API-650, Appendix R, Equation (g)

[gdubyuh]

Addendum 1 of API-650 11th Edition includes a new equation (g) in Appendix R. This equation combines the dead, live and external pressure loads attributed to an internal floating roof (IFR) to the dead, live and ext. pressure loads acting on the fixed roof.

Note 1 in the appendix indicates to add the IFR loads as point loads, since the loads from the IFR are concentrated on the fixed roof at the point where the IFR suspension cables connect to the fixed roof.

I don't know how to apply this equation since the loads on the fixed roof are uniform loads applied to the entire roof surface (pounds per square foot) and the IFR loads are point loads (pounds). I have inconsistent units. Advice?

Also, the variable (T) for equation 5.10.6.1 still indicates that it should be the larger of Appendix R equations (e)(1) and (e)(2). It was not expanded to read "or greater of (g)(1) or (g)(2) if applicable", which makes me wonder if equation (g) is intended to be applied to umbrella roofs.

Typically we suspend our IFR's by cables attached to the bottom flange of rafters on rafter supported roofs, but was curious how to apply the equation (g) if suspending the IFR directly to the roof plate of an umbrella roof.

 

[IFRs]

I use that particular load combination when analyzing or designing the rafters of a supported cone roof. I put one cable 5 feet from each end of the rafter and one cable in the middle of the rafter. This seems to reasonably represent the actual load condition, based on cable layouts I have seen and designed. If the rafter fails I reduce the center load by 1/2 and specify that the cables be attached between rafters in the roof plate, possibly with jack rafters spanning the gap. For umbrella roofs I would do a global and local buckling analysis of the curved plate using classical shell buckling theory and/or FEA and/or testing.

 

[JStephen]

I think the intent is that these are load combinations, not equations for calculating a uniform load. You can combine the effects of a factored point load and a factored distributed load without adding the two loads together first. Or in the case of seismic loading, note that the E and the DL terms will be vector quantities at right angles to each other, and are not just two numbers added together.

 

[TankDude]

gdubyuh,
IF you connect your IFR cables to the bottom flange of your fixed roof rafters, how do you adjust from high to low leg setting? Just curious.
I agree with the method IFRs describes. In fact, I typically obtain the actual proposed layout from the IFR manufacturer to determine where (worst case) the point loads will intersect the rafters. From there it's simply a beam analysis problem.
By the way, the required loading combinations are very conservative and when coupled with often necessary assumptions regarding existing roof structure material strength calculated bending stresses may push the allowable limit.