[A3] 2D Airfoil Aerodynamic Analysis With Fluent & Gambit
By Jerry He Dec 10, 2009
A little expiation before this article starts: I decided to write it in English because I want to use English more. If that reason is not enough, then I have another one: for trying to improve my readers’ English.
I think I should finish this article before starting my journey. Even though I haven’t solved all problems I’ve come across, but as time is limited, I will write down what I’ve got till now.
This is part of our teamwork of “Aircraft Design”, a course by our class advisor. Assume you have learned aerodynamics and fluid dynamics.
This article will divide into seven sections:
1、 Introduction
2、 Installing required software.
3、 Create geometry and mesh with Gambit
4、 Analysis using Fluent
5、 Results analysis and re-mesh
6、 More
7、 Acknowledgements
1、 Introduction
As a student from department of aerospace, this is the first time it seems that I really do. The problem we are trying to solve is:
We have a certain airfoil (given by a .dat file, which is ASCII text with (x, y) points list), we want to know the lift coefficient and drag coefficient in a certain attack angle and certain Mach number.
At this article, we will use SC (2)-0714 for example and assume the attack angle is 2 degree, the Mach number is 0.715. Fig 1 shows our airfoil.
Fig 1 SC(2)-0714
2、 Installing required software
Here we use Fluent to analysis and Gambit to create geometry and mesh. So before starting your work, get this three software’s installing program. They are Fluent, Gambit, and Exceed. Here I use Fluent 6.2.16, Gambit 2.2, and Exceed8.0.
Install Exceed first, then Gambit, then Fluent. After finished the basic installation, copy the crack\license.dat to the corresponding license folder (..\license\).
Finally, try if you can start Gambit and Fluent. If not, something must be wrong.
3、 Create Geometry and Mesh With Gambit
I won’t introduce this step by step, because there already have one detailed document teaching this. You can view it from:
http://instruct1.cit.cornell.edu/courses/fluent/airfoil/index.htm
For a fresher to Gambit, tricks below will help you to operate it (The gambit I use have no help files :<).
Left Mouse: Rotate
Middle Mouse: Move
Right Mouse: Zoom in/out
Menu: Edit-->Undo& Redo (you cannot use Ctrl+ Z to do that)
After this step, my resultant mesh is shown in Fig 2 and Fig 3.
Fig 2 Resultant Mesh
Fig 3 Result Mesh(Detailed view)
4、 Analysis using Fluent
You can also learn basic procedure with the detailed document I mentioned in the prior section. And I assume you’ve read it. I will only tell you the things that it didn’t introduce. So make sure you’ve read carefully about it before continue.
Define-->Models-->Solver: In the Solver box, select Coupled. For our problem using Segregated won’t have enough precision.
Define-->Model-->Energy: Make sure you checked the Energy Equation.
Define-->Model-->Viscous Model: Select the Spalart-Allmaras in the Model box. Because in our case, the Mach number is 0.715, which is really big, we cannot use inviscid or laminar flow. Take the default value for Model Constants options.
Define-->Materials: Make sure selected fluid in Material Type box. In properties box, select ideal-gas with Density options, and select Sutherland with Viscosity options. Remember always use the default value if I don’t mention it.
Define-->Operating Conditions: the Operating Pressure is your designed pressure. Here our plane flies at 11KM high, so the pressure is 22612Pa.
Define-->Boundary Conditions: For airfoil surface, select wall. For the front and side boundaries of your mesh, select pressure-far-field. And in the Pressure Far-Field options, input the Mach number with your Mach number, for me, it’s 0.715. The Temperature should be your designed temperature, for me, it’s 216.66. The X-Component of Flow Direction should be c*cos(attack angle), that’s 1*cos(2degree)= 0.99939, the Y-Component of Flow Direction should be c*sin(attack angle), that’s 1*sin(2degree)= 0.0348995. (So I think when create geometry, keep your c to 1 will save you a lot of work.) For the back boundary, select Pressure-outlet.
Solve-->Controls-->Solution Controls: There are two important settings: Courant Number and Modified Turbulent Viscosity. As William told me, the bigger these two numbers, the faster the simulation will run. But if you set it to too big, the result may disperse. So the suitable range for Courant Number is 1 to 20, while for Modified Turbulent Viscosity is 0~1.0. Here I used the default value.
5、 Results analysis
After the results converged, our Cl=1.0151,Cd=0.0373,L/D=27.21,Cm=0.4122
Plot-->XYPlot, select Pressure, then select Pressure coefficient, as Fig 4 shows:
Fig 4 Pressure Coefficient of Airfoil
Display-->Vectors, use the default options, click plot, as Fig 5 shows
Fig 5 Velocity Distribution
Display-->Contours: select Velocity, and then select Mach number. Click Plot, as Fig 6 shows. Of course you can display anything you are cared of.
Fig 6 Contours of Mach Number
6、 More
As the Fig 3 shows, the pressure coefficient of the up-surface of airfoil has some quiers. From Fig 5 we can see the flow in the top of airfoil is unstable. From Fig 6, we can find the velocity is not well-proportioned increasing from airfoil surface. All those just tell us one thing: our mesh near the airfoil surface is too sparse.
So of course I tried to solve this. I tried to re-mesh wanting the mesh near airfoil surface to be denser. But till now, I failed. It’s more difficult than I expected? May I’m more foolish than I thought. Every time I can let Gambit crash(maybe that’s a bug of Gambit.)
But at least I comfirmed one thing more. “Before you start doing it, never think it is too easy.” It always need more practices and need to know many details before you become the one can solve problems correct and quickly.
I named this article with a 1, I hope one day I will write a 2( this day may not come so fast, because I will have something more important to do in the next serveral weeks). In this article, I told you how to do, but in most case, I didn’t tell you why. And I don’t know. I’m really interested in how the CFD software worked. In my opionion, either CFD simulation or our laboratory’s multiple-body dynamic simulation, all FEA(finite element analysis) software, in the very base of the software, they do the same thing: They solve a matrix equation like Ax=b. The difference between them is how they make up the A and b with the basic theory of related field.
7、 Acknowledgements
I’ve spent more time than I expected doing this and I did something that proved stupid afterward. But here I’d like to thank the people help me a lot with that. They are Xiaozhi Xiang, William Wang, Xiaofeng Yang, Caijin Yang, Tian Ju, Qihong Ren.
P.S. I haven’t upload all figures, you can email to me getting a pdf.
P.S.2 My English is not good, so if you find any ‘bug’, simply point it out! If you cannot understand it, let me know.