26 Replies Latest reply on Jun 2, 2011 3:07 PM by Brendan Labrecque

    lid not on boundry between solid and fluid region

    Chris DiSalvatore

      Attempting to run a relatively complex model with a long water cooling flow path thru a variety of features (ss water traces, rubber tubes, etc.). I have refined the model to reduce cross sectional area changes and abrupt flow direction, however I continue to get the "lid not on boundry between solid and fluid region" message, and the program will not allow the definition of an inlet flow. I would rather not have to simplify the model, as it would not represent the configuration as it should, and I would sacrifice accurate looking results graphics. I am not sure if mesh optimization will necessarily fix the problem, or if the time spent doing that with the complexity of this model would be well spent. If anyone has had to deal with this kind of situation, I would be interested in knowing how it was finally resolved. Thank you.

        • Re: lid not on boundry between solid and fluid region
          Chris Michalski

          I'm assuming you used the lid generator instead of making your own lids? (I'm used to "selected surface is not on boundary....")

           

          When you run a geometry check and generate the fluid assembly is that lid touching a surface of the fluid component in question?  I've spent hours going through sections of a fluid assembly to find the one missing screw or weld gap that messes up a large model.

           

          What version of SW & Flow are you running?

            • Re: lid not on boundry between solid and fluid region
              Chris DiSalvatore

              Thanks again for your reply, Chris. Yes you are right, I am using lid generator. I'm guessing I may be able to resolve the matter if I generate my own lids? I have a circular inlet at a fitting, and when I select the surface, it auto generates a cylinder feature that prodrudes slightly into the flow path. I have run smaller sections of the model successfully doing this, where I had defined the endpoint with a static pressure in the same manner. I guess this is why I'm assuming it is due to the complexity of the model when I add more of the water trace. Whenever it gives me this message, and I run a geometry check, it always comes back with zero volume. I've gone back thru and carefully constructed the model geometry, making sure there were no gaps or openings, and I still get this message. I am using SolidWorks Premium 2011 with FLow Simulation, 64 bit version. I can try and mess with including my own lids. I am not seeing though why that may be creating the problem since it was able to run on a shorter simpler path, but maybe if I try positioning it differently in the path, the program may like it. Thanks again.

                • Re: lid not on boundry between solid and fluid region
                  Chris Michalski

                  Chris -

                   

                  unfortunately I don't have 2011 or I'd offer to take a look at the file.  If you can add BCs to a truncated model that means that somewhere in between the smaller model that works and the larger one that creates lid errors there is something wrong - a hole, a self-intersecting body, a zero-thickness feature.

                   

                  If you get zero fluid volume it means the model is open from inside to outside.  If you set it up as an external simulation (or clone it and make the clone external so you don't lose all your data) - then check the geometry of that - this way you can look at the fluid assembly and see where the outside connects to the inside.

                   

                  The automatic lids vs. manual shouldn't be the issue - I'd try looking at an external simulation first.

                    • Re: lid not on boundry between solid and fluid region
                      Chris DiSalvatore

                      Ok I will try and run it as an external model. Hopefully that will expose the source of the problem. What's still getting me is I have generated more of the trace to the simplified model, carefully a feature at a time even, and that message still came back. Unless I overlooked something, its not evident to me yet what's causing the problem, but maybe this will help to identify it. Thanks.

                      • Re: lid not on boundry between solid and fluid region
                        Chris DiSalvatore

                        I just discussed this problem with one of our ME's, and he thinks possibly there may be gaps in the traces where a weldment is not defined? I have stainless steel water traces that are rectangular in cross section, where 3 sides are welded to a surface plate. Would the program possibly be seeing gaps between the 3 sided trace and the plate? The parts have been mated, and the mated surfaces are on the same plane, so It doesn't seem like it should be the source of the problem, but in flow sim where there is a water flow thru the trace, is it maybe being treated differently to allow possible water seepage?

                          • Re: lid not on boundry between solid and fluid region
                            Chris Michalski

                            Weld gaps are a common problem - that is where I spent the most time recently in transitioning a model from build-able to simulate-able.  You specify an oversize hole so a tube fits and fill it with weld, but you have to create a part to represent that weld material for flow.

                             

                            If the surfaces are mated it should be fine as long as they do not form a line of contact - they must overlap.  If it is a U-channel mated to a flat plate that is larger than the opening it is fine, if the plate plate fits between the legs of the U it would have a problem.  If there is a line of contact it works down to the point of intersection and jumps across because there is zero thickness so as far as it can tell the fluid spaces are in contact.

                             

                            Are all of the changes in cross-section perfect mates or are these welded also? if they're welded they might need fillers there also.

                              • Re: lid not on boundry between solid and fluid region
                                Chris DiSalvatore

                                From what you are saying, I shouldn't have a weldment prob in this model. The 3 sided trace is mated to a flat plate and is overlapping so that there is not a line contact. The changes in cross section are perfect mates with no line contacts either. I'll try new manually generated lids and different positions for them in the model, as well as running as an external flow if I keep getting the same message.

                                • Re: lid not on boundry between solid and fluid region
                                  Chris DiSalvatore

                                  Hey Chris, I just thought I'd let you know I got the program to run. I'm guessing Flow Simulation does not like suppressed components. I had some in there, I deleted them out of my model, and it worked ok. I will be continuing to construct the path, and hopefully I can get it to run thru the entire thing. Thanks for your responses.

                                    • Re: lid not on boundry between solid and fluid region
                                      Chris Michalski

                                      were they overly complicated components that were surpressed?  I often surpress components from one configuration to another and the only hiccup I've ever seen is Flow complaining about "disabled components have been lost" or something.

                                       

                                      Just more proof that each simulation requires a unique perspective to setup and run correctly.

                                        • Re: lid not on boundry between solid and fluid region
                                          Chris DiSalvatore

                                          they were somewhat ... I've got some water traces in the model that are very snake shaped, and that was one of them. And yes I agree, each model is a completely different animal. This is my best guess as to what was keeping it from running .. either that or I was just lucky today. I hope that keeps up. Thanks again.

                                          • Re: lid not on boundry between solid and fluid region
                                            Chris DiSalvatore

                                            Hey Chris,

                                             

                                            I appreciate your replies on the issues that I raised on the flow sim program. I have resolved the solid and fluid region boundry issues, fortunately, and I have been running iterations on the complete model. The results I am getting are making sense, but one of the guys here built a similar system before, and thinks my numbers might be high. I wanted to know if you might want to check my inputs to see if there may be something overlooked. I believe I covered everything, but I'd be interested in knowing what is behind these result differences. Thanks.

                                             

                                            Chris  

                                              • Re: lid not on boundry between solid and fluid region
                                                Chris Michalski

                                                it all depends how complicated the simulation setup is.  I can only go up to 2009 so if you're running 2010 or 2011 you'd have to save it out as a parasolid and then I'd have to add back all of the simulation parameters.  I currently have some available time on one of my simulation computers I could get it onto though.

                                                  • Re: lid not on boundry between solid and fluid region
                                                    Chris DiSalvatore

                                                    That's great Chris, I appreciate it. I have a run going now, but when its done (in an hr or 2 hopefully) I'll save the model as a parasolid and let you know. Thanks.

                                                      • Re: lid not on boundry between solid and fluid region
                                                        Chris Michalski

                                                        ok, well seeing as it's Memorial Day weekend and i'm in Ohio  (and IT decided to change over our server today instead of waiting until tomorrow) I'm heading out shortly so I'll take a look on Monday.

                                                          • Re: lid not on boundry between solid and fluid region
                                                            Chris DiSalvatore

                                                            That's great Chris, thanks. I did check with one of the guys, and he thinks there may be a disclosure issue, so I'm not sure I can send a copy of our model. I can let you know my inputs though if you'd still like to help, and maybe you can tell if I've got everything covered, or if something has been left out. Thanks again.

                                                              • Re: lid not on boundry between solid and fluid region
                                                                Chris Michalski

                                                                yeah, even if you want to just make a PDF or JPEG with the BCs highlighted different colors and a summary of the situation.

                                                                  • Re: lid not on boundry between solid and fluid region
                                                                    Chris DiSalvatore

                                                                    Hey Chris,

                                                                     

                                                                    We have since discussed the model results vs the testing from the previous system, and have identified some possible reasons for the differences. One thing I would like to check is in the wizard where you define the heat transfer coefficient, where this is applied in the model. We want to select the exterior walls for natural convection to ambient, but the internal of the unit is in vacuum, and will need to be radiation only. I did not see in the wizard where this could be specifiedfied. Also, I do not have the program in front of me today, but I wanted to be sure where the inlet water temp is defined. I believe it was in the wizard, but I wanted to double check. Thanks for your help. It is appreciated.

                                                                    • Re: lid not on boundry between solid and fluid region
                                                                      Chris DiSalvatore

                                                                      Ok, I went thru the wizard with a new input. Concerning the convection issue where the unit is in a vacuum on the interior, I noticed in the wall conditions that it specifies default 'outer' wall thermal condition. Here I defined the heat transfer coefficient. Only radiation is otherwise input in the program, so I wonder if this is recognizing the exterior of my model only, defining that as natural convection, and leaving the interior as a vacuum since I only defined radiation there.

                                                                       

                                                                      Concerning the initial temp of the cooling fluid, I noticed in the initial conditions there is a section for input for 'thermodynamic parameters', and the temperature can be changed. It would appear this is for the cooling water temperature in when a flow rate is defined. However, if I specify a static pressure on a lid at the outlet of the cooling path, I noticed a thermodynamic parameter is also specified there defaulting to the initial condition that was set for temp. I wonder how this can be set so that the program allows the water to collect heat as it will thru the system, and just define the outlet resulting water temperature.

                                                                    • Re: lid not on boundry between solid and fluid region
                                                                      Rich Bayless

                                                                      Hello Chris,

                                                                       

                                                                      Many companies won't allow posting of proprietary models, because of disclosure issues.

                                                                       

                                                                      Perhaps you could create a non-proprietary model, with the important features that you are investigating, and then post that model?

                                                                       

                                                                      As some people say, create a safe sandbox model that you can play with and that you can share.  I've found this approach can help learn more about how Flo Sim works, without getting twisted up in details, such as a complicated suppress part.

                                                                       

                                                                      Just a suggestion.

                                                                       

                                                                      Rich.

                                                                    • Re: lid not on boundry between solid and fluid region
                                                                      Chris DiSalvatore

                                                                      Hey Chris, here are the inputs I have made to the run:

                                                                       

                                                                      Wizard unit system:     ips (in*lb*s)

                                                                      HTC Unit:                    w/m^2K

                                                                      Analysis type:             internal

                                                                      environment radiation:  72F

                                                                      Gravity:                       y=-386.22 in/sec^2

                                                                      Default fluid:                water

                                                                      Default Solid:               SS 321

                                                                      Wall Cond:                  HTC = 5.5 w/m^2K

                                                                      Default selected on remaining wizard items.

                                                                       

                                                                      The fluid domain was defined in the graphic when exiting the wizard.

                                                                      The boundry conditions were then defined:

                                                                      inlet mass flow 5.4 lb/sec (lid selected)

                                                                      outlet - static pressure 14.7 psi, 72F ambient

                                                                      radiative surfaces:     ss 310 smooth ox, (radiative part items selected from assembly model tree)

                                                                      radiative source:     900C (part sources selected from assembly model tree)

                                                                      Goals selected:

                                                                      surface goals selected, part surfaces selected from graphic, and temp of solid selected

                                                                       

                                                                      Solve and run.

                                                                       

                                                                      I believe I covered everything here, but if you see something overlooked, or have questions about the model, feel free to let me know.

                                                                      Very much appreciated. Thank you.

                                                                       

                                                                      Chris

                                                                        • Re: lid not on boundry between solid and fluid region
                                                                          Chris Michalski

                                                                          Chris -

                                                                           

                                                                          to make sure I understand what you're trying to model... you have a 900C heat source which is radiating to a water cooling block.  Radiation hits the absorbing surface, then there is a vcuum cavity in the block, then a portion of the block with cooling water passages.  The water flows in at 5.4lb/sec and flows out to an unrestricted tank at atmospheric pressure.  Did you define the inlet water temperature?

                                                                           

                                                                          Is there air surrounding the parts?  with a radiative heat transfer and water as the default fluid it is going to fill any non-solid space with water, do you need to define fluid sub-domains or is this accurate?

                                                                           

                                                                          I'm not sure what happens if you use the "ignore fluid bodies without flow conditions" option, if it would allow radiation to cross those volumes or what.  I've never tried to simulate anything with a vacuum.

                                                                            • Re: lid not on boundry between solid and fluid region
                                                                              Chris DiSalvatore

                                                                              Yes, you have it understood correctly. The 900C heaters are inside a vacuum cavity which is basically a rectangular box with water traces running on the surfaces. I did have an issue with the inlet water temp. It looks like it needs to be defined in the thermodynamic parameters in the boundry condition input. However, I notice that there is a temperature input also if there is an exit boundry condition defined. This is where I have the outlet to atmosphere, but I don't see why there should be a temp input here, since the temp of the water at this point should be defined by the heat it has collected thru the flow path.

                                                                               

                                                                              And yes, there is air surrounding the chamber, and the interior is at vacuum, and separate from the water traces which run on the outside surfaces. This is where I want to be sure that when I specify default outer wall thermal condition in the input wizard, this is giving me convection from the outside walls only, and my internal cavity at vacuum will have no convection since only radiation was defined there.   

                                                                               

                                                                              I haven't used the ignore fluid bodies without flow conditions option. Not sure what this would do to the exterior air, or what effect there would be on the inside of the chamber, which is just enclosed without a flow defined..

                                                                                • Re: lid not on boundry between solid and fluid region
                                                                                  Chris Michalski

                                                                                  On the outlet temperature, this is the starting temperature of the water that has exited.  It won't hold that constant, it just assumes that it is dumping into a large volume of that fluid so if the flow is low enough and ambient temperature high enough it could have an influence.  At 5lb/sec and 72F it won't have much impact.

                                                                                   

                                                                                  Unforunately with IT destroying our network I had to use a proprietary software package to generate this model.

                                                                                   

                                                                                  Because you define the heat transfer out of the exterior of the box at 5.5W/m2K then you shouldn't need the air information.  However, because you have water in the cooling loops and a vacuum inside I'm not sure how that will work.

                                                                                   

                                                                                  If I were doing it I would use a fluid subdomain to have water in the cooling loop and then air/Nitrogen/whatever fluid was in the cavity before vacuum was drawn as the inside of the box, but set it at a very low pressure.  You likely don't have absolute vacuum so use a realistic value based on how that vacuum is generated.  This way you don't have to worry about what Flow is going to do with the undefined volume.

                                                                                   

                                                                                  I've done a few simulations of a vacuum thermal process with <10Torr absolute and gotten fairly accurate results compared to thermocouple data of the real system so Flow can handle the low pressure stuff fairly accurately.

                                                                                   

                                                                                  When you take a cross-section of the results does it shows as a thermal gradient in the fluid inside the vacuum chamber?  if it is truly showing a vacuum you should have a hot source and zero thermal gradient in the "vacuum".