8 Replies Latest reply on Jun 20, 2017 10:35 AM by Amit Katz

    Asymmetric Results from Symmetric Setup

    Mengfan Li

      Hi there,

       

      An assignment from the course asks me to explore how wind velocity looks like in an ideal block in Solidworks Flow Simulation. The graph posed below is a looking-down view, as four grey building studs make a block. Wind is blowing from left side only, and all other setups that I am aware of are symmetric, but the velocity field is not symmetric at all, as shown in the graph shown below. I am trying to figure out what makes the simulation asymmetric with symmetric setups. Wondering if anyone can help me out

       

      Here are some setup details:

        Computational domain is 40m*40m*15m

        Block consists of 4 buildings, each of which is 10m*10m*10m

           Each building is 10 m away from another closest building, and the building edge is 5 m away from the computational boundary

        Fluid is air as external flow

        Wind enters horizontally the domain only from the left boundary at the velocity of 6 m/s

       

      Please let me know if any other detail is needed. Thank you!

       

      1.0_6_vector.png

        • Re: Asymmetric Results from Symmetric Setup
          Mark Keown

          I would make the domain significantly larger say 10x what you have in all directions.  This larger domain will capture the pressure bubble fore and turbulence / re circulation aft.  You will also get flow down the front of the building (from ~ center to bottom).

          The below publication and anything published by B. Blocken will be helpful to you.

            • Re: Asymmetric Results from Symmetric Setup
              Mengfan Li

              Hi Mark,

               

              Thanks for your reply!  I have read the paper you have suggested in your post, with several papers cited by the paper, and I made a second trial with a much larger computational domain, such that the inlet, lateral, and outlet region is at least 5 times the height of the block, and the top region 15 times. The results did come out more as expected, as symmetry can be observed in the velocity plot.

               

              Below is a plot generated from one of my trial, block being 10 meters high and velocity at 1.75 above the ground being 5 m/s. Regarding the wind velocity, instead of uniform profile, a log wind profile is introduced.

               

              What I found a little be confused is that after making the computational domain significantly larger than the previous trials, calculation time is much shorter. Could this be because larger computational domain can have a better-developed flow, compared to smaller domain?

               

              1.0_5_v.png

            • Re: Asymmetric Results from Symmetric Setup
              Bill McEachern

              You are running a steady state RANS calc with Flow Sim. If the Reynolds number is in the right range you will get vortex shedding. Where the solution converges if vortex shedding is present is any one's guess but chances are very good it won't look symmetrical. Just because the flow geometry is symmetrical doesn't mean the flow will be symmetrical. Symmetry in flow is a lot more tricky than in structural simulation where it's obvious. The other thing, even if the flow is symmetrical, sim programs can wonder off. If you know the flow is symmetrical you can run a wall down it an enforce symmetry with a BC.

                • Re: Asymmetric Results from Symmetric Setup
                  Amit Katz

                  Bill McEachern wrote:

                   

                  You are running a steady state RANS calc with Flow Sim. If the Reynolds number is in the right range you will get vortex shedding. Where the solution converges if vortex shedding is present is any one's guess but chances are very good it won't look symmetrical. Just because the flow geometry is symmetrical doesn't mean the flow will be symmetrical. Symmetry in flow is a lot more tricky than in structural simulation where it's obvious. The other thing, even if the flow is symmetrical, sim programs can wonder off. If you know the flow is symmetrical you can run a wall down it an enforce symmetry with a BC.

                  Yup, that's a great point Bill. Vortex shedding is a common fluid dynamics phenomenon in flow around bluff bodies, which this certainly is.

                   

                  I have to wonder exactly what "flow looks like an ideal block" means, in regards to Mengfan's assignment. I'm really not sure what that is, maybe it's some kind of translation goof.

                    • Re: Asymmetric Results from Symmetric Setup
                      Mengfan Li

                      Hi Amit

                       

                      Thanks for your reply! By "ideal block", I mean the building and ground surface roughness length is set to be 0.

                       

                      Regarding vortex shedding, honestly, I am not familiar with the phenomenon. I went over some examples of vortex shedding, as flow passes by a single obstacle. What I don't understand now is that, taking the flow plot below as an example, when flow is going by an object and Reynolds number is in the right range, why the very first vortex, Red "A", is formed from the flow at the bottom of the picture. Is it possible to produce a different vortex shedding look with the same flow in the picture, for instance, the very first vortex is formed from the top of the picture?

                       

                      vortex-shedding.gif

                        • Re: Asymmetric Results from Symmetric Setup
                          Amit Katz

                          Mengfan,

                          That picture you are looking at, it's not steady flow. What you are seeing is a "frozen snapshot" of a time dependent process. It's not that the first vortex always forms from the "bottom" (in fact you may even be looking at at top down view!) but rather that there is a pattern of vortices shedding from each side. It's like a pendulum, as each vortex forms and separates, the next vortex begins forming on the other side of the body.

                           

                          Vortex-street-animation.gif

                           

                          15. Flow Inbstabilities - YouTube

                    • Re: Asymmetric Results from Symmetric Setup
                      Thomas Wagner

                      You may probably also want to check the mesh if it is a symmetric as your solid bodies.

                       

                      Regards,
                      Thomas