Isolated Rotor Single Passage

  • 145 Views
  • Last Post 25 September 2018
Mohamed Abu-bakr posted this 26 July 2018

Currently, I am working on an unsteady analysis of a single passage of isolated rotor (no stator) of an axial compressor using cfx. 
I can't decide which to use Transient or Transient blade row model.
And if I should use transient blade row model, can I use time transformation in this case and if I can, how? 
Note my domain is 1 element including inlet blade passage and outlet, so there is no interface between them. The only interface exist are tip gap and periodic surface interface. 
Thank in advance.

Order By: Standard | Newest | Votes
MUSohail posted this 24 September 2018

I am trying to simulate Transient Simulation of Transonic Axial Compressor Rotor 67. I have already done Steady state simulation. But when I run transient simulation with time steps 1 sec and time steps 0.001 sec or 0.1 sec I get error “The ANSYS CFX solver exited with return code 1. No results file has been created”

i want to run unsteady simulation of rotor 67 from 0rpm to designed rpm

Kindly help me out

Bill Holmes posted this 25 September 2018

@Mohamed

Transient blade row is only used when you have blade flutter, inlet disturbance, or rotor/stator stage.  If you do not have this situation, then you should use Transient as the model.  For examples where Transient Blade Row should be used, see Tutorials 32-38 in the CFX Tutorial guide, found at this link:

https://ansyshelp.ansys.com/account/secured?returnurl=/Views/Secured/corp/v192/cfx_tutr/cfx_tutr.html

 

@MuSohail,

In this case the solver is failing. Likely due to timescale, incorrect boundary conditions, or a poor initial guess.  

Your timestep for a transient run should be a fraction of the blade passing period.  For rotor 67, I think the rotational speed is something like 16000 rpm and it has 22 blades.  So the blade passing period is:

16000 [rev/min]  means one revolution takes 0.0375 [s].

The blade passing period would then be 0.0375 [s] /22 = 0.000170455 [s]

So for your transient run you should probably have a timestep smaller than 0.00017 [s], maybe 1.7e-5[s].

Also keep in mind that at 0 rpm the outlet conditions will be much different than at 16000 rpm, so you will need to have the exit conditions a function of rotation speed.

You can also imagine that it will take many timesteps at 1.7e-5 intervals to ramp up to full speed from zero rpm.  It might be more practical to run multiple steady state simulations at different speeds to obtain the same sort of information.

You mention you are starting from a steady state simulation.  You must ensure that is at the same operation point that is definted at t=0 in your transient simulation.

Hope this helps,

Bill

 

 

Close