TLHS
Structural
- Jan 14, 2011
- 1,600
I have a fun situation. We are looking at the foundation design for a very large moment frame supported fabric shelter (no sidewalls) on a site adjacent to the water with poor ground conditions. The geotech folks are going to mitigate some of the liquifaction issues, but it's still not a pristine site. By my read of the National Building Code of Canada, the footings for this structure need to be tied together to avoid differential foundation movement.
There is an axial force requirement in the code. My issue comes from the span in question. The superstructure has a portal frame span of 20+ meters. That means that in one direction my tie beams will run that long. A slab on grade is not preferred from an economical standpoint, if it can be avoided, so there's no floor diaphragm. For any practical size of tie beam, this is a slender compression member. Realistically, though, how would I examine this? If the soils are competent, I could probably calculate out enough bracing capacity to keep it from buckling from selfweight friction and a bit of soil pressure. Counting on the soil to be competent, though, feels like it goes against the purpose of the tie beams. ACI and CSA A23.3 seems to have a maximum required size (450mm). If I use that, do I just get to assume there isn't a buckling issue? If I even wanted to consider buckling, it doesn't seem like there'd be a rational basis to use since this isn't a purely analytical design basis. How do you do a P-Delta analysis when there's an undefined lateral displacement that could be happening at the same time? I could attempt to brace the footings against skew to limit p-delta similar to how a floor diaphragm would, but I run into the same issue of having no real load basis for the design of that bracing, since it isn't really just bracing against buckling, it would be bracing against some sort of undefined differential ground movement.
There is an axial force requirement in the code. My issue comes from the span in question. The superstructure has a portal frame span of 20+ meters. That means that in one direction my tie beams will run that long. A slab on grade is not preferred from an economical standpoint, if it can be avoided, so there's no floor diaphragm. For any practical size of tie beam, this is a slender compression member. Realistically, though, how would I examine this? If the soils are competent, I could probably calculate out enough bracing capacity to keep it from buckling from selfweight friction and a bit of soil pressure. Counting on the soil to be competent, though, feels like it goes against the purpose of the tie beams. ACI and CSA A23.3 seems to have a maximum required size (450mm). If I use that, do I just get to assume there isn't a buckling issue? If I even wanted to consider buckling, it doesn't seem like there'd be a rational basis to use since this isn't a purely analytical design basis. How do you do a P-Delta analysis when there's an undefined lateral displacement that could be happening at the same time? I could attempt to brace the footings against skew to limit p-delta similar to how a floor diaphragm would, but I run into the same issue of having no real load basis for the design of that bracing, since it isn't really just bracing against buckling, it would be bracing against some sort of undefined differential ground movement.
