6 Creep effect on normal force
Task
During recent calculations we found out that the definition of creep factors of concrete slabs has a surprisingly high effect on normal forces in steel columns. The more statical undetermined a system is, the higher difference in normal forces can be expected.
In statical determined systems this effect does not exist.
Statical system
10 storey high building, floor height 3m
Slab: C35/45 (Norway B35), XC1, thickness 200mm
Geometry slab: 16 x 2m
Steel columns designed for 80% utilization in FEM-Design nocreep
Loads
selfweight
permanent load: 1 kN/m2
live load: 2 kN/m2
loadcombinations
ULS: Sd = 1,35 gk + 1,5 qk
SLS: Sd = 1,0 gk + 1,0 qk
Results
Results are calculated with these programs.
FEM-Design 19
no creep in slabs and walls
creep 2,5 / shrinkage 0,36
Focus Konstruksjon 2021 (2D)
creep is not taken into account during ULS calculations, just for SLS
Calculation is done with normal creep definition and reduced creep
OS-Prog K-bjelke
1D calculation / creep is not taken into account during ULS calculations, just for SLS
Calculation is done with normal creep definition and reduced creep
The results are shown in the following 3 diagrams.
Normal forces of center steel column for the whole structure.
Reaction force of center support (= steel column) per floor.
Maximum deformation of slab for the whole structure.
Summary
There is a considerable effect of creep factor of the slab on the normal forces of a slab.
The reason for this behaviour is that in finite element calculations, forces always tend to move to the support with the highest stiffness. The more stiff a construction behaves, the higher this effect tends to be.
In structural systems where there is one part that is much stiffer than another - like concrete walls vs. steel columns, there is a high risk of underdimensioning of the weak part - the steel column.
In this calculation, the difference in normal forces in FEM-Design between the structure with and without creep definition in the slab is as high as 18%. Comparing the most extreme values of this example we get differences as high as 30%.