Fluids

Fluids

simulation of two phase heat and mass transfer inside two phase closed thermosyphon

    • prashant kumar jha
      Subscriber

      i'm simulating heat and mass transfer inside  two phase closed thermosyphon using 2-d axysymmetric geometry.i've selected pressure based solver and transient.I am using ANSYS /FLUENT 19 student version .VOF method with Lee Model is used . Water is used as secondary phase and Vapor is used as primary phase. SIMPLE algorithm scheme for pressure velocity coupling and a first order upwind scheme for the determination of momentum and energy is included in the model. Geo –Reconstruct and PRESTO discretization for the volume fraction and pressure interpolation scheme, respectively, are also performed in these simulations. but,after running the calculation,contour is showing only in the evaporator region as shown in attached figure.I've also patched the liquid water fill ratio with volume fraction=1 in the evaporator region.Can anyone help me out?

    • DrAmine
      Ansys Employee

      First of all: Please avoid generating Entropy by creating several topics regarding the same questions.


      1/2D Axis Symmetric is only for the cases which are axis-symmetric. Can you confirm that from your side?


      2/The contour shows only that some vapor has been generated in the evaporation section. Try to clip the contour to some lower values to detected condensation regions. You can even check the Mass transfer under Phase Interaction  for post-processing.


      3/With the Lee Model you have coefficients (frequency) and those would require a high amount of tuning and this tuning won't make any sense without exp. data or references. Generally the condensation frequency is set to lower value than the one of evaporation but it is case dependent.

    • prashant kumar jha
      Subscriber

      sir.i am simulating the work done by B.Fadhl https://www.sciencedirect.com/science/article/pii/S1359431113004699.i've selected an axisymmetric domain as attached .The line at x=0 is the centerline.i've further splitted the right wall  into three parts namely evaporator,adiabatic and condenser in ANSYS SPACECLAIM 19 student version.Frequency is 0.1 in Lee model as given in the article attached. The problem that i'm facing is  that Contour shows the phase change only in the evaporator section.


       

    • DrAmine
      Ansys Employee

      I have said a couple of times: we do not have access to all journals across the world.  From the abstract: the authors are using an UDF.


      I wanted just to let you check if the case if rotational about the x-axis. If this is not fulfilled either you go to planar 2D or fully 3D. 


      You can clip the contour lines to something smaller to detect the condensation zones. Please check too the phase interaction post-processing variable. 

    • prashant kumar jha
      Subscriber

      it is rotional about the y-axis. Thermosyphon is cylindrical in shape.So,can i go for planar 2D.

    • DrAmine
      Ansys Employee

      No you do not need to do that. You can stick to axis-symmetric but we need to rotate it. So the "Y-direction" of the model will be your X-Axis.  You can then take only the half of the diameter and model the section to be aligned with the X-Axis.

    • Karthik R
      Administrator

       Hello Prashant,


      Just to add to Amine's comments:



      • Please read the paper and make sure you are using the same set of source terms as the authors. If I am to speculate, they are using sources terms responsible of evaporation and condensation in the mass and energy equations.

      • In addition to this, please make sure you are using consistent initial conditions, similar to what the authors have done.

      • Another important factor is the mesh. If there is a way to verify the mesh, please make sure you are using the same resolution of mesh.


      Replicating the results from a paper is always a challenging task. The most important suggestion I would like to give you is this - when in doubt about the exact steps used by the authors, you can always contact the authors of the paper via email. This is important as they might be able to clarify the exact model for you. Please let us know and we might be able to help you with questions regarding best usage and practices of the modeling tools.


      Thank you.


      Best Regards,


      Karthik


       

    • prashant kumar jha
      Subscriber

      So,is there any problem having vertical direction aligned with y-axis not with x-axis?

    • DrAmine
      Ansys Employee

      In Fluent if you want to use the 2D Axis-symmetric solver then the symmetry/ rotational axis needs to be the x-axis. You can then just apply gravity in the -X direction. 


       

    • prashant kumar jha
      Subscriber

      sir, i that case,what will be the sign of g.Will it be g=9.8m/s^2?

    • DrAmine
      Ansys Employee

      Example if the height goes from 0.0 to +HMAX in the positive x-direction. Then g will be -9.81 m/s^2.

    • prashant kumar jha
      Subscriber

      in ANSYS FLUENT19.0.Is is important to add source term UDF,if i have already selected the phase interaction as Lee model for evaporation-condensation in Fluent.Since,there is no momentum source term in the VOF model,when selected Lee model,will Ansys Fluent will take care of the source terms?


    • DrAmine
      Ansys Employee

      All the stuff described in the paper is already implemented in FLUENT. You do not need to add any UDF. All required secondary source terms are accounted for.

    • prashant kumar jha
      Subscriber

      for the energy source term,since latent heat of vaporization is 2455 kj/kg.I've incorporated the following udf for energy source term.Please,check it sir.


      #include "udf.h"
      #include "sg_mphase.h"
      #define T_SAT 373
      #define LAT_HT 2455.1345e3


      DEFINE_SOURCE(enrg_src, cell, mix_th, dS, eqn)
      {
      Thread *pri_th, *sec_th;
      real m_dot;
      pri_th = THREAD_SUB_THREAD(mix_th, 0);
      sec_th = THREAD_SUB_THREAD(mix_th, 1);
      if(C_T(cell, mix_th)>=T_SAT)
      {
      m_dot = -0.1*C_VOF(cell, sec_th)*C_R(cell, sec_th)*fabs(C_T(cell, sec_th) - T_SAT)/T_SAT;
      dS[eqn] = -0.1*C_VOF(cell, sec_th)*C_R(cell, sec_th)*LAT_HT/T_SAT;
      }

      if(C_T(cell, mix_th)<=T_SAT)
      {
      m_dot = 0.1*C_VOF(cell, pri_th)*C_R(cell, pri_th)*fabs(T_SAT-C_T(cell,pri_th))/T_SAT;
      dS[eqn] = 0.1*C_VOF(cell, pri_th)*C_R(cell, pri_th)*LAT_HT/T_SAT;
      }
      return LAT_HT*m_dot; }

    • prashant kumar jha
      Subscriber

      since latent heat of vaporisation is 42455kj/kg.I'm Using the attached udf for energy source

    • prashant kumar jha
      Subscriber

      sorry,latent heat=2455kj/kg

    • DrAmine
      Ansys Employee

      You do not need require any UDF. Furthermore if UDF then I will go for DEFINE_LINEARIZED_MASS_TRANSFER.

    • prashant kumar jha
      Subscriber

      sir, in paper they have considered latent heat of vaporisation as 2455kj/kg,how to incorporate that value in ANSYS FLUENT 19 STUDENT VERSION.

    • DrAmine
      Ansys Employee

      You use your working boundary conditions. Regarding mass transfer just use the implemented model and just provide the saturation temperature and change the frequencies.


      Latent Heat in Fluent is in J/kgmol and you have the molar weight so you can incoprarte theier setting in Fluent.


       


      You need to carry out some tutorials in order to get really started...


       

    • prashant kumar jha
      Subscriber

      sir,i've followed all your instructions ,simulation is started in 2-D domain but the coontinuity equation is diverging and the error appears.I've attached two pics


       

    • DrAmine
      Ansys Employee

      Same as usual: smaller time steps, lower URF's.. etc.


      If it does not help then perhaps other community members might help you because as ANSYS Stuff I am not allowed to give more that I have already done.

    • prashant kumar jha
      Subscriber

      thank you abenhadj and Kremella ,your guidance was very much helpful.Further,i need to refer some tutorials regarding the issue.

    • DrAmine
      Ansys Employee
      Or just take the molar weight in kg/kmol and multiply by the standard state enthalpy in j/kg.
    • giovanni_ianziti
      Subscriber

      Hello, I`m working on the same model. Can we discuss about it if you want.


      Thank you

    • pritam1688
      Subscriber

      Can you share the case file if it's working well for you? I also want to know whether UDF is required or not?

Viewing 24 reply threads
  • You must be logged in to reply to this topic.