plain flow could reenter the main channel just upstream
over the full width of the pier, can be the controlling
of the ICS, where water levels are below bank-full. This
mode on small streams (less than 300 ft wide). This
mode provides the closest analogy to the ice-arrest func-
flow can erode the ice floes arched between the flood-
tion of the proposed ICS. For vertical piers, the appli-
plain and the nearest structural element, causing early
cable equation is (AASHTO 1998)
release of the ice jam. Furthermore, this flow is suffi-
ciently fast to scour natural bank materials, threatening
Fi = Ca p t w Ki
(1)
catastrophic release of the jam.
Riprap placed along the right bank near the ICS can
where
prevent both problems. The riprap should extend at least
Ca = (5t/w + 1)0.5
(2)
one river width (about 150 ft) upstream and downstream
of the ICS and be tied-off to existing ground. To con-
and p
= effective crushing strength
trol reentering flow, the top of the riprap should be about
t
= ice thickness
2 ft above the local floodplain elevation, making it EL
w
= pier width
643 here. A similar riprap emplacement at the Hard-
Ki
= impact reduction factor.
wick ICS has worked well to prevent scour and reen-
tering floodplain flow.
Here, w = 5.0 ft, and we may assume t = 2.0 ft for
Although dynamic ice-jam surges can occur near the
design purposes. Equation 2 then yields Ca = 1.7. The
ICS, the banks do not experience persistent ice action.
largest value of p suggested in the standard is about
Consequently, local flow velocities rather than ice action
220 psi, recommended for cases where breakup can
should govern the riprap design. At Hardwick, the riprap
occur when the average ice temperature is measurably
consists of granite stone with D50 = 1.7 ft, and it has
below the melting point (e.g., mid-winter breakup).
shown no signs of failure. For the Cazenovia Creek ICS,
Lastly, Ki depends on A/r2, where A is the floe plan area
D50 = 2.0 ft should be adequate.
and r is the radius of the pier nose. The model tests sug-
gest that floes larger than about 30 ft across seldom strike
Ice-retaining posts
a pier at high velocity. For this case, Ki ~ 0.7 applies.
These values combine, via eq 1, to predict Fi = 0.4 106
The treed floodplain adjacent to the ICS allows flow
to bypass the structure but retains ice in the main chan-
lb, remarkably close to the design value suggested here.
nel. However, several gaps exists in the trees along the
The AASHTO standards do not provide quantita-
right bank at the Cazenovia Creek ICS site. In these
tive guidance on the elevation at which to apply the
gaps, wooden posts should be installed to retain ice
ice-impact force to the pier, especially as these stan-
pieces in the main channel.
dards are not intended for the design of ice-control struc-
Posts installed in the physical model (see Fig. 6)
tures. For this purpose, the model tests probably pro-
retained most ice floes in the main channel during sim-
vide the best guidance, Lp = 4.4 ft. Also, because they
ulated breakup events. Maximum ice elevations along
apply to bridge piers, the standards recommend simul-
this right-bank "tree line" were 57 ft above the local
taneous application of the downstream force and a trans-
floodplain elevation. We did not, however, measure
verse force equal to only 15% of that value. The simul-
loads on the posts. Much of the ice was grounded along
taneous transverse force recommended here, based on
the bank, and the water level drop across the posts was
model data, is 45% of the downstream force and is con-
much less than 5 ft.
servative compared to the standards.
Wooden posts placed 6 ft on-center along the gaps
in the right-bank trees are adequate to retain ice. The
Additional ICS features
posts should protrude 4 ft above the top of the bank.
The cylindrical-pier ICS uses the adjacent treed
We may conservatively assume that a 5-ft hydrostatic
floodplain, located along the right bank of Cazenovia
head acts across each post and that the resulting force
Creek (see Fig. 6), as a flow-bypass channel. This pre-
acts 2 ft above the ground. Thus, each post must resist
sents additional design requirements: preventing sedi-
12,000 ft-lb of overturning moment.
ment scour along the right bank, preventing floodplain
Dry, select-grade Douglas Fir and Southern Pine
flow from reentering the main channel upstream of the
(long leaf) have allowable tensile stresses for static loads
ICS, and retaining the ice pieces in the main channel as
of about 1800 psi (CRC 1970). Thus, 10-in.-diameter
flow diverts onto the floodplain. We may satisfy these
posts of these woods are adequate to resist the estimat-
requirements with additional ICS features.
ed overturning moment; use of 12-in.-diameter posts is
conservative. Note that allowable stresses may be in-
Right-bank riprap
Model tests of the cylindrical-pier ICS and the
creased by 100% for impact loads (CRC 1970), which
sloped-block ICS (Lever et al. 1997) revealed that flood-
should accommodate moments from ice impacts.
11
TO CONTENTS
Previous Page
