Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations TugboatEng on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
Effect of "Large deformation ON" 1
- Thread starter Haval
- Start date
- Thread starter
- #2
rickfischer51
Mechanical
What does the Ansys help manual say? Perhaps you should try it both ways on a simple beam model
Rick Fischer
Principal Engineer
Argonne National Laboratory
Rick Fischer
Principal Engineer
Argonne National Laboratory
-
1
- #4
I think the formulation is changed (as when you do a simple F.E. calculation on paper with small deformation and neglect 2nd+ order parts in your integration).
Plus, using large deformations, your load (only if you apply a pressure), can follow your deformation.
For example, say you model a 3D beam along X axis. The "vertical" is along the Z axis. You define a section at the end of your beam to apply a load along Z axis.
If you don't take large deformations into account, your load will be along Z axis even at the end of the study, whatever the deformation is.
If you take into account large deformations and use a force / remote force / nodal force, the load will still stick to the Z-axis, but if you use a pressure instead, the resultant load will be normal to the area of application of the pressure.
This is why a solve often uses iterations when taking large deformations into account.
(This part at least is explained briefly in the Ansys help)
Plus, using large deformations, your load (only if you apply a pressure), can follow your deformation.
For example, say you model a 3D beam along X axis. The "vertical" is along the Z axis. You define a section at the end of your beam to apply a load along Z axis.
If you don't take large deformations into account, your load will be along Z axis even at the end of the study, whatever the deformation is.
If you take into account large deformations and use a force / remote force / nodal force, the load will still stick to the Z-axis, but if you use a pressure instead, the resultant load will be normal to the area of application of the pressure.
This is why a solve often uses iterations when taking large deformations into account.
(This part at least is explained briefly in the Ansys help)
- Thread starter
- #5
- Thread starter
- #6
Dear Orel,
Thanks a lot. So if I want to simulate a three point bend test as I am doing now, I have to chose either Large deflection = OFF or to use remote force. As experimentally, the compact head of the machine (which is doing the bending)still vertical during the test operation?
I am doing a simulation to what I did experimentally in a three point bend test for an arrangements of beams. a three by three bundle of beams with a cross section 4 mm* 4 mm and a length of 250 mm.
Thanks in advance.
Regards,
Thanks a lot. So if I want to simulate a three point bend test as I am doing now, I have to chose either Large deflection = OFF or to use remote force. As experimentally, the compact head of the machine (which is doing the bending)still vertical during the test operation?
I am doing a simulation to what I did experimentally in a three point bend test for an arrangements of beams. a three by three bundle of beams with a cross section 4 mm* 4 mm and a length of 250 mm.
Thanks in advance.
Regards,
Hi Aval,
I don't think you need a remote force for what you are trying to do.
Also, as you describe it, I think you can use Large D OFF, but as you have experimental results and a model that should run fastly, it's a good case to test the two configurations and compare them for a good understanding.
I don't think you need a remote force for what you are trying to do.
Also, as you describe it, I think you can use Large D OFF, but as you have experimental results and a model that should run fastly, it's a good case to test the two configurations and compare them for a good understanding.
rickfischer51
Mechanical
Beam bending in FEA is based on basic beam theory. Get your basic stress analysis textbook and look at the part about the differential equation of the elastic curve. 1/rho = (Y")/(1+(y')^2)^1.5. If the deflection is small, (y')^2<<<1, so 1/rho = y". then EIy"=M(x). If your defleection is not small, then it gets nonlinear. Run your analysis with NLGEOM off and look at the deflections. If the deflections are small (typically less than one half the element thickness, then you are OK. Otherwise, turn NLGEOM on and run again and see what changes. Again, try it both ways on a simple model and compare results. Make a cantilever beam model, and keep increasing the load with NLGEOM on and off until you see a change in the results.
Rick Fischer
Principal Engineer
Argonne National Laboratory
Rick Fischer
Principal Engineer
Argonne National Laboratory
- Thread starter
- #9
- Thread starter
- #10
Hi Rickfischer51,
Do you mean the deflection equation of a simply supported beam (delta= FL^3/48 EI)? and about the deflection if it is small or not, I think its more than the half of the element thickness. The problem is that, I need the applied force to remain vertical during the solution to simulate the real case of the experiment and if I chose LD= ON then the applied force will start to incline with the deformed body as Orel explained previously.
Do you mean the deflection equation of a simply supported beam (delta= FL^3/48 EI)? and about the deflection if it is small or not, I think its more than the half of the element thickness. The problem is that, I need the applied force to remain vertical during the solution to simulate the real case of the experiment and if I chose LD= ON then the applied force will start to incline with the deformed body as Orel explained previously.
rickfischer51
Mechanical
1. It doesn't matter if you model a simple beam, cantilever beam or the Eiffel Tower. Make a simplified model of something and run it with NLGEOM on and off and plot deflection. After awhile, the two results will diverge.
2. If the issue here is to make modeling easy, then set NLGEOM to anything you want. If the issue is to get an accurate result, then you need to turn NLGEOM on if your deflections are sufficiently high. Per the manual, NLGEOM includes large deflections, large rotations, large strains, stress stiffening.
Rick Fischer
Principal Engineer
Argonne National Laboratory
2. If the issue here is to make modeling easy, then set NLGEOM to anything you want. If the issue is to get an accurate result, then you need to turn NLGEOM on if your deflections are sufficiently high. Per the manual, NLGEOM includes large deflections, large rotations, large strains, stress stiffening.
Rick Fischer
Principal Engineer
Argonne National Laboratory
- Thread starter
- #13
- Status