.
There on net is a tutorials of FLUENT. For fast results you must using a simetry of models and defining a right boundary conditions - all results depends a very of given boundary conditions, little changing a values in boundary conditions you may get a totally different results.
Sometimes you need be a lucky boy to get right answers.
I'm actually not working on anything specific at the moment, I was thinking to use Solidworks for learning something about aerodynamics. I think this would be a great way to see how different shapes and conditions would affect my design in the future.
This should be pretty straight-forward; create a "big" cylinder cut your part out of this cylinder and you've got your wind tunnel... It's not SW specific, it's the general approach in CFD... Make sure your cylinder is much bigger than your part to let the flow develop itself, use the front as an inlet, the back as an outlet and the sides as walls...
That's not the way to go unfortunately. The best approach is back to basics:
- set out objectives
- choose the correct method of CFD (ie SW offers 3 x different types of computational methods
- Set up file accordingly
In SW, using the flow simulation model you could run it as an 'external' simulation or internal by drawing a cylinder ( a rectangle is actually better) & create lids on either end. Then, specify the inlet and outlet conditions. This way, you can easily run the same Sim. at different Reynold's numbers.
The hardest part of any CFD is creating the mesh for accuracy. It's a nightmare in all CFD software. If possible, it's far better to do a 2d flow sim on a slice of the object. At one stage, my simulation runs we're cutting up a 0.3m^3 space with an object into 2.5 x 10^6 cells. And still the numbers were not that great compared to actual results. CFD is great to learn what exactly is going on and how to tweak a design, but the numbers cannot be relied upon.
Mabye you should read this doc. The good stuff starts on page 29.
Hi Ingmar (thanks David! I will get you back...),
This is a 'how long is a piece of string' kind of question.
But to give you a little structure and help I will do my best;
Modelling of any type has restrictions of course, but not as many as most think...
Too often people attempt to simulate too many variables at once.
Simply, if you clarify your objectives it's easier to knock them over one at a time. SW can practically do anything, but the tutorials only go so far. Fluids and Aerodynamics do overlap but not by alot.
So, what exactly are you look to find out / assess?
You know James, there is an expression that applies to your comment: "knowing what you don't know". I'm going to try to be as polite as possible, but your statements are plain and simple wrong. CFD cannot be relied upon? That is one of the most ridiculous statements I have heard. What you are describing is really either your model being inadequate or the solver not being capable of modeling the physics. Either way, the only one to blame is the user, either he made a poor model (2.5M cells is not that big for a CFD model) or he does not fully understand the physics to model and the limitations of the code. Meshing is not a nightmare in a lot of codes, it might be in XFlow, but let's not generalize. As for your comment on a rectangle being better than a cylinder, ask yourself what a sharp corner does to the continuity equations. If you can find the answer, you will then see what I mean :-) Technically speaking, it's actually a sphere that should be used to model the fluid...
It appears that you benefit from some kind of self-proclaimed or perceived expertise, but in this case I just cannot let it go any longer and must question it. I have been involved in this field for a very long time and am running a CFD consulting business.