Fluids

Fluids

CFD modelling of Flashing flow iN CD nozzle

    • amkum
      Subscriber

      Hello all..

      I am trying to simulate flashing flow of water in CD nozzle in Fluent using Eulerian with thermal phase change model. However, the solution is not converging. I tried changing the scaling factor, still no improvement. Mesh quality is good.

      I also tried default cavitation model (Schnerr-Sauer/ Zwart-Gerber-Belamri). In this case convergence is good, but outlet pressure changes to saturation pressure, which is strange (Fig is attched). I have tried both with constant vaporization pressure as well as with UDF. How can this model breach outlet boundary condition, which was imposed?

      Thanks in advance!

    • DrAmine
      Ansys Employee

      I will rather assume equilibirum momentum and use Mixuture Model. Which Nozzle is it? (Paper Reference?)

    • amkum
      Subscriber

      Hi Amine...

      Thanks a lot for your reply. We have already completed the simulation with mixture model using Zwart-Gerber-Belamri model for mass transfer. Zwart-Gerber-Belamri model for mass transfer hooked as UDF (same UDF crashing with Eulerian model). 

      Abuaf, N., et al. Study of nonequilibrium flashing of water in a converging-diverging nozzle. Volume 1: experimental. No. NUREG/CR-1864-Vol. 1; BNL-NUREG-51317-Vol. 1. Brookhaven National Lab., Upton, NY (USA), 1981.

      Other reserchers (https://doi.org/10.1016/j.nucengdes.2015.06.015,  https://doi.org/10.1093/jcde/qwaa020 and others) have used Thermal phase change with Eulerian approach for better matching. 

      We want to improve the accuracy. 

       

    • DrAmine
      Ansys Employee

      I ran that Abuaf nozzle with Cavitation model and Mixture Model without any UDF and the results are fine. I even used Real Gas Property Tables for that. I think with Thermal Phase Change the key will be to provide a smart heat resistance.

       

    • amkum
      Subscriber

      Thanks for your reply.

      Are you suggesting using RGP table to give the saturation pressure in the cavitation model?

      RGP Table: Lookup in saturation property tables for mass transfer exceeded the lower range 56197 times. Is this treated as error? To generate saturation table, I use the maximum and min pressure value from the previous study of same case. What is the right method to accurately generate this table?

      For thermal phase change model, I used the ranz-marshall. 

    • amkum
      Subscriber

      Dear Dr. Amine,

      I am simulating BNL309. A screenshot is attached for cavitation model. Please suggest where I am going wrong. When I used UDF with Mixture flow model, the pressure profile was correct. Fluent crashes if I use same UDF with Eulerian model.

      However, without UDF, outlet pressure changes to saturation pressure with Zwart-Gerber-Belamri with both mixture and Eulerian model. 

       

       

    • DrAmine
      Ansys Employee

      The UDF should be checked if it is crashing with Eulerian Model.

      I used RGP tables for Abuaf nozzle and the results were fine. The difference might be that I used expert commands to avoid clipping the pressure as psat. For that reason using UDF for cavitation is also okay. Also bear in mind that Fluent post-processing does limit the pressure to psat when using built-in cavitation model. There is a TUI command to disable that. It is always best to expand the tables with total ranges not only static. The warnings are only printed for static values. I used these limits in my case (again perhaps we are not referring to same case)

      Pmin:   1.000000000000000e+05
      Pmax:   2.500000000000000e+07
      Tmin:   3.000000000000000e+02
      Tmax:   7.000000000000000e+02

       

      I obtained this pressure curve: 

       


    • amkum
      Subscriber

      Dear Dr. Amine, 

      Thank you so much for your reply. Now, I can generate the correct RGP table.

      But I am still struggling with the cavitation model. I tried setting up the pressure limit by using /solve> set limit. Still, the minimum pressure is clipped to the saturation pressure. Can you please share the TUI command to avoid clipping pressure to psat when using built-in cavitation model? . I tried to find TUI command to disable that. However, I could not find it.

       Please help me with this.

      Thank you

    • amkum
      Subscriber

      Dear Dr. Amine, 

      I can resolve the issue which I was facing in the cavitation model. I found the command in Fluent user guide manual. 

      solve/set/multiphase-numerics/heat-mass-transfer/cavitation/min-vapor-pressure

      To see the actual pressure contours, use the following text command:

      solve/set/multiphase-numerics/heat-mass-transfer/cavitation/display_clipped_pressure?

      display clipped pressure in the post-processing [yes]: no

      Can you please give some suggestions for thermal phase model. What should be the reasonable value of From Phase Scaling Factor and To Phase Scaling Factor. By default, these are equal to 1.

      In some research papers, nozzle pressure, volume fraction, liquid and vapor temperature were initialized with initialization function. Please refer to attached

      please suggest. 

    • DrAmine
      Ansys Employee

      Initialization can assist the steady state solver to quicker convergence without as using Real Fluids + Phase Change might be numerically very stiff.

      For the Thermal Phase Change Scalin Factors: they do not exist in Ansys CFX but do exist in Ansys Fluent. I will leave them as default to avoid decoupling the energy balance from mass balance. You can reduce them if you prefer tuning that more than tuning Interfacial Area and HTC alone. I usually do not change that. The referenced paper uses Ansys CFX and so the coefficients are set there to 1.0 You can contact the authors of the paper for more details.

       

      The minimum vapor pressure you can tune is not the one I have in mind (In built in cavitation the material properties are calculated at saturation pressure by default whenever it cavitates -> Incompressible kind of treatment).

Viewing 9 reply threads
  • You must be logged in to reply to this topic.