Connection of storm shelter to surrounding builidng

[byoung8708]

I am trying to design a storm shelter inside a school per ICC-500. I am using the corridor walls with a concrete slab on top as the shelter. There are light gauge metal trusses spanning between exterior classroom walls, but I need to use at least one of the corridor walls (or the concrete slab) as a bearing wall for these trusses (gravity load support as well as uplift). Section 308 of the ICC-500 is not very clear on how to design these connections. It seems to say that the connections should be designed stronger than the tornado winds, but that seems odd to me (as I would think you want the surrounding building to disconnect with the storm shelter). Any help or clarification about this section would be very helpful.

 

[JAE]

I read 308. It seems to be saying that the "framing" of the host building that is directly connected to the shelter must be designed for the severe storm wind forces.

So in your case, it seems to suggest that your light gage trusses need to be designed to the higher wind speeds. Also the connections of the trusses to the shelter.

This seems counter-intuitive to me. I've always thought of the connecting building as a "tear away jersey" (for those of you who remember US Football back in the 70's).

The concept is that during a storm, the surrounding building just gets sucked away and the LAST thing you'd want is that collapsing building to be significantly fastened to your storm shelter.

The older FEMA 361 does discuss the idea of internal shelters a bit and includes a design example for a Kansas classroom addition - the details don't seem to imply that the connection between the addition and the new school were designed for the higher loading.

 

[paddingtongreen]

The interpretation is that everything necessary to support the shelter must be design for the purpose, it is not disposable, but everything else is.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[JAE]

Huh?

That's not what I read. It specifically says that if you have non-shelter framing attached, that framing must be designed to the storm loads.

The non-shelter framing doesn't have to "support" the shelter, or parts of the shelter.

 

[paddingtongreen]

@JAE, sorry, expressed badly. In practice, the load on the framing is so reduced after the cladding is gone that it survives, it must be designed to resist the loading until the fabric tears away, after that, the load on the frame is diminished. We limited the "framing" to main wind resisting framing and the shelter support.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[byoung8708]

So are you saying that the trusses and their connections need to be designed for 250 mph tornado loads so that they will stay attached to the storm shelter until the metal deck rips off? Or are you saying that the connections should be designed for that load and the added force from wind on the trusses should be considered in the design of the storm shelter (this would seem more plausible)?