I am an FEA noob in training so I apologize in advance if this is user error (probably is), but I am pulling my hair out. I have an alu 6061 solid frame that I simulated under two loading conditions separately: point load and acceleration (gravity) load. I ran the simulations using Inventor 2019 stress analysis and PrePoMax v2.0.0. The point load scenario shows similar displacements/stresses, but the gravity loading in PPM shows crazy results. See the attached pictures.
Some considerations/ observations:
Due to lack of experience, I probably messed up the setup (GIGO). For example, the mesh does not look quite right in both programs.
I double checked the units.
For both results of the point load, the frame does not deflect as I would expect. I thought the frame would deflect like a cantilever beam, where the face of the free end would rotate with respect to the moment about the fixed end.
I did a quick search in the forum but could not find a remedy. Please point me to the right thread if I missed it.
Apart from the input units, check the dimensions of the geometry imported to PrePoMax. Especially since you are using imperial units. Itâs hard to say more without seeing the model or at least (if you canât share the files) some screenshots with boundary condition and load symbols to know how they were applied.
I tried uploading the PPM file, but I canât because I am a new to the forum and this is my first thread/post. So currently I can only post pictures and one at a time.
I checked the geometry using the query tool and dimensions are correct. I did have to scale down the model after import, but I did so prior to setting up the simulation.
I took some screenshots (attached) of the boundary condition and load. The only BC I have is a fixed constraint at the base. Gravity is 15 gâs in the x direction.
You can upload the .pmx file(s) to some hosting website like Goolge Drive, Dropbox or WeTransfer and paste the link here.
It would be also helpful if you shared the indended global dimensions of the frame, load values and Youngâs modulus. Those are typically incorrectly defined in such cases when there are orders of magnitude differences in results.
That didnât even cross my mind; here is the link to the file:
If you canât open the file for whatever reason:
Frame is 22" wide (x) x 46" tall (y), 26" deep (z). Legs are 1" wide x 2" deep. Top and bottom thickness is 1".
Acceleration is 15 gâs in the x direction.
I selected the material from the PPM library. Youngâs modulus and other properties seem to be consistent with alu 6061.
The error seems to be how I am applying the acceleration load in PPM. I did point load and distributed load scenarios as an experiment, and both results were in the ballpark of my Inventor 2019 simulations.
It seems to be the fault of the density unit. Iâm not familiar with the imperial system but I know that density units become particularly tricky when a consistent imperial unit system is required in FEA. It seems to be the same problem as here: Centrifugal Loadings - Imperial Units Error? - V1.4.1+
and, for whatever reason, dividing the density by 386.1 gives reasonable results.
Iâm glad to hear that. Youâre definitely more familiar with the imperial units so what do you think - should the materials in the PrePoMax library all have their densities divided by 386.1 or does it depend on the case ?
Yes, I think the imperial density values should be updated AND so should the units. 386.1 is a conversion factor (1 lbfâ s^2/in = 1 âblobâ = 386.1 lbm) so once you divide you are no longer dealing with lbm. It would also be a kind reminder that the user is in the realm of imperial inches, and thatâs what the program is looking for. Ultimately, I should have caught the anomaly on the very first step of creating my new PPM analysis since it shows you the base and derived units.
Separately, what is interesting is that with the point load and pressure load scenario, I got similar results to Inventor without correcting the gravityâŚ
I think so (maybe apart from the specific heat unit which may also have to be changed) but someone who actually uses those units should confirm this just in case.
Then thereâs also the US Unit (ft) system (with ft and slug instead of in and blob) but I guess we have enough fun with the currently available imperial system
I looked up the blob term and it seems it is not very common. slinch might be a more common one. But I do not know.
If I use lbf¡s²/in for mass and lbf¡s²/inⴠfor density I would have to use lbf¡in /([lbf¡s²/in]¡°F) = in²/(s²¡°F) for specific heat. Does the unit in²/(s²¡°F) makes any sense?
It should be consistent since in SI units J/(kg ¡K) converts to m²/(s²¡K).
I have changed the default mass unit in the new PrePoMax version. I have no idea if all units are correctly addressed so please try it out. Since the mass unit changed, I had to change the density unit, specific heat unit, Stefan-Boltzman unit and Newton gravitational constant unit.
