7/30/2013
Snow Loading for Non-Standard
Roof Shapes
Michael O’Rourke P.E. , Ph.D.
Email: orourm@[Link]
Non-Standard Roof Shapes
Itseem that architects no longer feel constrained by
the tyranny of straight lines / rectangular shapes
By their nature, building codes and load standards
can not address all conceivable roof shapes
Laboratory studies ( wind tunnel or water flume ) are
possible but not common practice
CFD and other numerical simulations are possible
but again not common practice
What is the Structural Engineer to do?
1
7/30/2013
Non-Standard Roof Shapes
ASCE 7-10 Snow Load provisions can be
extrapolated to estimate design snow loads for
one of a kind roofs. This webinar will cover six
Mansard Shape
The Wing
Multiple Roof Steps
Roof Top Tent
The Slide
Asymmetric Gable
Non-Standard Roof Shapes
Disclaimer : The roof shapes discussed in this
Webinar are not covered by ASCE 7-10. There is
no guarantee that the suggested roof snow loads
will be accepted by your local building official.
However they are logical extensions of and
reasonably consistent with the ASCE 7-10
provisions for standard roof shapes.
2
7/30/2013
Non-Standard Roof Shapes
Restriction: It is assumed that all roofs consist of
simple supported, single span beams without
overhangs. That is, reducing the load somewhere
does not result in an increase in moment, shear
or deflection elsewhere
Non-Standard Roof Shapes
Mansard Shape
The Wing
Multiple Roof Steps
Roof Top Tent
The Slide
Asymmetric Gable
3
7/30/2013
Mansard Roof
Although the
mansard shape is
fairly common - it is
not currently
covered in ASCE 7-
10
ASCE 7-10 does
cover gable roof
shapes which are in
some ways similar
Mansard Roof
In
ASCE 7, Gable drifts apply when the slope change
(½ on 12 to ½ on 12) is large enough to develop flow
separation – area of aerodynamic shade
Forour mansard, the slope change is flat to 6 on 12 –
large enough that separation and drift formation is
expected
4
7/30/2013
Mansard Roof
In ASCE 7, significant drifts do not form on Gables
with slopes steeper than 7 on 12 – related to
angle of repose for drifted snow
Drift
loads are expected to form on our 6 on 12
mansard slope since roof slope < 7 on 12
Mansard Roof
Fora gable roof , the upwind fetch is the eave to
ridge distance
Forour mansard, the upwind fetch distance would
be the horizontal extent of the sloped + flat
portions that is lu = W = 20 + 75 = 95 feet
5
7/30/2013
Mansard Roof
Trappingefficiency is % of windblown snow flowing
past a aerodynamic shade region that stays
Trappingefficiency for leeward step similar to that for
gable roof drift
Trapping efficiency for mansard expected to be same
Mansard Roof
We are now in
a position to
apply the
ASCE 7-10
Gable Roof
provisions
6
7/30/2013
Mansard Roof
Consider location with Pg=35 psf,
Ce=Ct=Is=1.0 and Cs = 0.67 for the 6 on 12
slope (same as 1 on 2, hence S=2)
Balanced load on sloping portion
Ps=0.7CeCtCsIs = 16.4 psf
‘Gable’ with lu= 95 ft. we have
hd=0.43lu.33(Pg+10).25-1.5 = 3.67 ft
horizontal extent for gable drift
8hd √S/3 = 13.5 ft
Mansard Roof
Forour site with Pg = 35 psf, the snow density γ is
0.13 Pg + 14 = 0.13(35) + 14 = 18.6 pcf
The intensity of the drift surcharge is hd γ / √S
(3.57ft)(18.6 pcf)/√ 2 = 46.8 psf
The gable roof provisions call for 0.3 Ps on the
windward side. However given the large fetch there
is no guarantee that snow on the flat roof portion
close to the downwind mansard will have
significant snow removal. Will use the balanced
snow on the flat portion 0.7(35) = 24.5 psf
7
7/30/2013
Mansard Roof
The estimated design loads are shown below
Non-Standard Roof Shapes
Mansard Shape
The Wing
Multiple Roof Steps
Roof Top Tent
The Slide
Asymmetric Gable
8
7/30/2013
The Wing
Itis unlikely that this
shape will ever be
covered by ASCE 7
The Wing
Inrelation to the
nominally flat roofs
on either side of the
Wing, consider a
rectangular shape
For wind normal to
the wing, we now
have something
reasonable for
analysis
9
7/30/2013
The Wing
Some snow would
be captured at the
windward roof step
and possibly within
the V
However
conservative to use
LA + LB as the
upwind fetch. LB
likely large enough
so it isn’t a RTU
The Wing
Wind front to back –
gable/mansard-like
drift atop balanced
Roof drift starting @
½ on 12 point (gable)
or 1 on 12 point
( mansard)
Upwind fetch W from
eave to ½ on 12 or 1
on 12 point
Windback to front,
same idea with larger
W –all 4 drifts are
separate load cases
10
7/30/2013
The Wing
Distribution across
the ‘V’
Balanced load case
Saw-tooth load case
ASCE 7 has both of
these – should be
checked again as
separate load cases
The Wing
Distribution across
the V
Assuming V is not
too steep(>7 on 12)
gable-like drift at tip
of V is possible
Againexpect drift to
form on lower level
but use LA as
upwind fetch
11
7/30/2013
The Wing
Wing may be
sensitive to
imbalance
Consider balanced
on one side plus
sliding load from the
other
This
imbalance
condition is a
separate load case
Non-Standard Roof Shapes
Mansard Shape
The Wing
Multiple Roof Steps
Roof Top Tent
The Slide
Asymmetric Gable
12
7/30/2013
Multiple Roof Steps
ASCE 7 addresses single roof steps, but not the
multiple roof step geometry shown below
What would be an appropriate leeward drift for
roof C
Multiple Roof Steps
Clearlythe upwind fetch is more than 120’(roof B
alone) and less than 220’(roofs A & B combined)
Upwind fetch less than 220’ since some of the
snow originally on A ends up in drift on B
Trapping efficiency about 50 % - roughly half the
snow which blows over an edge ends up in the
drift immediately beyond the edge – drift on B
doesn’t need to fill before snow on A can
contribute to drift on C
13
7/30/2013
Multiple Roof Steps
From drift equation in ASCE 7 - between 20 & 40 %
of roof snow ( Pg x lu) ends up in leeward drift.
20% corresponds to large fetch and ground load,
40% corresponds to small fetch and ground load
O’Rourke & Kuskowski (2005) conservatively
assume 25 % capture at the upwind leeward drift
and determine the effective fetch for C is 0.75 LA
+ LB
For our problem, upwind fetch is
lu = 0.75(100’) + 120’ = 195’
Multiple Roof Steps
For a design ground snow load of 20 psf and a
upwind fetch of 195’ , Figure 7-9 in ASCE 7-10
gives a drift height of hd = 4.32’
For20 psf and a fetch of 220’ (step at B neglected)
we get hd = 4.56’
Ifthe step at B is small and is easily filled , the
modification factor would be closer to 1.0, that is
lu= 1.0 LA + LB = 1.0(100) + 120 = 220 ft
14
7/30/2013
Multiple Roof Steps
A roof step downwind of a gable is similar
Some snow originally on AB ends up in drift beyond
the ridge line
Trapping efficiency of gable similar to leeward step
Reasonable to use 0.75la + lb as fetch for drift on DE
Non-Standard Roof Shapes
Mansard Shape
The Wing
Multiple Roof Steps
Roof Top Tent
The Slide
Asymmetric Gable
15
7/30/2013
Roof Top Tent
Itis unlikely that the Roof Top Tent - a cross
between a RTU and a gable – will ever be
covered by ASCE 7
However current provisions for both RTU’s and
gables provide a starting point
Roof Top Tent
Ifthe slope is steep,
reasonable to
consider it as a RTU
with sloping walls
As per current
ASCE 7 - windward
drifts on both sides ,
based on larger of
the two fetch
distances
16
7/30/2013
Roof Top Tent
Knowing the
height of the right
angle RTU drift
and an assumed
1:4 rise to run, one
can solve for the
dimensions of a
‘non-vertical wall’
drift with the same
area
Roof Top Tent
Ifthe tent slope is
moderate, the tent
resembles a gable
In that case , one should
add a leeward gable roof
drift to the balanced load
downwind of the ridge
Although we expect a
windward drift at the tent
‘pegs’, use the distance
from ridge to roof edge
as fetch
17
7/30/2013
Roof Top Tent
For drifts steep &
shallow are with
respect to the angle
of repose of drifted
snow
Angle of repose for
fresh fallen snow
quite large
Roof Top Tent
Vertical axis is height for
drifted snow –
horizontal axis is length
Typical drift has rise to
run about 1:4 (3 on 12)
Largest about 1:2 (26˚) or
6 on 12
In ASCE 7 – no drift on 7
on 12 slope
18
7/30/2013
Non-Standard Roof Shapes
Mansard Shape
The Wing
Multiple Roof Steps
Roof Top Tent
The Slide
Asymmetric Gable
The Slide
The distinctive feature of this roof is the large
concave portion – with a 25 º slope on the left and a
flat slope on the right – resembling a playground slide.
Encloses heated high occupancy space, Pg = 20 psf
Will assume a slippery roof surface
19
7/30/2013
The Slide
ASCE 7 addresses arch/domes with a convex shape
and saw-tooth roof shapes.
Sliding
for an arch/dome results in less load on the
arch/dome
Sliding for a saw-tooth results in a redistribution of
load
The Slide
Will
assume Ce =Ct =1.0 with Is = 1.1 due to the
high occupancy. Since sliding snow may well
remain will use Cs = 1.0
Flat roof snow load 0.7CeCtCsIsPg = 19.3 psf
Minimum roof load for Pg > 20 is IsPg = 22 psf
20
7/30/2013
The Slide
For
wind from right to left, no abrupt change in slope,
separation not expected, drifts not expected
For
wind from left to right we have multistory step.
Conservative to base drift on gable with slope of 25 º
and fetch as shown. Place atop flat roof snow load.
The Slide
Slidingloads are tricky – more likely to occur if low
end is to the South. In that case snow on the roof is
warmed by sun and more likely to slide.
Sliding
loads in ASCE 7 (0.4 Pf) only if not blocked
“by snow already on the lower level roof”
For“Slide”, lower snow could block sliding, or rigid
body movement with snow being pushed off the
low end, or upper snow somehow over-riding lower
snow. First two not governing design load -will
consider sliding/over-riding case
21
7/30/2013
The Slide
For
sliding/over-riding - what slides? what is
shape of slid pile? what gets over-ridden?
Seems reasonable to adopt 0.4 Pf as sliding load
from ASCE 7 or 0.4(19.3) = 7.7 psf in our case
Because of uncertainty, reasonable to consider a
number of roof slope boundaries between sliding
part and over-ridden part
Againbecause of uncertainty, reasonable to
consider multiple shapes for the snow that slides
The Slide
Slopefactor Cs for
warm roofs in ASCE
7 suggests sliding
important for roof
slope of 5º and
steeper
‘Over-riddensnow’
boundaries of 10º,
15º and 20º seem
reasonable
22
7/30/2013
The Slide
Seems reasonable to
consider a
rectangular and a
triangular shape
Over-ride distance
could be 25 to 75 %
of length of sliding
snow
Figures show 50%
over-ride with 15º
boundary between
sliding and over-
ridden snow
Non-Standard Roof Shapes
Mansard Shape
The Wing
Multiple Roof Steps
Roof Top Tent
The Slide
Asymmetric Gable
23
7/30/2013
Asymmetric gable
TheSymmetric gable shape has always been
covered in ASCE 7
Asymmetric gables are currently not covered
although they may be the next ‘new’ shape
Asymmetry can arise due to the different roof
lengths or different roof slopes each side of the
ridge
Asymmetric Gable
Different
roof
lengths are easily
addressed
The up-wind fetch
for a drift on one
side of the ridge is
the eave to ridge
distance for the
other side
24
7/30/2013
Asymmetric Gable
Differentroof slopes
are a bit more
complex
Gable drifts not
expected if the
change in roof slope
is less than ½ on 12
to ½ on 12 ( ~ 4.8 º)
Inthat case
streamlines remain
attached, no flow
separation & no
drifting
Asymmetric Gable
Gabledrifts are not
expected on roof
slopes steeper than
7 on 12
Such drifts would
require an angle of
repose of drifted
snow greater than 7
on 12
25
7/30/2013
Asymmetric Gables
The horizontal
extent and intensity
of the gable drift is a
function of the roof
slope parameter S
ForAsymmetric
gables use S for the
side in question
Non-Standard Roof Shapes
Questions?
26
7/30/2013
Live P.E./S.E. Exam Review Courses, Fall 2013
Pass the P.E./S.E. Exam and Save with ASCE’s New Individual Pricing for
Review Courses. Three review courses are offered:
P.E. Civil Exam
P.E. Environmental Exam
S.E. Structural Exam
For more information and registration visit [Link]/pereview
Pay a single site registration fee and an unlimited number of people in your organization can
attend at that site or attend as an individual.
S.E. Structural Exam Review Course - with Seismic
Design Preparation
Mon., Aug. 12 – Gravity Loads Wed., Aug. 14 – Lateral Loads
Mon., Aug. 19 – Structural Analysis Wed., Aug. 21 – Bridge Loads
Mon., Aug. 26 – Bridge Design Wed., Aug. 28 – Masonry Design
Wed., Sept. 4 – Timber Design Wed., Sept. 11 – Concrete Buildings
Mon., Sept. 16 – Prestressed Concrete Wed., Sept. 18 – Seismic Design
Wed., Sept. 25 – Steel Design
27
7/30/2013
P.E. Civil Exam Review, 12-Part Course, Fall 2013
Tue., Aug.13 – Structural Analysis Thur., Aug. 15 – Strength of Materials
Tue., Aug. 20 – Structural Design Thur., Aug. 22 – Construction Materials
Tue., Aug. 27 – Geometric Design Thur., Aug. 29 – Engineering Cost Analysis
Tue., Sept. 3 – Hydraulics Thur., Sept. 5 - Hydraulics
Tue., Sept. 10 – Waste & Water Treatment Thur., Sept. 12 – Geomechanics
Tue., Sept. 17 – Foundation Engineering Thur., Sept. 19 – Construction Scheduling
and Estimating
P.E. Civil Exam Review, 17-Part Course, Fall 2013
Tue., Aug.13 – Structural Analysis Thur., Aug. 15 – Strength of Materials
Tue., Aug. 20 – Structural Design Thur., Aug. 22 – Construction Materials
Tue., Aug. 27 – Geometric Design Thur., Aug. 29 – Engineering Cost Analysis
Tue., Sept. 3 – Hydraulics Thur., Sept. 5 - Hydraulics
Tue., Sept. 10 – Waste & Water Treatment Thur., Sept. 12 – Geomechanics
Tue., Sept. 17 – Foundation Engineering Thur., Sept. 19 – Construction Scheduling
and Estimating
Mon., Sept. 9 – Water Resource Depth Mon., Sept. 23 – Structures Depth
Tue., Sept. 24 – Geotechnical Depth Thur., Sept. 26 – Transportation Depth
Fri., Sept. 27 - Construction Depth
28
7/30/2013
P.E. Environmental Exam Review Course
Fri., Aug. 16 – Environmental Assessment Fri., Aug. 23 – Hazardous Waste and
and Remediation Emergency Operations
Fri., Aug. 30 – Storm Water Fri. Sept. 6 – Waste and Water Treatment
Fri., Sept. 13 – Water Quality Fri., Sept. 20 – Air Quality
29