7 Replies Latest reply on Apr 12, 2010 10:05 PM by David Paulson

    Settling

    Alex Schmidt

      Hello,

       

      I am currently working on flow in settling tanks for biodiesel production.  The process uses large tanks to allow separation of Glycerin from the Biodiesel.  I am inputting the two fluids with .9 concentration of Biodiesel and .1 concentration of Glycerin.  The more dense Glycerin should settle and drain at the outlet at the bottom of the tank and there is another outlet at the top of the tank to drain the less dense Biodiesel.  I guess my question would be if Flow Simulation calculates the separation, and how would I do it?

       

      I have set outlet goals for “mass fraction” and “volume fraction” of the two fluids, but I have been getting the same concentrations that I have specified for the inlet.  I noticed that under my environment pressure boundary condition (outlets) that it also specifies the concentrations, I thought that that might be the cause for the problem. 

       

      Any help would be appreciated.

      Thanks,

      Alex

        • Re: Settling
          Kristofer Skyttner

          I dont think separation of fluids in the same volume due to gravity are supported in Flow Simulation.. Thats the answer I got from the support when I tried to simulate a while back.

           

          //Kristofer Skyttner

            • Re: Settling
              Bill McEachern
              I would disagree, though I have not tried it I think the code will do it. You would need to put gravity on with the vector in the right direction and have different density liquids but with a density variation less than 10. My understanding is that gravity will drive the flow and separate the liquids if the physics is right and by this I mean the diffusion potential is much less than the gravitational and the viscous effects work out as well. You may need to turn on the pressure potential switch as well. I would just do a simple test with material that are know to work and start out with a mass fraction of 50/50 and see if it redistributes itself - oil and water might be a good combo to test.
                • Re: Settling
                  Rich Bayless

                  Hello,

                   

                  When you add the concentration of two fluids, such as 0.9 for one, 0.1 for the other, Flow Sim will calculate the properties of the mixture using the mixture laws.  So you will have density, viscosity, etc., calculated using a weighted average to get the mixture properties.  Select a lid and create a fluid volume with the mixture.  This means you will have a 'single' fluid inside your volume.  This also means that the fluid will not separate, since it is a 'single' fluid.

                   

                  I'm not sure how you could apply one fluid to half of an interior volume and a second fluid to the other half of the interior volume in flow sim?

                   

                  I seem to remember that to follow a fluid to fluid interface, such as water sloshing in a glass with air, you need a moving mesh of some sort, which flow sim can't do.

                   

                  Flow sim will definitely show motion due to temperature induced density differences in a single fluid, such as hot and cold water in a storage tank.  Perhaps you could try to 'fool' the math by using a single fluid, and with hot and cold representing your two fractions?  Maybe use a very low thermal conductivity in the fluid to slow down conduction?

                   

                  Just my two cents.

                   

                  Cheers,  Rich.

              • Re: Settling
                David Paulson
                I don't think this can be done with two distinctive fluid flows.  Even if you could, I don't see how you could differentiate them on the flow trajectories.  But you can do this by treating the glycerin as a particle injection.  I just ran one with water as the fluid and water as the particles.  I don't have glycerin in my material database so can't help you on that account  But it does indicate that you can have liquid particles in a liquid.  You just need to define the user defined materials.  And you probably want the walls set to reflect instead of absorb.
                  • Re: Settling
                    Alex Schmidt

                    Thanks for your input everyone,

                     

                    And thanks David for the idea.  I am able to get settling with the glycerin, but one variable I am unsure of.  That is the Diameter of the injected particle.

                     

                    Help says:

                    Diameter. The initial particle diameter. The flow influences the particle temperature and, as a result, the particle density. Since the considered particles are particles of constant mass, the particle volume is variable.

                     

                    As I decrease the Volume of the particle, it becomes buoyant. As I increase the Volume of the particle, it sinks.  I would think that, with a constant mass, a larger volume would increase the buoyant force.

                    Also considering a liquid injection, what would I use as the particle size?

                     

                    thanks

                    Alex

                      • Re: Settling
                        David Paulson

                        Alex,

                         

                        I always try a number of different particle sizes.  Most particle injections that I deal with in real life are not of a constant diameter.  So, I run the analysis on various sizes which are within the realm of reality.  Flow will not predict a distribution of particle sizes, but it will tell you how a given particle size will respond.  And make sure that gravitional effects are turned "ON".  And realize also that partticles that go into solution with the main fluid will never settle.