This content has been marked as final. Show 4 replies
Do you mean that you are going to simulate a 2D or 3D flow around an airfoil? If so, the shear force should be your induced drag and the total normal force will give you the total reaction force. About the boundary conditions, make an aerodynamic tube with inlet and outlet, the diameter should be 2-3 times bigger than your wing chordlength, and apply all that you need on its inlets and outlets. And, use a real wall conditions.
The following would be my simulation.
1) Import the aerofoil NACA0015 from autocad to soilidworks.
2) Extrude the following aerofoil.
3) In 2D simulation, I have created the computaional domain.
4) Now my question, can I create the boundary condition on the wall of the aerofoil and how? U were mentioning about real wall condition, may be u want go into the details.
5) How do I select the outlet of the computaional domain to have the same pressure as the inlet pressure?
6) When using the initial condition, what doe the laminar and turbulent function means because there is also a function u can select either laminar or turbulent?
7) The ither question is that, I need to find the transition point where the laminar actually converts to turbulent. How can I do it in cosmosfloworks?
Please when answering to these questions try as much to give more details. I would like to include that my campus have purchased the Cosmosfloworks and want us to evaluate further whether the Cosmosfloworks would be suitable for the analysis of aerofoil.
I have done a NACA 0012-64 and with mesh refiniments and at low alpha it nailed the L/D as an equation goal. I have my suspcisions about how much of hte range over which it would do this but hey it was just for a demo and worked great.
If the section is aerodynamically smooth you do not have to specify anything on the walls of the aerofoil when using an external flow or internal analysis for that matter. Though if you build the wind tunnel you could put convecting conditions on the wall. This approach can be more computationally efficient in terms of cell count. If you want to specify a roughness ont eh surfaces you can do this either as the global or you can apply a real wall condition. Note though that the real wall condition if the value is grey then it is Not what you are getting - puch the button and specify the value you want.
Hi Kannan: I have done this in two ways; 1) put a hollow box (part) around the airfoil to simulate a wind tunnel test section, and 2) use "External" flow (this is simplest). For 1), you can simulate upper and lower box surfaces moving at free-stream velocity (just apply a wall boundary condition to them with tangential flow), and put inlet velocity on the inside of the front surface of the box; outlet on the back surface of the box. For 2, choose the option to use "External" flow, then in dialog "Initial and Ambient Conditions", just specify the free-stream velocity and that's it. I would stick to "Laminar and Turbulent" solution because it is supposed to automatically predict when flow transitions from laminar to turbulent. As far as identifying where on the airfoil the flow transitions: that might be difficult because I understand the program uses an algorithm call "Law of the wall" with something called "y+" values. If you look in the Fundamentals guide, I recall they show how to plot that variable and how to interpret the plots. I recall from some readings that the transition is extremely difficult thing to predict and you may not really be able to find it. Good luck on this. As far as the program capability is concerned, it is really good at predicting performance changes in designs. For instance, try airfoil A and then try airfoil B with the exact same conditions and compare results. The changes in performance can be predicted very accurately, as the tool is best at this. I have tried correlating airfoil test data lift coefficients and it is exceedingly difficult (external flow), as I know it is with all CFD codes. I had difficulty getting "exact" lift coefficients to match test data (which has uncertainties in measurement, too), however I could match the lift slope nearly dead-on (i.e., I tried several small angles of attack and all the lift coefficients differed from test data by the same percent, but the slope of the line was the same as test data). Sincerely, Tony