Understanding bending moment diagrams

[Kerdni]

Hi

I'm having some questions about understanding bending moments from panel cuts.

Background information:
I am designing a three span rigid frame concrete bridge that has a deck with variable depth. The bridge is modeled of concrete panels. Variable depth is achieved giving different thicknesses to the panels. Materials are chosen the same as they will be used in the capacity calculations.

Problem description:
Calculating a 3D model I get bending moments (MYY) at the intersection of the inner pier and the deck that are distributed as shown in picture_3D_1 and picture_3D_2.
picture_3D_1 - no smoothing is used
picture_3D_2 - global smoothing is used - this one does not seem right at all.

Calculating a 2D model I get bending moments (MYY) at the same location that are distributed as shown in picture_2D.

From picture_3D_1 the sum of moments is:3595,07-2499,31-670,67=425,09
From picture_2D the sum of moments is:546,98-288,07-258,91=0

Questions:
1. Shouldn't the sum of moments in the 3D model be 0 also?
2. If so, how the problem could be fixed?
3. If not, wich moments should I choose from the 3D model for determining the usage of rebars?

Link to the 3D model: http://we.tl/RQYsnQlk9I

Best regards.

[Rafacascudo · Accepted solution]

In this case (panel cuts) , results get a lot better as you refine your mesh. If you reduce your mesh from 0,50m to 0,25m, your 425 difference will go to half(212). If you reduce it to 0,05cm(only in the adjoining panels)  , it will go to only 42kn.m/m.

So it seems to be directly proportional.

Best results here were achieved  keeping 0,50m mesh for the other panels and for the panels at the intersection I defined them also with 0,50m(Delaunay) but with emitters at the intersection nodes. delaunay and Kang with H0=0,05cm

Difference went to 24kn.m/m

[Kerdni]

Thank you! Problem solved.