Why the symmetrical structure under self weight has unsymmetrical deformation?

[netsonicyxf]

It's a two span bridge structure modelled by shell elements. The structure itself, boundary conditions are symmetrical about the bridge axis. But the vertical deformation under self-weight is unsymmetrical.
I seperate the model into two sub-models, one for superstructure and the other for substructure.
Both sub-models shows symmetrical deformation under self-weight. Then why the original structure has unsymmetrical deformation?

This is the unsymmetrical vertical deformation of original structure under self-weight

This is the symmetrical vertical deformation of superstructure under self-weight

This is the symmetrical vertical deformation of substructure under self-weight


The attached are zipped sap2000 models.

 

[Denial]

I have never been a SAP user, so I cannot offer any specific comments. But if you genuinely have a symmetrical model and you genuinely apply a symmetrical load to it, then all aspects of the analysis results should exhibit the same symmetry. If your results do not show symmetry then your structure and/or your loading and/or your boundary conditions cannot be symmetrical. But of course you know this, which is why you lodged your post.

If I had this problem I would perhaps create a sub model by ripping off the right hand side, replacing it with a plane of symmetrical boundary conditions, then analyse that. It might reveal something. Then if necessary repeat the exercise by ripping off the left hand side instead. Maybe your SAP software has some direct or indirect means of subtracting one set of output actions from the reflection of the other set of output actions, which MIGHT draw your attention to where the original full model is not truly symmetrical.

 

[SWComposites]

there has to be something about the mesh, material properties, loads or boundary conditions that are not symmetric. show us pictures of the loads, boundary conditions, etc

 

[netsonicyxf]

There are 53 post-tensioned tendons in the superstructure. All the tendons have the same material, jacking force and losses, but with different geometry profiles.
All the tendons are paralle to the XZ plane. Tendon 1 and 53 are symmetrical about the XZ plane, then Tendon 2 and 52, then Tendon 2 and 51, and so on. Finally the tendon 27 is in the XZ plane.

The tendon 6 and 48 are culprits. Their y-coordinates are -6.912m and 6.912m, respectively. If I put them in the model according to their origianl y-coordinates, Sap will give me unsymmetrical results under self-weight and other loadcases. But if I change the y-coordinates to -6.91 and 6.91, or -6.92 and 6.92, then the result are symmetrical.