I am student and now doing a project on a NACA 23012 airfoil. I am required to calculate the lift and drag coefficients. PHOENICS is used to do this. 1. The domain is rectangular with height=4.5m, length=7m and 1m to the z-direction. The airfoil is situated at the centre. 2. The flow is set to laminar. 3. The inlet velocity is 10m/s. 4. The airfoil's chord length is 4.5m. 5. The density is set to constant and is 1.189kg/m^3. 6. The kinematic viscosity is 1.544X10^-5. 7. The calculated Reynolds Number is 2.9X10^6. 8. Cartesian grid is used.

My approach: 1. the lift force is calculated by substracting the the term [P+rho*v^2] between the left and right stations of the domain. 2. the drag force is calculated by substracting the term [P+DEN*u^2] between the upper and lower stations of the domain. 3. lift and drag forces are computed simutaneously using the same control volume.

The results are then compared with the experiment done by NACA.

My problems are: 1. The control volume seems to be not giving good approxiamation for lift and drag forces simutaneously.

That is as lifts are good approxiamted then drags are not

and vice versa. Why? 2. The good results are only happen in some angle of attacks that is 2,4 and 6 degrees. Angles larger or smaller than that produce bad solutions. Why? 3. I try to change the domain size or stations used. However, the good solutions shift to certain angle of attacks or even not good solutions at all. Why?

1. Should I used different domain size to calculate the lift and drag force? 2. Any standard domain size should I use? 3. The experimental lift coefficients done by NACA are:

2 degree------ 0.2

4 degree------ 0.4

6 degree------ 0.5

Based on this data, what is the maximum % error on the solutions that are still be accepted?

(1). I think you are doing fine, because you are able to obtain solutions from a CFD code. That is good enough. (2). Then you said that you are not satisfied with your own answers, and the results are not consistent with the test data. (3). Since you are using CFD code to obtain the solution, the results should be different from the test data. (4). To make the direct comparison, you need to know the exact test conditions used in the airfoil test. Then you can use that condition in your simulation. In this way, you will have the same boundary conditions. (5). There is not much you can do about the test data, because that was done by someone long time ago. One way to eliminate this problem is to try other configurations. (6). For the CFD solution, at least there is one thing you can do, that is to make sure that the solution you obtained is mesh independent. You can double the mesh size (increase the total number of cells or grid points) and plot the lift vs mesh size curve. In this way, you can determine whether your solution is a function of the mesh size or not. If your solution is a function of mesh size, the it probably will be a function of other parameters such as the angle of attack. (7). Since it is a project, you also need to know that there is a smaple case available which actually shows the results and good comparison with test data. If you have not done that, it is a good idea to check out that part. At least you need to make sure that the knife you are using is sharp. (8). And if you can't find any such good examples using this code, then you are inventing something. You need a lot of luck in the invention business. So, the above suggestions should be enough for you to get started.