Load distribution on concrete slab

[lars_hatlen]

I have a two storey concrete structure above an open garage. One side of the building is very stiff with concrete walls and an elevator shaft. The other end of the building consist of steel columns in wooden walls that comes down on a concrete beam - column system in the garage roof. Building floors are concrete slabs resting on concrete walls, elevator shaft walls and concrete columns. When looking at reaction forces in steel columns and some of the concrete walls in the back around the elevator shaft somehow I get an uplift instead of downward facing forces. It looks like the elevator shaft "steals" forces through the concrete slab. Some force distribution towards the shaft is expected as there is it a stiffness center but the final results are not making sense. Any idea as to why this happens? Steel columns are pinned-pinned and slab-wall connections are fixed. 

[lars_hatlen]

See picture of an example building of a concrete slab supportet by concrete walls and steel columns. The forces shown are Fz (vertical reaction forces). Could someone please try to explain to me why the load distributes itself this way on the walls? I struggle to understand how the self weight and a 5 kN/m UDL on the slab would result in an uplift in the corner of the concrete walls.

[Romanich]

Hi @lars_hatlen ,

Please zip and attach your model

[lars_hatlen]

Here is the zip file. Top floor is a wooden structure. I have used solid wall shells with reduced stiffness to simulate stud wall and wooden beam roofs stiffnesses. If you have any tips on have to modell this in a better way, please let me know. Thank you in advance for any help!

[Romanich]

Please check these two articles:

https://www.autodesk.com/support/technical/article/caas/sfdcarticles/sfdcarticles/Negative-reactions...

https://www.autodesk.com/support/technical/article/caas/sfdcarticles/sfdcarticles/How-to-neglect-ten...

[Rafacascudo]

In this picture/model the applied loads would make the corner of the slab lift from the wall support if it was not attached firmly to it. Being attached, it transfers the uplift to the wall which generates the negative reactions down below. To avoid these negative reactions , you have the solution already proposed by @Romanich . Another option , if the the slab is not firmly attached to the walls, is to set linear UZ+ releases(unilateral) on those slab edges

[lars_hatlen]

Hi, 

I have tried to use the proposed solutions from you and @Romanich but struggle with the calculations not converging. To solve this I have tried to implement the content of this article without success: 
https://www.autodesk.com/support/technical/article/caas/sfdcarticles/sfdcarticles/ROBOT-No-convergen...

When neglecting negative reaction forces the modell still faces a problem when it comes to stabilizing self-weight against horisontal earth pressure. As the modell behaves now, only part of the concrete walls self-weight actually end up as vertical reaction force at the bottom of those walls. When concrete slabs are added to the modell the uplift is created. It seems that the concrete slabs redistribute loads away from the walls in stead of using the wall as a support. Is there any other way to work around this problem or do I have to use UZ+ releases and non-linear analysis. If so, any tips with the convergence issue other than what is in the article.