SAFETY FACTOR FOR WIND RESISTING BALLAST CALCULATIONS

I would like to get help on the following topic:

We have tested the aerodynamic characteristics of a ballast-only-supported-structure in a Wind Tunnel.

The structure is to be installed on top of buildings.

For sliding and overturning calculations: Which safety factor on Loads and/or ballast should I use (Hint: ASCE 7-98, which allows usage of Wind Tunnel for load calculation does not explicitely mention a safety factor for sliding and overturning. I am reluctant to use the same safety factor that ASCE recommends for strength/stress calculations because the level of uncertainty related to stability-only analysis is much less : for example the rate of load application does not play a roll in stability analysis , while it plays a significant one in strength/stress analysis)

THanks

 

[dlew]

sanfro,

The safety factor against sliding and overturning specified in several building codes is 1.5. Normally only 90% of the dead load is considered when checking for overturning, to account for any overestimation of the dead weight.

This safety factor is consistent with ASCE 7 load combination factors:
For Strength Design, 0.90D + 1.60W + 1.6H
For Allowable Stress Design, 0.6D + W + H

where D = dead load, W = wind load, and H = lateral soil pressure or ground water pressure.

AEF

 

Thanks AEF,

That is the way I approached it and end up with a "safety factor " of 1/0.6 = 1.67.

Still I am struggling because in the case of (for example) sliding analysis , once Wind Tunnel is used, the sources for uncertainty are:

* Wind Tunnel uncertainty in the order of 10% forces and 5% moments

*Friction coefficient uncertainty (This is ballast only, i.e., no anchoring points) which we reduce to negligible by measuring case by case (many trials) friction coefficients in the roof we are going to install the structure.

*Ballast level uncertainty (the weight you put vs. the weight you think you put)(10% max.)

...and we are using the 3sec gust wind velocity as the steady-state one to determine the acting loads.

If you compare this analysis with a ultimate strength or alloable stress analysis you see that in the later the mirroring uncertainties to those above are (I think) much bigger , specifically the one related to the lower effective strenth against rapidily applied loads.

I am affraid of being too conservative here.