3 Replies Latest reply on Oct 17, 2010 2:45 PM by 1-TZBQ32

    Flow simulation and computational domain problem

    1-TZBQ32 Oct 15, 2010 7:13 PM

      I'm running a simulation of a land vehicle, and I'v added a ground plane but I have found some problems while adjusting the computational domain settings, here are my basic settings for the Solidworks 2010 flow simulation:


      -The vehicle is set on a big solid rectagle, which is the 'ground'.

      -The computation domain is set to .001 m below the  surface of the 'ground'.

      -The top surface of the 'ground' has the boundary conditions of "real wall" and is set to a velocity that the air is moving. 
      -The 'ground' settings are: 20 m/s along the Z axis, Surface roughness of 50 Micrometers

       

      the study is  set to an external study, excluding internal volumes.


      -The goal for the study is Force in vehicles moving direction axis Z

       

      Everything else is pretty much clear, but the problem comes with the computational domain and it's position.

      Results vary depending on how much Computational domain is set below the ground, and if its set to 0 or above the ground, computing is not possible due to the missing boundary wall conditions.

      Also when inserting a cut plot, its clearly visible that the front area of the ground also interferes with the flow, velocity is decreased in the very front area, why is that?

      Can you explain what should be done to make ground act as a real ground without messing the results ?

      Below is the picture of the simulation, vehicle is removed from the picture.

       

       

      Boundary.JPG

      • 3528 Views
      • Categories: Flow Simulation, Simulation
      • Tags:
        • Re: Flow simulation and computational domain problem
          Anthony Botting
          Anthony Botting Oct 16, 2010 12:57 PM (in response to 1-TZBQ32)

          Hi Gunn:
          I got this to work with a tire on a road surface. Like you, I buried the Comp. Domain into the road, and set the road surface to "real wall", then turned on "Wall motion" with a velocity value input to the surface (you may have to put in a negative sign to get the correct direction). It worked OK. Did you already try that? I did a presentation on it (some shots included). I'm not sure what to do about the stagnation point you see at the front of the road, but I assume it may disappear with the wall motion BC. Let me know if that works. We can all figure it out together. Note that I did it in 2D initially to observe the results with a very fast turn-around (specify 2D Plane flow in Comp. Domain settings)- and once I got the 2D running satisfactorily, I moved on to 3D and everything appeared to work well.  - Tony

            • Re: Flow simulation and computational domain problem
              1-TZBQ32 Oct 17, 2010 2:45 PM (in response to Anthony Botting)

              Hi Tony, thanks for the quick reply ! Since my post, I'v spent a lot of time in front of my pc, but after trying things out, I think I got it, you were right, my wall motion was set to opposite direction, the wall coordinate system was opposite than my global system, and that confused me, should have checked that in the first place. But the front end stagantion point you mention did not go away with this change, but when I changed my comp. domain box bigger than the actual ground plane, it disappeared. Great! Big thanks!

               


              Now that the basic things should be in order, I got the results of some goals I'v set, global goal force in Z-component which is the air flow direction, and I got ridicilously small force of 5 Newtons, and from that force we can calculate the wattage needed. 5 newtons mean that it only requires 100Watts of power to go 20 m/s (72km/h) speed. How can it be...? I know this vehicle is very streamline and its narrow and low, aprx. 2.5 meters long, 75 cm high and 50 cm wide, but that small reading just amazes me, am i hitting a goldmine of low drag or is it a miscalculation . I dont have much practical information how much the force and wattage should be, all I can do is to trust these rreadings? hmmm.. am I really doing all correctly, readings just show really small forces involved. My surface roughness was set to 0, practical value could be very close to that since surface is highly smooth and polished and waxed.How much 1-5 microns would affect, should be tested.

               

              The other amazement is the force directional to Y direction which is from ground directly upwards. 11 Newtons, over two times more than forward to wind direction force Z. This would mean that the vehicle has more lift than is required to go forward. When looking at the vehicle and its shape, it just sounds odd.

               

              Tyres I'm using (not yet simulated) should not increase wattage too much (high pressure, narrow) and they are hidden inside the body and it was interesting to see that the pics you sent were about tyres, fascinating! Is it possible to simulate rotational force needed to turn the wheel? Tyres used in this vehicle are slick-tyres with no pattern of anykind, so only meaningfull is the surface roughness, what surface roughness did you use on that model of your tyre simulation ? Did the simulation correspond to the real thing ? Highly appreciated of any information you can provide me.

               


              -Gunn

            • Re: Flow simulation and computational domain problem
              Bill McEachern
              Bill McEachern Oct 17, 2010 9:20 AM (in response to 1-TZBQ32)

              Issue could be "face coordinate system" Maybe you should try global coordinate system.