7 Replies Latest reply on May 1, 2013 10:58 AM by Jared Conway

    Greenhouse flow going the wrong way?

    Alastair Donnelley

      This is an odd one.


      I'm attempting to set up a simulation with solidworks flow simulation that is essentially modelling the flow of air though a greenhouse-like structure.


      So the boundary conditions are all set to environmental pressure, and the only source of energy is solar.


      The results of the simulation are entirely believeable, that is the trajectories and turbluence seem right: except that the direction of the airflow appears to be exactly opposite to what I would expect.


      Rather than air entering the greenhouse at the base and flowing upwards, it's entering the top of the greenhouse and flowing downwards!


      What have I missed? I have attached an elevation view of the greenhouse structure. It is open at the left and right verticals, and at the verticals under the 'eaves' (so environmental pressures at all vertical elements shown in the elevation view.) Also attached is a detail showing the direction of airflow - downwards??


      It looks like air is rushing down the 'chimney', and out the base...



      I am eternally grateful for any assistance in working out what the heck is going on here. Also attached is the SLDPRT file.


      Thanks in advance,



        • Re: Greenhouse flow going the wrong way?
          Bill McEachern

          When I have encountered similar issues increasing the mesh density worked for me. I would also make your domain a bit taller, but that is just an opinion. It might work the way it is.

          • Re: Greenhouse flow going the wrong way?
            Jared Conway

            you're doing this as an external analysis so there is really no need to have the lids with environmental pressure applied to them. they can just be open to atmosphere. otherwise they act like portals to environmental pressure. this might be fine but it might cause some funnyness to happen. if you do this, Bills suggestion about making the comp domain bigger is a good one. (And generally a good one in this case because I think the affected area will cross the boundary of your comp domain which is a no no)


            >>sorry, spacing on Friday afternoon here. you have it as an internal analysis, there is really no need to have the comp domain be any bigger than what you have. and you can keep the environmental pressure conditions.


            the other thing I'd be a bit careful about is the use of non-radiating surfaces. generally the developers recommend going with black body and then using radiative surfaces to tune things. are you doing this because they are glass?


            also, you have 800W/m^2 on the outside faces plus environmental radition and solar radiation. what are your real boundary conditions? Is there really a heater on the outside faces?


            another thing to consider is that you have the bottom of your comp domain on the bottom of your green house. flow simulation doesn't consider that a "Ground" unless you add it.


            i'm running the analysis now but suspect that one of these BCs isn't necessarily real and causing your trouble.

              • Re: Greenhouse flow going the wrong way?
                Jared Conway

                also, is there a real need for a transient analysis or can it be steady state?

                  • Re: Greenhouse flow going the wrong way?
                    Alastair Donnelley

                    Hi Jared,


                    Thanks for your reply, it was most helpful.


                    I've now got a number of valid-looking sims going. I've removed the 800W heater. Not my most brilliant moment, that.

                    Also I've left the wall conditions as the defaults.


                    The most important realisation (and at this point, I'm flaking out a little myself, so I'm not sure if I had it right in the version posted perviously...) was that the material definitions needed to be edited to be transparent to solar radiation, thermal radiation, or both.


                    (For those following along at home, right click on 'Solid Materials' in the Flow Sim menu, 'Add Solid Material from Model' -- or, if your model doesn't have a material, add one -- and then in the rollout click 'Transparent' and edit the parameters to your liking.)


                    The other thing that I've found gives weird results is if you apply a material to your lids. When you add a material to your model, make sure you delete the lids out of the selection.


                    The only thing that I don't know how to do from your suggestions, Jared, is to add a ground in Flow Sim. Any pointers?


                    I'll post a Solidworks file and/or a few salient images when I'm a little less 'under the pump'.


                    Thanks again, Jared and Bill.

                      • Re: Greenhouse flow going the wrong way?
                        Jared Conway

                        Hi Alastair, glad you're making some progress.


                        on the walls being transparent, yep, that's a good call. With the HVAC module that you can buy, you can actually make glass semi-transparent to radiation if that is the case. (Not sure if you have a commercial or EDU license, but it is a separate add-on to Flow Simulation)


                        On lids, I usually make them insulators so that they don't affect anything. In general we teach to make the default solid insulator in complicated cases because it makes it more obvious where there are mistakes in your analysis. (you won't see any heat in those parts)


                        For ground, just model a plate and apply a material to it. That will be your ground. Or since you're working with an internal analysis, you could make it non radiating as an external wall condition. For an external analysis, just model a plate, make it bigger than the greenhouse, apply a material and the appropriate radiative surface and you're good to go. Your comp domain should end at the boundaries of the ground plate and extend upwards.


                        Good luck!