19 Replies Latest reply on Jul 1, 2013 11:27 PM by Jared Conway

    Pressure Vessel Simulation

    Lincoln Welch

      Hey guys,

       

      This may be a seemingly uneduated question, but I am having trouble wrapping my head around it so please bear with me on this one.

       

      I am wanting to simulate a highly pressurized tank, an accumulator to be exact, that will be at 30k psi. I do not have the pressure vessel simulator on the system I am currently using so I am simply using SimulationXpress (SW 2012) to simulate the pressure of the tank. I am getting results, but they are very hard for me to believe. I have made a custom material to simulate 410 SS and am ultimately trying to come to a safe OD and ID of the vessel that we will potentially manufactur. I am not seasonsed in the design of pressure vessels so I have been somewhat shooting in the dark on this one.

       

      When I run my simulation, I get super low FOS but it appears that SW is showing maximum localized stresses along the ID of the tank. I realize that this will be the case but even creating a tank that has a VERY small ID compared to the OD (we're talking something around a 2" ID, 12" OD just to test) I still get a FOS less than 1.0. Is the method I am using not appropriate for what I am wanting? I've attached a picture of what I am seeing. Any help or general guidance on this is greatly appreciated.

       

      ADDITIONAL INFO:

      Modeling after a 12" OD and a 36" length just to get a baseline of what will be safe wall thickness using the 410SS.

        • Re: Pressure Vessel Simulation
          Jared Conway

          where is your model fixed in relation to your restraint?

            • Re: Pressure Vessel Simulation
              Lincoln Welch

              I have it fixed at the ends in that image. I have tried adding in the end caps as well as a stand for the vessel to rest on with similar results. The actual vessel will have caps that screw into the ID to creat the vessel.

               

              When I added in the stand I fixed the bottom face that would be bolted to the concrete floor. Still not an understandable test result.

                • Re: Pressure Vessel Simulation
                  David Paulson

                  Lincoln,

                   

                  I have designed pressure vessels on SW, but they were nowhere near 30,000 psi.  Typically 300-500 working pressure. 

                   

                  Are you building this to ASME Code?  SW will not advise you of ASME Code requirements as to welds and nozzle reinforcements and I see none on your model.  I would be very surprised if you could use screw in caps at that pressure, but if your company is in this specific line of business they would know better than I.  You can probably easily cross check the vessel wal thickness to sustain that pressure from independent sources, but the stress points will be on the welds and nozzles.

              • Re: Pressure Vessel Simulation
                Lincoln Welch

                The vessel I am interested in is a high pressure piston acculumator. Honestly, do not know much about them so my design is more so research.

                  • Re: Pressure Vessel Simulation
                    Jared Conway

                    take a look at discussions about stress concentrations and singularities here on the forum. if your high stress is anywhere near the fixed BC that is likely the cause. you didn't comment on where things are with relation to each other so it is hard to say. but like david says, use BCs that are real. if you're concerned about the sim, check hand calcs first, work on a problem with a known answer and then translate to sim. there are a bunch of examples, tutorials and validation examples in the software that you can look at but most likely this is a setup or expectation problem. you're going to be extremely limited on setup using xpress, and expectation you can deal with using calcs.

                  • Re: Pressure Vessel Simulation
                    Jerry Steiger

                    Lincoln,

                     

                    I'm pretty sure that your problem is because of the fixed ends that you are forced to use becuase you are using SimulationXpress. Check the stress in the middle of the part, or leave one end free and check the stress on that end. You should see numbers very close to the standard equations for a thick walled cylinder with internal pressure.

                     

                    Jerry S.

                      • Re: Pressure Vessel Simulation
                        Lincoln Welch

                        Jerry,

                         

                        That's what I was thinking too. Doing calculations by hand for a thick walled vessel (12" OD, 6" ID) gave stress of 344.75 MPa when at 30k PSI. Running in SW shows a max stress of 483 MPa when I add a stand and fix at the bases as noted in the attached .jpg. Still trying to figure this one out. Any more suggestions?

                         

                        Capture 2.JPG

                          • Re: Pressure Vessel Simulation
                            Jared Conway

                            lincoln, your limited by the tool that you are using

                             

                            by moving the boundary condition away from your area of interest, it looks like you've solved your issue. the stresses are close to what you've calculated by hand, you simply have outliers that are liked right near your fixed condition. in simulation we could use other tools to isolate that stress or we would add more components to the assembly or different conditions to avoid the issue.

                              • Re: Pressure Vessel Simulation
                                Lincoln Welch

                                I believe it is showing the maximum stress on the ID. I cannot see anything near my fixtures that would indicate the high stress (i,e, no red). Am I misunderstanding your analysis?

                                  • Re: Pressure Vessel Simulation
                                    Jared Conway

                                    it might be on the inside. unfortunately with xpress you're limited in your ability to interrogate the model to determine where the high stress is. if you want to post the model, we could look deeper with simulation.

                                      • Re: Pressure Vessel Simulation
                                        Lincoln Welch

                                        I've posted the file in the original topic message. It is a simple cylinder with caps at the ends. I had to keep it as a part and not assembly since I am limited to Xpress. I am just wanting to see a safe wall thickness to operate at 30k psi with a good factor of safety. I've done hand calculations already, now it's just about not letting the machine beat me. Haha.

                                          • Re: Pressure Vessel Simulation
                                            Jared Conway

                                            can you post your most recent one with the feet on it? i think that is the closest to what you want because it moves the restraints away from your area of interest.

                                              • Re: Pressure Vessel Simulation
                                                Lincoln Welch

                                                Jared,

                                                 

                                                That is the model that I have posted. Also, I realize that I am limited in using Xpress. I am just wanting to see what results you guys are able to obtain with a tool more suitable for my simulation.

                                                  • Re: Pressure Vessel Simulation
                                                    Jared Conway

                                                    344.JPG

                                                     

                                                    the results don't change with simulation with the exact same setup

                                                     

                                                    you have very high stress concentrations in the sharp corners, that is where your really high stress is coming from. adding a radius there will bring them down but you'll also need to add mesh controls to bring the effect of the discontinuities down.

                                                     

                                                    overall, an improved number of element across the thickness will help you understand where the stress lies.

                                                     

                                                    i've cut your model in half, applied a symetric boundary condition to limit the size of the problem and used 5 loops of adaptive meshing. you can see that the stress concentration has actually gone up (expected...see singularities here on the forum) but the area that is above your threshhold of 344Mpa is lower than your original model. elimination of the sharp corners and improvements in the mesh overall could help better understand the stress state.