Solenne, Best case, it will likely be a couple of days before I can dive deeper into the files you sent.
Looking at your latest message, I was wondering what kind of strains you are seeing in each of these models with ABAQUS. Is your enclosed volume changing significantly, similar to inflating a balloon with the large pressure and not significantly with the lower pressure? If so, that would mean that the surface area of the elements significantly changed with the second pressure, which might narrow down where I need to start looking. Could you share a few images from ABAQUS of the undeformed vs. deformed shape (true scale, not exaggerated) for each of the two pressures for additional visual context. Maximum principal strain contours from ABAQUS might be helpful as well. Best Regards, Mike On Wednesday, January 8, 2025 at 3:31:55 AM UTC-6 SoMdt wrote: > Dear Mike, > > > > Thanks you very much for your answer, here are our latest advancement on > the topic. > > To answer in the order : > > - I tried both non linear static and dynamic tests and obtained the > same results > - I specifically launched a simulation with a very fine mesh + very > small timestep, but this did not improve the results. > - Eventually, we implemented ANCF_3833 elements, but the result is the > same as 3423 and Reissner. > - I also tested using SetStiff(true) for the pressure load but the > results are the same. > > > > I am thinking we might have missed something in the implementation. > > > > To complete my message, here are all the results for the maximum > displacement in mm, both in Abaqus and in Chrono for two different values > of pressure (p=0.001MPa -> small deformation, p=0.03MPa -> large > deformation): > *Abaqus * p=0.001 : d= > 1.51e-2 p=0.03 : d= 7.81e-1 > *Chrono - Shell Reissner * p=0.001 : d= 1.46e-2 > p=0.03 : d= 4.18e-1 > *Chrono - ShelANCF_3423 * p=0.001 : d= 144e-2 p=0.03 : > d= 3.73e-1 > *Chrono - ShellANCF_3833* p=0.001 : d= 1.42e-2 p=0.03 : > d= 3.73e-1 > > > > I join you all the simulation files (cpp + mesh + nodesets for boundary > conditions). If you have time to look into it, I would be immensely > grateful: we are currently completely blocked in our project due to this > issue. > > Best regards, > > Solenne > > > Le jeudi 12 décembre 2024 à 23:51:13 UTC+1, Mike Taylor a écrit : > >> Solenne, >> >> I'm most familiar with the ANCF elements. The ANCF formulation supports >> both geometric and material non-linearities. Currently most of the ANCF >> elements in Chrono are formulated with only a linear viscoelastic material, >> but they do inherently support moderately large deformations (i.e. large >> angle bending, but not plastic like flows.) >> >> Here are a few thoughts on things to try: >> >> If you've only tried to run a static analysis, you might want to >> try running a dynamic analysis with a slowly applied load with a relatively >> large amount of damping to dynamically settle to the "static" solution? >> >> If you are running a static analysis, are you running the linear static >> or nonlinear static analysis version? If you are using the nonlinear >> static analysis, what argument are you giving that function? >> >> If you haven't already examined refining your mesh (using more elements), >> you might want to give that a try as well. >> >> Finally, the ANCF Shell 3423 element has a relatively low order (linear) >> interpolation function. You could also try an element with a richer/higher >> order interpolation function like the "ChElementShellANCF_3833"element. >> >> Best Regards, >> >> Mike Taylor >> >> On Tuesday, December 10, 2024 at 10:26:15 AM UTC-6 SoMdt wrote: >> >>> Hello, >>> >>> I am implementing a simple test of a cylinder with fixed extremities >>> under pressure (length of 42 mm, radius of 2.5 mm and thickness of 0.6 mm). >>> Material has a Young's modulus of 0.6 MPa and nu=0.45. >>> I tested ShellBST, ShellReissner and ShellANCF4234 in chrono and >>> compared the results with Abaqus. >>> >>> With a pressure of 0.001 MPa, the results are the same as Abaqus. >>> However, when increasing the pressure to 0.03 MPa, the deformation of the >>> cylinder is roughly twice lower in chrono than in Abaqus (results are in >>> the same magnitude for all 3 shell elements). >>> >>> To completed the analysis, in Abaqus, when setting the NLGEOM parameter >>> to NO, the deformation goes down to the range of chrono deformation. >>> >>> My question then : are there shelle elements able to consider large >>> deformations ? Did I forget a hidden parameter somewhere ? >>> >>> Thanks a lot for you help, >>> Solenne >>> >> -- You received this message because you are subscribed to the Google Groups "ProjectChrono" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/projectchrono/d580c19c-4aff-48e2-822f-b848fef0c85bn%40googlegroups.com.
