6 Replies Latest reply on Apr 5, 2013 11:19 AM by Jared Conway

    Finer mesh gives the worse results

    Choon-Sik Jhun

      Dear Group,


      Does anyone have the problem like this, i.e., given the same model, a finer mesh gives the worse results.

      We're simulating a pump and the rotor is imbedded in the rotating region. We have two inlets and outlets, and expect to see the balanced (i.e., equal) amount of flow from both outlets. Total volume flow coming in through the inlets is 4 LPM, and the outlet flows are expected to be 2 LPM from each outlet. With a 50,000 elelements we get a fairly even flow like ~2.2 LPM from each outlet, but with a 0.3 million elements, we get ~2.5 LPM (which is farther off from 2 LPM). Any advice will be greatly appreciated.



        • Re: Finer mesh gives the worse results
          Mark Kaiser

          I'm confused.  If you have 4 total of something coming in, and 2.2 x 2 = 4.4 coming out, it looks like your violating some basic rules somewhere.  Am I not interpreting something correctly?  I'm not a flow expert, but this doesn't sound right to me.

          • Re: Finer mesh gives the worse results
            Jared Conway

            Mark's right, something doesn't sound right.


            2 inlets at 4LPM

            2 outlets, 2.2 LPM each, total of 4.4 LPM total


            first suggestions:

            1. what does your mass flow rate tell you? (remember we conserve mass, not flow rate)

            2. how are you measuring this? goals or surface parameters? please create goals for each surface and then goals for the inlets together and outlets together. and if you want to get fancy, setup some equation goals to check everything is working the way you expect it.

            3. check convergence. how did your solution end. did the results converge? post the convergence plots for us to take a look at.


            if you haven't done it already. i'd suggest running without rotation just to see what happens. also can you describe your setup. outputting the report would be a good first step if you can't upload the files.

            • Re: Finer mesh gives the worse results
              Radoslaw Koneczny

              1. Did you saw "Vortex crosses pressure opinengs "error within solver screen?

              • Re: Finer mesh gives the worse results
                Andrew Vassallo

                Never use VOLUMETRIC flow rate to measure inlet/outlet flows.  Always use MASS flow rate, this is what is conserved in the analysis.


                It sounds to me as though you have a pressure drop in the assembly, so the volumetric flow rates aren't adding up.  This is because the volumetric flow rate is calculated at the absolute pressures of the goal surfaces.  For example, a mass flow rate of 0.01 lb/sec Nitrogen at ambient pressure (14.7 psia) will be ~218 Liters/min of volumetric flow.  But at a pressure of 12 psia (slightly below ambient), the volumetric flow increases to 267 Liters/min for the same mass flow rate of 0.01 lb/sec because of the lower gas density.  So if you have a pressure drop in your system, you may be seeing higher volumetric flow values at the outlets due to the lower pressure (i.e., reduced gas density), even though the mass flows are properly conserved.