Problems w/ Simulating Rocket Engine Exhuast

  • Last Post 08 May 2020
edgue1997 posted this 18 April 2020


My name is Edwin Guerrero and I am an aerospace engineering student. I am having difficulty with simulating my rocket nozzle, particularly, the residual plots for my simulation will not converge, as in, they will not plummet down to below my convergence criteria. Also, it may be related to this problem, my velocity contours look very strange? The velocity seems to be increasing again when approaching the outlet. Below are photos of my simulation and residual plots. 

Residual Plot

Mach Number Contour

I have simulated other similar rocket engine nozzles and they perfectly expand the exhaust gases to be axially. Below, are some key parameters of my simulation:


  • Type: Density-based
  • 2D Space: Planar


  • Energy: On
  • Viscous Model: k-epsilon; realizable


  • Fluid: (solid-fuel-one)
    • Density: ideal gas   
    • Cp: 1881.15 j/kg-k
    • Thermal Conductivity: 0.0242 w/m-k
    • Viscosity: sutherland
    • Molecular Weight: 23.77 kg/kmol

Boundary Conditions:

  • Inlet: pressure-inlet
    • Gauge Total Pressure: 3447378.64 Pa
    • Supersonic/Inital Gauge Pressure: 3447368.64 Pa
    • Total Temp: 2603.15 k
  • Outlet: pressure-outlet 
    • Gauge Pressure: 10 Pa
    • Backflow Total Temp: 288 k 
  • control volume walls & nozzle walls: wall

Reference Values

  • Density (kg/m^3): 1667.73
  • Pressure: pressure-inlet gauge total pressure
  • Temp: pressure-inlet total temp
  • Ratio of Specific Heats: 1.2284

 My ansys workbench file is attached to this post.

Attached Files

Order By: Standard | Newest | Votes
killian153 posted this 21 April 2020


The problem comes from your CFL number set to 5. You must reduce it.

I ran your simulation with AUSM and CFL=1 (and transformed your farfield into a simple wall) and here's what I get (I stopped it at 5200 iterations so it's not finished yet):

Best regards.

  • Liked by
  • rahkumar
edgue1997 posted this 21 April 2020

Hello killian153,


Thank you for you response. I just got a few questions.


Why did you did you edit the 'Flux Type'? What give you that indication that it should be altered? Also, for 'AUSM', did you edit any of default settings within 'Spatial Discretization'? 


Edwin G.

killian153 posted this 08 May 2020

Sorry, I wasn't notified of your reply!


I usually use AUSM for this kind of application, as it accurately captures shocks. It is a well suited model for the simulation of nozzles.

In the Spatial Discretization, I selected 2nd order. It more accurate than 1st order (but harder to converge).