Contribution of Roof Ballast to Effective Seismic Weight

[dbnerds]

Friendly debate in our structural design group:
Do you consider roof ballast as part of the Effective Seismic Weight?

-some say yes, since it IS a dead load and their are no caveats/exclusions for it per code.
-some say "maybe not" since it may act like ball bearings (i.e. not stay put in their original location) during a seismic event and thus not fully contribute to the pendulum/oscillating momentum of the structure.

Anybody seen any texts refer to this or have an opinion of their own?

Thx.

 

[dhengr]

Use the ball bearing analogy if you wish, but assume frozen bearings, and a multiplier of 1 on the weight of the ball bearings.

 

[msquared48]

The same could be said of live loads, but any uniform live load over 100 psf has to be factored into the seismic, even though it may not be stable, as well as any snow load over 30 psf.

Add it in.

Mike McCann
MMC Engineering

http://mmcengineering.tripod.com

 

[structSU10]

agree with above - even if the object is moving around at a different frequency its still a mass unless if flies off the building.

 

[PicoStruc]

msquared48 : The live load is factored with the seismic load to estimate force effects but the live load is NOT included in seismic weight !

 

[PicoStruc]

Note : I am only refering to canadian code! Not sure about US codes though !

 

[MiketheEngineer]

Its a dead load - consider it.

 

[MiketheEngineer]

Like plywood and shingles

 

[msquared48]

picostruct:

Sorry, but for live loads over 100 psf, like storage loads of 125 or 250, you DO have to include a portion in the seismic calcs - I believe it is either 25 or 30% of the total love load.

Mike McCann
MMC Engineering

http://mmcengineering.tripod.com

 

[JAE]

since it may act like ball bearings

MAY act - if you don't know its usually better to be on the safe side right? How would you ever know if it does indeed roll about adequately?

Even if it did roll about some, how much weight do you discount?

And if you choose a certain path, which one is closest to what a reasonable engineer would do in similar circumstances?

 

[PicoStruc]

@msquared48 oh, The seismic mass is different in Canada then !!!

Here, the seismic mass is defined as
- 100% Dead Load
- 25% Snow Load
- 100% Reservoir weight capacity
- 60% Storage Live Load, Parking is not considered as storage.

Any other live load is not included in seismic mass. Of course as engineer if I have a heavy equipment anchored to the floor... I will accout for it !

 

[dbnerds]

@JAE
good question about how much do you discount.
Lets look at definition of Effective Seismic Weight inclusions -- like the snow one.
(I'm quoting AISC) ..if snow load > 30 psf, then 20% of that load can be considered part of the effective snow load.

I would argue that the reasoning of that is that the code (writers) believe the shaking of the seismic event will disrupt the "layering" of the snow and the top (80% in the AISC case) would be not positively connected to the building structure.

I suppose thats how I'm seeing this question:
If an object isn't positively connected to a structure, can it count as the effective weight?
Water in a silo/tank can be counted because its contained within walls.
If the water was standing/ponding on a roof, would you consider it effective seismic weight? NO.
Can't that be carried onto other objects that are not CONNECTED to the building structure?

which comes back to JAE -- HOW MUCH??
perhaps it is a function of dynamic coefficient of friction?

 

[JAE]

I think you are diving way to deep into the theoretical here and beyond what the code even suggests.

Snow is discounted more for statistical probability of being at design depth at the same time as a major seismic event...not that the snow somehow rolls about on the roof.

 

[dbnerds]

Sometimes (way too often) the code doesn't cover some subject matter and you have to use engineering judgement (theoretical as you said).

If the snow was discounted for "statistical probability" purposes we should be looking at load combinations, not the discount of the load itself, depending if its over a certain level (ie 30 psf).

 

[connect2]

my understanding is that the probability is not high that you will get a max seismic event with a max snow load, but that in areas of sufficient snowfall (winter) it is possible so that 25% of the snow is included. I suspect if you live in the Artic it would be wise to carry more than 25% of the snow in the calc of effective seismic weight.

 

[msquared48]

Especially considering layered ice crusting of the snow with intermittent snowing, freezing and warm/rain.

Mike McCann
MMC Engineering

http://mmcengineering.tripod.com