CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   transonic compressor Convrgce pb with transient .. (https://www.cfd-online.com/Forums/cfx/26254-transonic-compressor-convrgce-pb-transient.html)

 Noureddine August 13, 2008 15:23

transonic compressor Convrgce pb with transient ..

Hi,

I would excute a transient simulation of an isolated transonic Rotor 37 using ANSYS CFX 11 with rotation along one pitch, i did not have any problem to get results for steady fluid flow but i could not converge with transient any way.

However, I noticed that the value of pitches displayed by the solver doesnt correspond to that i specified.

my setup is as fellow :

- Rotor 37 design speed = 17188.7 rpm = 1800 [rad/s]

- 36 blades ===> 1 pitch = 2*pi/36 = 2*3.1416/36 = 0.1745333 [rad]

- total time to rotate along one pitch = 0.1745333/1800 = 9.7e-5 [s]

- if i choose 100 time steps ===> DT = 9.7e-5/100 = 9.7e-7 [s].

I should see in each time step an increase of pitches by 0.01, but it is not the case !!!!!!

what it can be the problem ?????

Regards,

 sfallah November 10, 2014 11:58

Dear Noureddine
I can not simulate steady case of rotor NASA 37, all my simulation leads to overflow. Can you give me your boundary condition containing: total inlet pressure, mass flow rate or static outlet pressure for steady simulation of one blade(with periodic condition) of NASA 37 in CFX??? I guess that overflow arisen from fault boundary condition. :confused:

 ghorrocks November 10, 2014 17:05

 sfallah November 11, 2014 03:03

Quote:
 Originally Posted by ghorrocks (Post 518407) Have you read the FAQ on overflow error? http://www.cfd-online.com/Wiki/Ansys...do_about_it.3F
Yes, I read it but my sticky problem is not solved. My operating condition is selected according to suggestion of AIAA paper:"Fully Coupled Fluid-Structural Interaction of a Transonic Rotor at Near-Stall Conditions Using Detached Eddy Simulation":
Total inlet pressure=17.7 (psi)
Outlet Mass flow rate= 20.19(kg/s)
Total inlet temp=519 (R)
My grid contain about 600000 element for on blade passage, physical time step in steady simulation assumed 0.0001s,
I used Geometry of NASA 37 which exist in turbogrid tutorial. I test it in different inlet and outlet domain length(by extending original geometry in Bladegen)
In solution procedure, first, Mach Number increased gradually, then Notice:"a wall hase been placed at portion of an outlet..." appears in monitor screen and finally :Overflow!!!!!!

 ghorrocks November 11, 2014 17:55

Well, that's your problem. If you are running near stall conditions you are unlikely to have a steady state solution. You will probably need to run it transient.

 sfallah November 13, 2014 07:34

Quote:
 Originally Posted by ghorrocks (Post 518612) Well, that's your problem. If you are running near stall conditions you are unlikely to have a steady state solution. You will probably need to run it transient.
ghorrocks
Thank you ghorrocks
Very useful comment.
1 technical question: Does the length of the computational domain in inlet and outlet is important in turbomachinery?
Simulation of Original geometry of NASA67 as exist in turbogrid tutorial(Small inlet and outlet domain) leads to smooth but low slop convergence curve, in the other hand, using geometry with extended inlet and outlet length leads to steep and oscillatory convergence curve. Which of them is correct and optimum?????

 ghorrocks November 13, 2014 18:28

Quote:
 Does the length of the computational domain in inlet and outlet is important in turbomachinery?
Yes, it does. This is one of the normal things to check with a sensitivity analysis. A longer domain is more accurate, but will result in a larger model. If the smaller model is converging better then I would be suspicious it is artificially damping the result causing inaccuracy.

 sfallah December 23, 2014 12:48

Outflow boundary condition NASA37

Dear All
What is the best outlet boundary condition for transonic(subsonic inlet and outlet but transonic passage) compressor and in general transonic turbomachines? why?
I would like to have specified inlet mass flow rate. I use total pressure(because of more stable and better convergence behavior than inlet mass flow rate) at inlet but by applying static pressure at outlet, desired mass flow rate is not be obtained.

My case is Nasa 37 rotor which outlet length is short. using k-omega sst and steady-state option, I have not converged results but using k-epsilon convergence attainment is easy. I guess that its reason is static outlet boundary condition which forced at non-uniform flow location (outlet).