Fluids

Fluids

convergence issue for combustion application with complicated chemistry

    • Weiqiang Liu
      Subscriber

      Hi all,


      I am doing a methane combustion simulation with fluent. Gas phase and surface mechanisms are included in the case. Surface reaction rates are converted into volume rates through surface area per volume parameter.


      My propblem is the case is never converged. The flame position moves forwards and downwards and the maximum temperature varies with 100K magnitude. Also the monitored variables keep changing no matter how many iterations are set. I put the temperature contour and area averaged temperature of outlet below:




      I tried a lot of methods to aid convergence. For example, to get a cold flow first and then enable reactions recommended by fluent  or keep the leading edge of channel inert. However, nothing worked. Can anybody give me some suggestions on how to get converged results for the my case. 


      Or it's the problem of mechanisms I use. In other words, I can never get converged results with this mechanism?


      Best regards


      Weiqiang


       

    • RK
      Ansys Employee

      Hi Weiqiang, 


      May I know the boundary conditions and the solution method you using?

    • Weiqiang Liu
      Subscriber


      Hi, Rahul,


      I put the solution methods above. SIMPLE method is used for velocity and pressure coupling. Second order upwind is used for spatial discretization.


      As for boundary conditions, very normal boundary conditions are used. specificly, velocity inlet, pressure outlet, wall boudnary conditions and coupled walls for fluid and solid interface.


      Best regards


      weiqiang 

    • Rob
      Ansys Employee

      Use the Execute Commands to save contour plots every some time steps. It's possible the monitor is in a region where the flame is moving so the monitor looks worse that it actually is. 

    • Weiqiang Liu
      Subscriber

      Hi, 


      my case is steady-state. Do you mean my case itself is unsteady?


      Best regards


      Weiqiang

    • DrAmine
      Ansys Employee
      Looks like intermittent transient behavior. Assuming correct mechanisms worth switching to transient solver. We do warranty for the mechanisms.

      For current setup please use coupled pseudo transient plus node based gradients.
    • Weiqiang Liu
      Subscriber

      Hi Amine,


      do you mean I should use coupled pseudo transient plus node based gradients with steady state solver?


      Best regards


      Weiqiang

    • DrAmine
      Ansys Employee
      yes as best setting for running steady. If still oscillating then unsteady. As I said assuming you kinetic is correct which we do not know
    • Weiqiang Liu
      Subscriber

      Hi Amine,


      Thanks for your suggestions. I would try this. By the way, the kinetic I use is widely used in published papers. I think my kinetic is correct.


      Best regards


      weiqiang

    • Maruko_u
      Subscriber
      Hi weiqiang,nI have the same problem as you do, I wonder how you finally solved your problem.nBest regardsnMarukon
Viewing 9 reply threads
  • You must be logged in to reply to this topic.