Fluids

Fluids

Incorrect pressure from fluent

    • DetroitKiddo
      Subscriber

      Hi all,

      I am using fluent to run a simulation which I will explain below which is giving me the incorrect pressure and I am not sure what the issue is with my simulation.

      The simulation consists of a box. There is a inlet hole into the box. Inside the box there is a tube with 2 holes inside and the tube leads outside the box. So the fluid enters the box via a circular inlet and fills up the box and enters the tube via the two holes and flows inside the tube and exits the box. It is a steady state problem, laminar model. Inlet mass flow of 1 kg/s with constant pressure outlet.

      Based on values from our benchtop the pressure inside the box should be around 270 pa however in ansys I am getting 9 pa total with a convergence at e-6. The fluid is water. Based on this I wanted to check the potential issues. My total skew is 0.41 and ortho quality is 0.50. I am using a polyhedral mesh. So from what I understand and could be wrong is since my volume and inlet is fixed the pressure will be solved to give me the exit flow rate. So should I try a different mesh ? smaller element size ? different solver ? I am using a couple 2nd order upwind to solve it with hybrid init.

      I am new to fluent and not sure if the simulation is wrong due to my poor meshing or incorrect BC's or what I can do to check why the pressure is so low.

      Thank you all for any advice/tips!

    • Rob
      Ansys Employee

      The inlet pressure with a velocty inlet is set to get the flow through the domain against the value of the outlet pressure boundary. So, if you have zero on the outlet, inlet pressure will be some Pascals above zero. 

      If the flow is constant density use a velocity boundary. Mass flow is designed for compressible flow. It works, but it's lazy modelling: I'm all for that but not where hand calcs can be used. That also allows you to check the flow rate more easily: if you know the velocity and mass flow and run the model are they correct? That gives you confidence that the model size (scale) is correct and that you've used the correct material. 

      So, check domain scale (Domain tab, Mesh section) and in Cell Zone what fluid are you using? If it's single phase try deleting the air material. 

    • DetroitKiddo
      Subscriber

      Thank you for your reply Rob. I checked the cell zone and it is water. I am using water to simulate the model. The model should be to scale and was imported using the correct units. If I understand your reply correct using the mass flow might make Fluent think the fluid is compressible when its not so it is better to use velocity inlet which might give me the higher pressure I am looking ? Thank you Rob.

       

    • Rob
      Ansys Employee

      No, the solver will model using the density you define. Using a mass flow boundary just adds a couple of extra calculations into the solver: the result will be the same as using a velocity bc. 

      Please check the scale in Fluent. The import functions can scale things due to CAD formats and the like, hence seeing models of a 9km high water tank and 9mm high building at various times over the years. 

    • DetroitKiddo
      Subscriber

      Hello Rob. Thank you again for your help. I did some checking and the original cad was created in MM from Rhino 7. It was imported to spaceclaim in MM and volume extracted. It was meshed in fluent with meshing with changed it to Meters. So based on your point I need to open the scale tool in fluent solver and change the units back to mm since the mesh was in m so its uniform all the way. This could be the issue based on your helpful feedback. Just wanted to double check if this is correct. Thank you for your help Rob.

    • DetroitKiddo
      Subscriber

      Hello Rob I have a quick update on my last message. Originally as mentioned above I was importing a CAD STEP from Rhino 7 into spaceclaim under MM units. This was sent to fluent with  meshing which automatically converted the units from MM to M. Total skewness was 0.41 with a ortho quality of 0.50. This produced a total of 2.8 million elements. This was then imported to solver and run. Now once I created my volume in space claim I saved my geometry file as .sdoc from spaceclaim. I opened fluent with meshing as a individual module and imported the spaceclaim doc which allowed me to manually set my units which I did to MM. Now I ran the mesh with the same sizing just different for units and ended up with a 42 million element mesh. Clearly this is a massive change. I feel I am doing something wrong but my units are definitely now in MM as the CAD and not M as the first simulation. Could this be why I was getting such an error in my pressure ? Perhaps the element sizing was off due to fluent switching my units around ? 2.5 million to 42 million seems like a massive jump. I went with the settings recommended to me by ANSYS in the mesher. If you have any suggestions that would be helpful. Thank you. 

    • Rob
      Ansys Employee

      When you move between codes check the sizes. As you've found the scale can get messed up. My approach is to mesh to whatever size the geometry thinks it is (mm, m, km) and in Fluent scale the mesh. Scaling CAD tends to cause problems with the tolerance controls, whereas the mesh doesn't have that issue. 

      Note, changing the units in Fluent just changes the reporting, you need to scale the mesh. 

Viewing 6 reply threads
  • You must be logged in to reply to this topic.