14 Replies Latest reply on Dec 6, 2018 11:39 AM by Bill McEachern

    drag coefficient for sphere?

    Anton Jonsson

      Hi I am trying to validate Solidworks Flow simulation using this tutorial http://www.webpages.uidaho.edu/mindworks/Adv%20Solidworks/CFD/Drag%20coefficient%20of%20sphere%20-%20Final.pdf I seem to get reasonable values for Cd for Re 10 but for Re 10,000 the Cd value seems quite off 0.17 (solidworks)  instead of about 0.4 (theory). Any help with trouble shooting this to get the correct value would be much appreciated. I have attached the model I did in the post.

        • Re: drag coefficient for sphere?
          Eric Christison

          Looks like SW thinks Re is 10^5 rather than 10^4.

           

          Could this be a simple typo error in the data?

           

           

           

            • Re: drag coefficient for sphere?
              Amit Katz

              The initial settings are OK, 3m/s of air moving past a 50mm diameter sphere, that gives Reynolds number of 9720 so that's OK. I'll take a look at your setup and see if I can find any obvious reason why the calculation is off.

                • Re: drag coefficient for sphere?
                  Amit Katz

                  I can't really tell where the issue is, everything seems to be OK but the behavior of the wake is similar to high Re flow with turbulence dominated drag. Maybe you should ask your TA what's going on, this may be an object lesson in the limits of the software.

                    • Re: drag coefficient for sphere?
                      Anton Jonsson

                      Thanks for the reply Amit and Eric. It just seems weird that Solidworks flow simulations is close to dead on for Re of 10 but is so wrong for Re 10,000 I mean surely it should be able to give a reasonable output to such a simple object. Unfortunately my TA doesn't have much experience in using solid works flow simulation. Even if it was the typo and they meant a value of Re 10^5 it should still give a value of around 0.4 since it is a smooth not rough sphere (I wish it was just a typo). Still many thanks for the suggestions and help.

                • Re: drag coefficient for sphere?
                  Amit Katz

                  When I employ a high resolution mesh and a smaller time step I see the Cd creep up to 0.3 and the wake becomes unsteady periodic, as you'd expect to see from a bluff body in steady flow. It's still far off from laboratory results as shown by that chart though.

                  • Re: drag coefficient for sphere?
                    Ashton Abright

                    I know this is old, but I wanted to let anyone still curious know, if you set up a local mesh on the surface, you get a significantly more accurate result. I got about .38 with these mesh settings.

                     

                    Capture2.PNGCapture.PNG

                    • Re: drag coefficient for sphere?
                      Mark Keown

                      There are two big problems and two other problems.

                      1. For the force on the sphere use a surface goal, force in x.  Not even sure what the GG force would be.
                      2. The plot is a log log plot so difficult to read the one you have - this one is better.
                      3. Increase the domain size down stream.
                      4. Put a local mesh around the sphere.
                      5. This is not solved yet but for 3 m/s this is where it is at.  Reynolds is ~ 11,900 at 3 m/s say 12^3.  The Equation Goal 1 has N for units which is not correct should be unit-less.  how did this get in bold?

                        • Re: drag coefficient for sphere?
                          Ashton Abright

                          Mark, awesome insight, a couple of things I might clear up.

                           

                          1. Surface goal and global goal should be the same when you select all the surfaces (in this case the sphere)

                          5. Equation goals default to whatever goal unit is inputed first, so in that case newtons, you can manually change it.

                           

                          .313969 seems a bit low for a reynolds number of 11,900 , what did you get on finishing the run?

                           

                          Thanks for the input.

                            • Re: drag coefficient for sphere?
                              Mark Keown

                              Ashton,

                               

                              I may be reading this wrong but looks like Cd is around 0.3 - 0.4

                               

                              Shown below is the domain size that may need to be a bit longer lee of the ball.  Flow development still 'pulling' air.

                              Again still not finish solving but this should be close enough.

                              Also note that Force on ball X and GG Force x are not the same value.  Still not sure what GG Force x is referring to.

                              The Cd value will increase if:

                              - The domain is larger.

                              - The ball has some surface roughness.

                                • Re: drag coefficient for sphere?
                                  Ashton Abright

                                  Mark,

                                   

                                  This is how I was reading the chart, 12,000 = 1.2*(10^4)

                                   

                                  pastedImage_1.png

                                   

                                  Which would mean Cd is a little larger than .4

                                  The flow development is something I did not consider, and would affect my number.

                                  Im now curious how gg and sg are different, and I will run both when I get the chance to verify.

                                   

                                  Thanks for entertaining this discussion, as a student I really appreciate the chance to learn this.