37 Replies Latest reply on Aug 25, 2013 11:10 PM by Jared Conway

    Gravity in Flow Simulation not performing?

      Hi,

       

      This is my first post, so please forgive any forum  gaffs I make.

       

      Gravity doesn't seam to have any effect on my flow  simulations. I have run a simple test in exactly the same way, having  gravity on first and and off second, but the results are identical. And  yes, gravity is acting in the right direction.

       

      As I am  working on a gravity powered device, it's crucial that this force be  included.

       

      I have been using this for only 4 months, but I have  managed to use it successfully to design a carburetor which, when  prototyped in  reality was very close to the flow simulation results. However, I had to  fake gravity by working out the additional pressure myself, and adding  it to the atmospheric pressure cap on top. I did this on advice from  another flow simulation user, but I still don't understand it and I  can't fake it this time.

       

      I've attached a screen shot of the set up. I hope someone can help.

       

      reservoir test cross section.jpg

      If i've not explained something or its a bit unclear, please do ask.

       

      Thanks

      James

        • Re: Gravity in Flow Simulation not performing?
          Bill McEachern

          try turning on the pressure potential switch in the initial conditions section of the general settings tab. Gravity is for density variations like bouyancy.

            • Re: Gravity in Flow Simulation not performing?

              Hi Bill,

               

              Thanks for the quick response and the idea on how to solve my problem.

               

              I looked at the setting and it was already switched on,  will turning it off help? I'm running a simulation with it off, now.

               

              Could you explain practically, what you mean by:

                   'Gravity is for density variations like bouyancy.'

               

              Thanks for the help.

               

              James

                • Re: Gravity in Flow Simulation not performing?
                  Bill McEachern

                  As far as I know, the pressure potential switch is for "head" type effects where the weight of the fluid is in play (you need to supply a direction vector so you may have to have gravity on for that). For effects like natural convection in say a thermal problem (hot air rises) you need to turn on gravity but not pressure potential to get the proper physical simulation (hot air rising)

              • Re: Gravity in Flow Simulation not performing?

                Also useful to this question.

                Default (-10).jpg

                This is a pressure plot.

                It shows the highest pressure at the bottom due to the mass of water, due to gravity.

                 

                the total pressure difference between top and bottom is about 350pa which I also get when I do a hand calc on paper, P=pgh (pressure = density*gravity*height).

                  • Re: Gravity in Flow Simulation not performing?
                    Bill McEachern

                    Whats's the problem again? I thought you could not get this effect which is what your original post was about?

                      • Re: Gravity in Flow Simulation not performing?

                        Hi Bill,

                         

                        While gravity appears to be on, and the effects can be seen in that pressure plot, it isn't having the effect on the water i think it should.

                         

                        Both lids are set to environmental pressure.

                        The bottom lid I leave unchanged,

                        The little-pipe lid set to be 10Pa lower in pressure.

                         

                        In a gravity free world, (imagine it were on its side so gravity wouldn't effect it), the slight negative pressure would cause flow out of the little pipe.

                         

                        However, turn on gravity, and the force of water due to gravity should cause it to flow the other way.

                         

                        Sorry for any confussion.

                        Thanks again Bill

                          • Re: Gravity in Flow Simulation not performing?
                            Bill McEachern

                            stuff is going to move from high pressure to low pressure which is what I see happening in all you shots. If you want it to go somewhere else you need to have the lower pressure of that location (height) in the model.

                              • Re: Gravity in Flow Simulation not performing?

                                thanks for your patients Bill.


                                I'm not sure I understand your answer.
                                I know 'stuff is going to move from high pressure to low pressure' but only if the pressure difference is great enough to overcome the force from gravity of the water in the device which wants to fall down.
                                I've drawn a diagram that I hope will help you and others.

                                 

                                thanks

                                Forum diagram.jpg

                                  • Re: Gravity in Flow Simulation not performing?
                                    Bill McEachern


                                    the way you have the problem formulated the answer is yes, in my view. as the head is added to the lower BC, you your 1 Pa is going to move it. You might want to put in a bigger tank and move the BC to surface of the water outside the vessel - like you have in your sketch. then the head will act on the outer free surface (modeled as a BC).

                                      • Re: Gravity in Flow Simulation not performing?

                                        Thanks Bill.

                                         

                                        Sorry, I don't understand what you mean. I take it BC is Boundary Condition?

                                         

                                        How are you saying to model this? and why?

                                         

                                        'move the BC to [the] surface of the water outside the vessel'

                                        I don't understand what difference that will make, given the pressure is the same on the inside as the outside. In modelling terms, won't the BC be at the same level as shown in the below diagram?

                                         

                                        I have tried increasing the tank size and reducing the pressure difference between the Lids, to try to find a point at which the pressure difference (now only 1Pa) will not produce at upward flow. But even with a massive tank, that 1pascal of negative pressure is capable of lifting 10Kilos (22pounds), Although the flow rate is getting very weak compared to the first small tanks with large pressure differences.

                                         

                                        Thanks for your help Bill, I'm nearly bald from scratching my head over this.

                                         

                                        James

                                        Forum diagram2.jpg

                                        • Re: Gravity in Flow Simulation not performing?

                                          Hi Bill,

                                           

                                          I decided to go to 0 pascals. No pressure difference. This should surely cause water to flow down under the force of gravity. But I get no flow at all, up or down. So Gravity can't be working.

                                           

                                          Got any more ideas of how to switch it on?

                              • Re: Gravity in Flow Simulation not performing?
                                David Paulson

                                James,

                                 

                                On this Earth your 10 Pa difference in pressure should yield a differential heigth within your tank of about .89 micron.  If you selected gravity and pressure potential you should have no flow, just a slight change in the liquid level in your tank.  Have you defined your fluid as water or as a gas??  What do you get if you RMB click View Settings and set to velocity instead of pressure??  How high is your tank??

                                  • Re: Gravity in Flow Simulation not performing?

                                    That's good to hear David ! that's exactly what I thought, I hadn't  worked it out in mircons but I guessed it would be something like that.

                                    Of   course, in the simulation, the model starts full of water, so rather  than a .89 mircon rise, we expect a dramatic flow down?

                                     

                                    So   I'm wondering what I'm doing wrong in the simulation??

                                     

                                    I've  screen  grabbed the settings I know, and attached a velocity plot as requested.

                                    I've also done a calculation sheet, which I hope, worked out the  pressure required to be in equilibrium, 1.26MPa.

                                     

                                    I've  been increasing the tank height to see if makes any difference, its now at 100mm.

                                    What do you  think?

                                     

                                    Forum diagram Fluid.jpg

                                    Forum diagram Initial conditions.jpg

                                    Forum diagram Gravity.jpg

                                    Forum diagram Velocity Plot.jpg

                                    Forum diagram maths.jpg

                                      • Re: Gravity in Flow Simulation not performing?
                                        David Paulson

                                        James,

                                         

                                        I modified your model to make it taller so that we can better observe the efect of gravity on the "flow".  In the pictures that follow, the cylinder is 120 inches high by 24 inch diameter just for graphical display purposes.  For all cases both BC's are set to an environmental pressure of 14.69 psi at 68.09 F.  When gravity is turned "OFF" I get the following flow trajectories:

                                         

                                        Default (1) copy.jpg

                                        As you can ee there is no flow because there is no pressure difference.  However when we "enable" gravity we get the following result with identical BC's:

                                         

                                        Default (1)_3 copy.jpg

                                        So, it now seems that we have entered the Twilight Zone where gravity will cause an upward flow, esentially defying gravity.  (And my gravity setting is -32.18 ft/sec2 in Y dim)  Pressure potential was turned "OFF" for both BC's.  I cannot comprehend the efect that Flow is predicting with a zero pressure difference between the BC's where only the influence is causing flow.  Perhaps this imaginary flow could be called iFlow or jFlow, depending upon the imaginary operator you might wish to add to this result.

                                         

                                        I hope that someone with more devine wisdom of Flow can offer a down to Earth explanation.

                                          • Re: Gravity in Flow Simulation not performing?
                                            Rich Bayless

                                            Hello Dave and James,

                                             

                                            I think at this point, that this may well be a bug in Flow Sim.

                                             

                                            I also think that someone should file a service request for this, unless someone already has posted this as a request?

                                             

                                            I'm willing to post a request, if either of you gentlemen hasn't already.  Let  me know either way, please.

                                             

                                            Cheers,  Rich.

                                              • Re: Gravity in Flow Simulation not performing?

                                                Hi Rich and David,

                                                 

                                                Thank you both for your contributions and help in tying to resolve this issue.

                                                 

                                                I liked the idea of creating a centrifuge to create my own gravity but this doesn't work either, the mass of the fluid is not in the calculation.

                                                 

                                                Rich, I would very much appreciate you logging this problem to Solidworks, because being on a student license myself, doesn't entitle me to do so. If you get a response, I would love to hear it.

                                                 

                                                I've had another idea of how to model this, which I will share with you if it proves successful.

                                                 

                                                Thanks again

                                                 

                                                James

                                                  • Re: Gravity in Flow Simulation not performing?
                                                    Rich Bayless

                                                    Hello James and Dave,

                                                     

                                                    I was looking for something else and found the below article in the SW knowledge base.  I suspect this may be the answer to the original question, but I don't have time right now to check.

                                                     

                                                    Cheers,  Rich.

                                                    -------

                                                    Solution Id: S-015123 
                                                     
                                                      Product:  
                                                     
                                                      Updated: 8/27/2009 04:41:05 AM
                                                     
                                                      Area: COSMOSFloWorks
                                                     
                                                      Sub-Area: FW Fluid mechanics
                                                     
                                                      Question: Can Flow Simulation model the pressure hydrostatic effects of a drain where there is an elevation change within the drain?
                                                      
                                                     
                                                      Answer: To correctly model the hydrostatic effect of the water the developers wrote the following to calculate the pressure potential (grZ).

                                                    Absolute pressure is related to the pressure potential by the formula: Pabs=Ppot+grZ.

                                                    Here Pabs – absolute pressure level, Ppot – pressure potential, g=9.81 m/s2, r - water density, Z – coordinate (representing the height term.

                                                    The pressure potential term for COSMOSFloWorks is equal to the absolute pressure at Z=0. For example a drain model with the Z coordinate from top to bottom of the drain is equal to 0.3 m. The gravitation pressure term is then equal to grZ= 2943 Pa. Looking at the pressure potential at the top point where absolute pressure is equal to 101325 Pa. Pressure potential is equal to Ppot=Pabs- grZ=98382 Pa.

                                                    Therefore, to solve this problem the following boundary conditions can be set:

                                                    -  on the top surface: absolute pressure is equal to 101325 Pa;

                                                    -  on the exit of the drain: absolute pressure is equal to 101325 Pa.

                                                    The initial condition in the General Settings can be set with a pressure potential of 98,382 Pa.

                                                    Pabs=Ppot+grZ
                                                    Ppot =  Pabs - grZ
                                                    grZ  = 9.81 * 1000 * .3 = 2943
                                                    Ppot = 101325 – 2943
                                                    Ppot = 98,382 Pa

                                                  • Re: Gravity in Flow Simulation not performing?
                                                    David Paulson

                                                    Rich,

                                                     

                                                    My previous flow model that illustrated flow was deceiving.  When you depict the flow with lines with arrows instead of just plain lines you will see that the flow is downward and caused by gravity........

                                                     

                                                    Default (1)_1 copy.jpg

                                                    Environmental pressure is 14.69 psi in both cases, pressure potential not enabled.

                                                     

                                                    Dave

                                            • Re: Gravity in Flow Simulation not performing?
                                              Rich Bayless

                                              Hello James,

                                               

                                              From your model and your pictures, you intend for this device to use a reduced pressure applied to the outlet to pull water up and into the cylinder?   (Such as, vacuuming water and oil up from the surface of the Gulf?)

                                               

                                              It seems to me that you are expecting Flow Sim to show a free surface between the water and the atmosphere?  Showing the water level inside the cylinder rising as more vacuum is applied to the outlet?

                                               

                                              Flow Sim won't work with a free surface, you have to take this into account as you pick your boundary conditions.  Such as apply sufficient negative pressure at the outlet to obtain a zero pressure relative to the environment at the bottom of the cylinder.

                                               

                                              Cheers,  Rich.

                                                • Re: Gravity in Flow Simulation not performing?

                                                  Hi Rich,


                                                  Thanks for your reply.

                                                   

                                                  "From your model and your pictures, you intend for this device to use a  reduced pressure applied to the outlet to pull water up and into the  cylinder?   (Such as, vacuuming water and oil up from the surface of the  Gulf?)"

                                                   

                                                       Yes, sort of, I'm just trying to see if gravity is having the effect it should.

                                                  Having a one pascal pressure difference should not cause water to be sucked up. as David Paulson says.

                                                  But in the flow simulation, the solver finishes with a steady flow up.

                                                   

                                                  "It  seems to me that you are expecting Flow Sim to show a free surface  between the water and the atmosphere?  Showing the water level inside  the cylinder rising as more vacuum is applied to the outlet?"

                                                   

                                                       Sorry, not sure what a free surface is.

                                                       I realise that in my diagram, the cylindar is shown with air in it, with water at the bottom of the cylindar. However, I know that in Flow Sim, the device is full of water. The same thing should happen though, if Gravity is switched on and working correctly. Case A- full of air-  no water is sucked up. Case B- full of water- water flows down.

                                                   

                                                  "Flow Sim  won't work with a free surface, you have to take this into account as  you pick your boundary conditions.  Such as apply sufficient negative  pressure at the outlet to obtain a zero pressure relative to the  environment at the bottom of the cylinder."

                                                   

                                                       Again, I'm sorry Rich, but I don't understand this suggestion.

                                                   

                                                  Many thanks again for your work.

                                                   

                                                  James

                                                   

                                                  If anyone felt like skype or google talk, I'm jmebartlett

                                                  and my email is jmebartlett@gmail.com

                                                    • Re: Gravity in Flow Simulation not performing?
                                                      Rich Bayless

                                                      Hello James,

                                                       

                                                      perhaps you should post your model?

                                                       

                                                      cheers,  Rich.

                                                        • Re: Gravity in Flow Simulation not performing?

                                                          Model is attached.

                                                           

                                                          Also, the simulation files can be downloaded here to save re-running the test.

                                                          http://dl.dropbox.com/u/2160796/1.0_links/7.zip

                                                            • Re: Gravity in Flow Simulation not performing?
                                                              Rich Bayless

                                                              Hello James,

                                                               

                                                              I ran your model and it seems okay to me.  It is a 100 mm diameter cylinder that is 100 mm tall.

                                                               

                                                              To see what is going on,  I evaluated surface parameters for the inlet lid and for the outlet lid, for pressure, velocity, and mass flow.

                                                               

                                                              Environment pressure is set to 101,325 Pa.

                                                               

                                                              inlet lid:

                                                              Pressure  100,393 Pa

                                                              Velocity     0.0194 m/s

                                                              Mass flow  -0.000726 kg/s

                                                               

                                                              outlet lid:

                                                              Pressure  101,277 Pa

                                                              Velocity     1.44e-6 m/s

                                                              Mass flow  0.000726 kg/s

                                                               

                                                              So the inlet lid has dropped below environment pressure by 884 Pa, it has pulled a vacuum and water is flowing into the chamber.

                                                              The inlet mass flow matches the outlet mass flow, except for the direction of course.

                                                              the velocity probably makes sense, with a 10 mm inlet diameter and a 100 mm outlet diameter.

                                                               

                                                              In my conversion program, 884 Pa is just under 100 mm of water column.

                                                               

                                                              Cheers,  Rich.

                                                                • Re: Gravity in Flow Simulation not performing?
                                                                  Rich Bayless

                                                                  one more note,

                                                                   

                                                                  I turned off gravity and the pressure at the inlet and the outlet now are the same at 101,324 Pa,

                                                                   

                                                                  rich.

                                                                  • Re: Gravity in Flow Simulation not performing?
                                                                    David Paulson

                                                                    Rich,

                                                                     

                                                                    You are modelling the flow with a 884 Pa pressure difference, which is 9x the pressure difference that James modeled.  But if 884 Pa is under 100 mm water column, how can you have the flow out of a nozzle at 100 mm??  The nozzle may be slightly lower which accounts for the very low flow.  But if you increased the heigth of the tank, or decreased the pressure difference the flow should be 0 due to the effect of gravity, should it not??  If gravity is working properly the flow would be limited by gravity and would be bouncing off the liquid level in the tank.

                                                                      • Re: Gravity in Flow Simulation not performing?
                                                                        Rich Bayless

                                                                        Hello Paul,

                                                                         

                                                                        I loaded James' model and ran it, as is, without changing the settings.  The model has environment pressure on both lids, set at 101,325 Pa.

                                                                         

                                                                        Why did my run come out 9x higher than his?  Now we have another mystery.

                                                                         

                                                                        Also, when I turned off gravity, the pressure difference basically went to zero, and yet the mass flow rate did not change.  One would think it would go to zero flow.

                                                                         

                                                                        What do you get when you run his model?

                                                                         

                                                                        I'm running SW2010 SP3.1, and his model was from an older version of SW.

                                                                         

                                                                        Cheers,  Rich.

                                                                          • Re: Gravity in Flow Simulation not performing?
                                                                            David Paulson

                                                                            Rich,

                                                                             

                                                                            I didn't run James' model because I am on SW 2009 and I presumed he was on SW 2010.  In his initial post he used a 10 Pa pressure drop, which should only change the level in the cylinder less than one micron if gravity is enabled.  If the pressure difference is only 10 Pa there should be no flow.  But Flow depicts a flow.  Independently, I modelled his situation as a 4 inch vertical pipe with atmospheric pressure on the top lid and pressure low enough on the bottom lid that if it were a pump applying that pressure it would only lift the water half way up the pipe.  But in this case as well Flow predicts flow to the top of the pipe.

                                                                             

                                                                            So, in answer to James' original post that gravity in Flow doesn't perform correctly, I have to agree with him.  But, Flow assumes flow in a dynamic state and seems unable to predict a static condition.  It also assumes the vessel is full of water between the BC's at the start of the calculation.  But even then, when gravity is greater than the pressure difference, there should be no flow.

                                                                             

                                                                            I think that this is a limitation in the program and there may be no work around that will lead you to the correct answer.  I do not know if this would be handled correctly in other CFD programs because my experience is only with Flow.  But in my opinion Flow should establish a virtual boundary based upon the static head and gravity and limit flow to the volume between the inlet and the virtual boundary....which would be the surface of the water.  One would not think that prediction of fluid level in a vessel should be beyond the limits of a CFD program.  What do you think??

                                                                              • Re: Gravity in Flow Simulation not performing?
                                                                                Rich Bayless

                                                                                Hello Dave,

                                                                                 

                                                                                as you said, limitations.

                                                                                 

                                                                                As a test, I set up a cylinder with the long axis in line with the Y-axis so gravity applies a force to the fluid.

                                                                                 

                                                                                I added two lids, one at the top, one at the bottom.

                                                                                 

                                                                                the cylinder is 0.5 m tall, and I set the units for pressure to be millimeters of water.

                                                                                 

                                                                                setting the upper lid and/or the lower lid to environment pressure, static pressure, or total pressure, all show hydrostatic pressure on the column of water as 500 mm of H2O, with essentially zero flow in all cases.  Flow sim is calulating the hydrostatic pressure due to gravity, but does not use that pressure to drive flow.

                                                                                 

                                                                                Then, I changed the model to be external flow, deleted the lower lid, and set the upper surface of the upper lid to environment pressure.

                                                                                 

                                                                                Gravity causes flow of the water downwards around the outside of the cylinder, and pulls a vacuum inside the cylinder.  The velocity inside the cylinder is essentially zero.

                                                                                 

                                                                                Next, I added a small hole in the upper lid, and you can see flow into the small hole, due to the vacuum inside the cylinder.  Interestingly, if you delete the environment pressure from the upper lid, flow goes to zero throughout the domain and only hydrostatic pressure shows up.

                                                                                 

                                                                                to summarize, use an external study and apply an environment pressure somewhere on your object.  the attached model is from SW2010 SP3.1.

                                                                                 

                                                                                Cheers,  Rich.

                                                                    • Re: Gravity in Flow Simulation not performing?
                                                                      David Paulson

                                                                      I think that the issue here is that Flow has inherent assumptions which lead to limitations in the software.  Flow in your model will assume the cavity to be full of water, even though gravity will make this impossible.  And as Rich Bayless pointed out Flow will not work with a free surface.  Flow is into "flow" and if gravity makes the flow impossible, Flow must be ignoring gravity.

                                                                       

                                                                      I think that your model is very simple and straight forward.  We could come to the same conclusion if we just had a centrifugal pump cconnected to a vertical pipe.  Water in that case should flow upward in the pipe to the heigth of the maximum shut-off head the pump will deliver.

                                                                       

                                                                      I hope someone on here cam help us to understand why gravity does not seem to have an impact on the flow.  Your model may get a little more confusing because you are using an atmospheric pressure difference to lift the water.  But everyone should understand a "centrifugal pump"  and a vertical pipe.  When you apply a minimal pressure at the bottom of the pipe and atmospheric pressure at the top of the pipe Flow predicts flow out of the top of the pipe.

                                                                       

                                                                      Shouldn't happen.

                                                                  • Re: Gravity in Flow Simulation not performing?
                                                                    Diego Munoz

                                                                    I'm just wondering are the red arrows that are representing the flow, is that created in SolidWorks when you use motion study? Or was it created in another software.