Fluids

Fluids

water flow, nozzle

    • Lance
      Subscriber

      Dear all,


       


      I have a question about modeling the water flow through a nozzle. When I run the calculation I get the following error: Floating point exeption. Is that because my mesh is not fine enhough and/or are my selected models not right.


      Thank you in advance!


      Mesh:


      502980 elements


      orthogonal quality average: 0.81 


      skewness average: 0.17


      skewness max: 0.97


      inlet conditions:


      mass flow: 0.36 m^3/s,


      Pressure: 127530Pa


      velocity: 3.75 m/s.


       


      Model:


      solver: pressure based, transient flow, gravity added


       


      Model SST, aimed for y+ >30 --> first cell height 0.32mm


      Cell zone conditions: water-liquid


      Boundary conditions: 


      inlet-velocity, 3.75m/s , operating pressure: 127530Pa


      outlet-pressure


       


    • raul.raghav
      Subscriber

      Lance, the mesh with a max skewness of 0.97 is a red flag for sure. Fluent recommends the skewness to be less than 0.95.


      Ansys Fluent Mesh Quality


      What were the initial conditions for the flow?


      Before you go transient, try to sort out a steady state simulation. And why do you to include gravity in your simulation?

    • Lance
      Subscriber

      Hi Rahul, 


      Thanks for your help!


      1. skewness: okay than I have to refine my mesh a bit more, I thought an average of 0.17 would be sufficient enough.


      2. Inital conditions: The nozzle is placed on a pump which has a volume flow rate of 0.36 m^3/s, so I transfered this to a velocity of 3,75 m/s according the continuity equation Q=V*A. The pump has a head pressure of 127530Pa. So thats what I inserted as intial conditions in the inlet. 


      3. I have included gravity, since the flow needs to be pushed up in the nozzle? or can this be excluded from the model?


      4. Could you explain the steady state simulation a bit more? Do you mean that I have to calculate the exit velocity and insert this in the model, run the simulation in steady state, when converged go transient?


       


      Lance 

    • raul.raghav
      Subscriber

      Lance,


      1. The mesh is "almost always" the issue. A better mesh wouldn't hurt after all . See if you can fix the max skewness elements.


      2. Initial conditions. Forget about the pressure at the moment. Just initialize with the inlet velocity. And the for the outlet BC, I assume the nozzle is open to the atmosphere. Curious, did you set the static pressure at the outlet to 0 Pa?


      3. Again ignore gravity. You'll later see that gravity doesn't have a huge influence on this simulation. I can explain more if you have questions regarding gravity.


      4. What are you trying to analyze by this CFD simulation? What are the outcomes you are looking from doing this? This will help me answer your question in more detail.

    • Lance
      Subscriber

      Hi Rahul,


      Thanks again for your reply!


      1. I am almost on the limit of 512000 elements. The bad skewness is probably a cause from the inflation at the wall. Do you  think I could use a mesh without inflation? Than the skewness is much better. And then maybe use a k-e model with scalable wall functions or k-w?


      2. I will use steady state as you suggested and initialize the velocity inlet. Yes, the outlet is open to the atmosphere and did set the static pressure to 0 Pa.


      3. I would like to optimize the nozzle on efficiency/pressure losses. And to validate the output of the nozzle, will it produce the right water film with the required characteristics, like velocity etc. 


      Looking forward to hear your opinion about the model.


      Lance 

    • Lance
      Subscriber

      Hi,


      I have simulated steady state, velocity inlet 3.75 m/s, pressure outlet 0 gauge pressure. Used two different meshes 340k & 510k both with a skewness average of 0.2 and max 0.78, orthogonal quality average of 0.78 and min. 0.26. I have used different models, the k-e standard, k-e rializable with scalable wall functions, k-w sst. Simple scheme with first order upwind and second order upwind. But in all simulations the solution wont converge, where the k-e rializable is very close to converge. When monitoring the mass flow, velocity and yplus, they are quite constant.  Is this due to the number of elements?


       


      I hope you can provide me a solution to the convergence problem and thanks for your time!


       


      Lance 


       


       


Viewing 5 reply threads
  • You must be logged in to reply to this topic.