[Keskin]

Hello everybody,

I am a new member in the forum. In my graduate years I solved some heat transfer and flow problems but I had never tried a compressible flow problem. I am trying to solve a problem from Hoffman's CFD book. (Vol.2, p. 201)

The book suggested a modified Runge-Kutta but it didn't work while I was trying. I will start from the beginning. What I request from you:

Occasionally I will upload my progress. It will include my code and a presentation showing what portion of the code is doing what. Could you please check and give me advices?

I have got obsessed with the problem and my main problems are:

*The problems I have solved in the past had boundary conditions and initial conditions clearly stated. It seems different when it comes to compressible flow.

*I am not much familiar with the computational space. Therefore I cannot easily see my mistakes.

As there are many things hard to understand about the subject I think doing it with help from experienced people will contribute to me a lot.

The first files are attached. It includes a presentation and the grid generator MATLAB file.

The presantation Shows:
P1: The cover page
P2: Shows the dimensions of the problem (The red ones are the ones I use and the purple ones are the ones in the book). Also the parts declaring the dimensions are being shown.
P3: Shows the boundary conditions for the elliptic PDE.
P4: Shows the code piece solving the PDE with Gauss-Seidel and the generated mesh can be seen.
P5: Shows the code piece calculating Jacobians and metrics.

What I need for this first attempt is not you to inspect everyline of the code. You can run the MATLAB code and check if it is OK to use these grids. Do you see any problems? Also it takes too much to generate the grids. Is it normal?

[Keskin]

Hello again,

I have decided to go with a simplier geometry and less node numbers. Therefore I can check the calculations by hand.

The geometry is below and the flow enters the channel from the left side:



The inlet boundary conditions are Ux=1, Uy=0, density=1 and pressure=0.20 as dimensionless numbers. The inlet Mach number is about 2.

My questions are:

*As the inlet flow is supersonic and the crossection area decreases can I assume the outlet flow will be supersonic too?

*I have specified all the inlet boundary conditions. As I will solve inviscid 2D-Euler Equations do I need to specify the boundary conditions for one of the upper and lower surfaces? As the problem is a first order 2D it seems to me I cannot make both of them extrapolated.

[LUQILIN]

Originally Posted by Keskin

Hello again,

I have decided to go with a simplier geometry and less node numbers. Therefore I can check the calculations by hand.

The geometry is below and the flow enters the channel from the left side:



The inlet boundary conditions are Ux=1, Uy=0, density=1 and pressure=0.20 as dimensionless numbers. The inlet Mach number is about 2.

My questions are:

*As the inlet flow is supersonic and the crossection area decreases can I assume the outlet flow will be supersonic too?

*I have specified all the inlet boundary conditions. As I will solve inviscid 2D-Euler Equations do I need to specify the boundary conditions for one of the upper and lower surfaces? As the problem is a first order 2D it seems to me I cannot make both of them extrapolated.

Here are some test cases you need.
http://turbmodels.larc.nasa.gov/index.html

[thibor]

Hello everyone!

I need help! I am not expert in matlab.

I have a similar problem to find the solution, please see enclosed.

How can I change the geometry?

Do you have some similar solution or a matlab code to share with me?

Tks,

Regards,

Thibor

[Keskin]

Originally Posted by thibor

Hello everyone!

I need help! I am not expert in matlab.

I have a similar problem to find the solution, please see enclosed.

How can I change the geometry?

Do you have some similar solution or a matlab code to share with me?

Tks,

Regards,

Thibor

Hello,

Is it mandatory to use the specified methods in the document to solve this problem? I was trying different methods when I asked this question. For example I can generate grids for your example but it won't meet the requirements stated in the document as my code use elliptic grid generation.

By the way what book do you use for this lecture? The PDF mentions Chapter 12. I may have a look if I can find it.