well, that looks consistent with the stress level you got. Maybe the tension part in the concrete is taking a significant amount of bending moment, or the hand calc method is too conservative (neither I am familiar with Civil Engineering typical figures)
1 Like
I just noticed in your inp . Only Compression material is like only compression support. It’s NLGEOM material behaviour.
Yeah, but the stress in concrete is correct. That’s the strange part.
Right, makes sense. I enabled Nlgeom but it didn’t help:
The deflection is 1.997 mm now.
I’m starting to think that the analytical solution is wrong but, out of curiosity, I asked ChatGPT to solve this problem and it gave me the following results:
- normal stress in each reinforcing rod: 160 MPa
- maximum normal stress in concrete: 8.33 MPa
So pretty close to Hibbeler’s solution.
You could plot the axial stresses thru the thickness to check if neutral axis position is in accordance with book calculations.
This can be indicated of analytical calculator from book reference is violating equilibrium. Couple force of tension and compressor is not in balance, but section force in CalculiX doing by integration for each lqrge element mesh and reported properly.
I have review the procedure I followed with the T some months ago when learning some more about the Mohr Coulomb material model. The analytical calculation was more rigurous and based on EuroCode. Agreemen was very good then so there was no reason this should not work. I like very much your BC. It’s simpler than mine and provides even cleaner results.
Let me clean the inp and I will send it later.
Regards
I’ll appreciate it. So you got the right values ? Cool. Was there anything special that you had to change to get there or did it just work right away when you modeled it in Mecway ?
It works right away except I’m using hexa elements (not tested with tets) and custom material cards. Mecway doesn’t have only compression
¿Are you doing a tutorial?. If yes I will send it to you. This could bring many users to try ccx/Mecway/Prepomax.
I switched to hexa as well (C3D8). With some wedges for filling.
That’s interesting. I wonder how you defined it. Maybe CalculiX’s compression only is not working well here.
Sure, it’s for another PrePoMax tutorial.
Really interesting, section force of solid surface reported properly but it’s not for stress. How could it be, made curiosity.
It’s an isostatic beam, internal forces are obtained from equilibrium equations only, however the stresses depend on the internal arrangement and the material models.
The neutral axis looks in the right position, consistent with book analysis. However in Calc_em’s model looks different but difficult to say from the figure.
Material model, of course but this threads discussed of compression only material not something else. A simple beam subjected to pure bending without axial loads, internal tension resisted by rebars and compression by concrete, there’s couple force from stress and areas to balance in equilibrium.
when peak concrete stress, neutral axes and section force are reported properly, it should be in rebar stress also.
p.s I’m not yet to open and running input file attached since to large, not reproduced the problem also, only following result from discussion and some pictures provided.
i do simple check, these values from reference books are in equilibrium since couple force of tension and compression in balance (T=C=166.84kN). So, the problem in model and probably due to over-constraint by rigid body. However, i’m not sure the reason of peak compression stress, depth of neutral and section force are reported as expected but not for rebar stress.
*edited (updates)
it’s not good to leave the treads unclear for just simple case, i decided to test by reproduced and removing one translation along beam axes seems to solve the problem.
You are right, thank you. I tested it before when Disla mentioned it but then I kept doing more tests with previous supports. Now when I remove that one translation, set PR=0, use 0.1 MPa for max allowed tension (0.01 MPa doesn’t converge) and enable Nlgeom, the results are pretty close to the analytical ones:
When I compute the overall Bending Moment on the midsection through the nodal forces, the value is very accurate 59.99 KNm
If I request ccx to compute it by Section Print the result is completely wrong. ¿Are you experiencing the same.?
statistics for surface set MCR and time 0.1000000E+01
total surface force (fx,fy,fz) and moment about the origin(mx,my,mz)
-3.209614E+05 1.220696E-10 -1.458681E+01 -7.604734E-11 1.507528E+05 -7.649881E-12
moment about the center of gravity(mx,my,mz)
-8.917191E-11 1.162439E+05 -8.890390E-12
area, normal force (+ = tension), shear force (size), torque and bending moment (size)
1.350000E-01 3.209614E+05 1.458681E+01 8.917191E-11 1.162439E+05
EDITED: Solved. Section Print doesn’t work properly when sections of different materials share commond nodes.
it can be simple for uniform moment along the beam length, but can be problematic for non-uniform since the inertia of cracked or transformed section vary. An approach by element division and effective inertia assign can be use, beam length may divide by three or six segment.