Professional Documents
Culture Documents
1
School of Natural and Built Environments, University of South Australia
(terry.lucke@.unisa.edu.au)
ABSTRACT
KEYWORDS
Siphonic roof drainage, stormwater harvesting, water reuse, building services design
Commercial siphonic rainwater drainage systems were first developed by Ebeling and
Sommerhein in the early 1970s in Scandinavia (May, 1995). Since then, many
thousands of new buildings worldwide have been designed incorporating siphonic
roof drainage systems. The advantages of these systems over conventional roof
drainage systems are numerous and have much appeal for architects and designers.
Siphonic Drainage is growing from a once “obscure curiosity” in Europe to an
emerging market in the United States (Rattenbury, 2005). Because of the height
requirements needed for siphonic roof drainage systems, the technology is only viable
for larger commercial buildings and structures over about four metres in height (Refer
to Plates 1 and 2)
Siphonic systems are designed to exclude air from the pipework and, once primed,
cause the pipes to flow under pressure. Syphonic roof drainage systems have strategic
advantages over conventional systems, and particularly so in respect of their cost-
effectiveness to quickly remove large volumes of rainwater safely and effectively
(Brahmall and Saul, 1999). A major advantage of siphonic systems is the greatly
increased driving head of water and consequent reduction in pipe diameter sizes. The
driving head in this case is effectively the difference in level between the water in the
gutter and the ultimate discharge point, which is usually near ground level.
The increased driving head in siphonic systems offers much potential for stormwater
harvesting and reuse. Because the building’s total roof runoff is normally discharged
from only one or two downpipes with high velocity, the water can easily be directed
to most places on a development site without the need for pumping. This means that
Air
Water
H = head of water
in gutter D = diameter of
downpipe
Research has shown that the water flowing in the downpipe in a conventional roof
drainage system is annular in nature (Wright, Jack and Swaffield, 2006). This means
that the water spirals down the inner edges or walls of the pipe and there is a hollow,
air-filled core down the centre of the water flow (Arthur, Wright and Swaffield,
2005). The air that is drawn down by the water actually restricts the water discharge
in a pipe to between one quarter and one third of the pipe cross section area. This
means that large diameter pipes are required to enable the gutters to drain quickly
without risk of overflowing. These types of conventional roof drainage systems are
very inefficient and require extensive underground pipework systems (Figure 2).
SIPHONIC DRAINAGE
Siphonic roof drainage systems are currently designed to operate with full bore flow
without the need for any mechanical pumping either to prime or operate the system.
The gutter outlets in siphonic roof drainage systems are specifically designed to
restrict the inflow of air into the pipework and this allows a much greater volume of
water to flow in the pipe. Because there is limited air in the pipework, the falling
water generates a vacuum behind it which “sucks” the water into the gutter outlet.
This vacuum effect around the outlet allows more water to be drawn into the pipe
resulting in much greater flow rates (Arthur and Swaffield, 2001).
Unlike conventional roof drainage, in siphonic systems gutter outlet pipes are directed
into a horizontal collection pipe which often then flows into a single downpipe
(Figure 3). This collection pipe usually runs at roof level close to the gutters collecting
all the water flowing out of the gutter (Arthur and Swaffield, 2001). The total volume
of water collected then flows into the downpipe and is often discharged at a single
outlet at ground level. This outlet can be located directly in the underground
stormwater system or redirected to a rainwater tank for harvesting and reuse (Figure
3).
Because the roof runoff from siphonic systems is usually directed into a single
downpipe, the normally extensive, underground drainage pipe system is virtually
eliminated. Besides the obvious benefit of lower excavation and pipe costs, siphonic
systems also minimise potential building damage associated with footing movements
in reactive soils. As there are generally no drainage pipes under the slab or parallel to
footings, soil heave problems caused by leaking pipes are also eliminated.
Siphonic drainage systems are designed using pipe full flow conditions which assume
no air in the system. As such, most siphonic outlets are specially designed to reduce
the amount of air entering the system. This is often achieved by means of a horizontal
baffle plate configuration at the siphonic outlet installed in the gutter floor. These
baffle plates restrict the formation of a vortex above the outlet which would suck air
into the system and break the siphon action. Other outlet configurations include a type
of “upturned dish” arrangement which forms an airlock to restrict the air entrainment
into the flow (Figure 4).
SYSTEM PRIMING
Siphonic rainwater drainage systems are designed to operate under both part-full and
pipe-full conditions. The transition between these two states involves priming or
unpriming of the system, both of which involve considerable air entrainment.
Priming is the term used to describe the process where resistance to flow is sufficient
to cause the pipe system to become full of water. It is the friction and form losses
which are present in every pipe flow which resists the movement of the water and
assists in the development of pipe-full flow conditions.
The design of siphonic systems involves dynamic balancing of the pipe systems when
they are flowing under pressure. Friction and form losses in pipes are proportional to
the square of the velocity of the fluid and are cumulative. This means that the further
the water travels, the more energy it loses and consequently the less volume of water
can flow. In order for the siphonic outlet flows to be balanced, smaller tail pipe
diameters are often used closer to the vertical downpipe to reduce the flow volumes to
similar flows upstream.
The priming sequence usually occurs in three phases. As the water level in the gutter
increases, so does the flow into the vertical tailpipe. This flow enters the lateral
collector pipe in a supercritical state and thereafter forms a hydraulic jump. As long
as the subcritical depth downstream of the jump is less than the pipe diameter, the
flow in the vertical stack remains annular, as for conventional gravity flow systems.
As the flowrates increase, the subcritical depth downstream of the jump also increases
As part of an ongoing PhD research project in siphonic roof drainage systems titled,
“The role of air entrainment in the performance of siphonic roof drainage systems”
, a full scale siphonic drainage rig has been constructed in the hydraulics laboratory at
the University of South Australia (UniSA) Mawson Lakes Campus (Refer Plate 3).
This research is being undertaken in collaboration with Syfon Systems of Melbourne,
a leading Australian siphonic drainage company since 1992.The dimensions of the
UniSA rig are 32m long by 6m high by 3m wide. This is, according to the literature
reviewed so far, the largest laboratory-based siphonic testing facility in the world.
In order to allow visual observation of the flow patterns, perspex was chosen for the
pipework material. Hydraulic calculations undertaken for this system using
commercial siphonic software predicted that the expected maximum flow rate in the
laboratory will be 69 litres per second. This is equivalent to the rainfall of a 1 in 300
year storm event falling on the entire roof of the hydraulics building in which the
model is housed. The maximum flowrate measured through the rig has been 70 litres
per second which thoroughly agrees with the calculations. This demonstrates the
accuracy of the commercial software used in designing this system.
Research into siphonic roof drainage systems has been undertaken at Heriot Watt
University in Edinburgh, Scotland. The principal researchers have been Swaffield,
Wright and Arthur of the Heriot-Watt Drainage Research Group. They described the
formation of hydraulic jumps within the pipework and the role they play in priming.
Arthur and Swaffield identified three main effects that air entrainment has on system
performance. Air affects system operating pressure, propagation velocity and friction
losses. Their research has identified various interesting phenomena of siphonic
drainage which will be further investigated on the UniSA rig.
The design of the testing facility at UniSA has evolved from studying the limitations
of existing experimental models identified in the research literature. The UniSA rig
has much larger pipe diameters than previously built models and has four gutter
outlets. At six metres high, the apparatus is considered to be representative of real
siphonic roof drainage systems in use.
This results in low energy stormwater distributed along the entire length of the
underground drainage pipe system. In order to collect this water for re-use, some type
of collection pit and pumping system would normally be needed. This not only
increases system costs and energy consumption but also places landuse restrictions on
the building’s surrounding area.
As previously discussed, the discharge from siphonic roof drainage systems is usually
from a single, full-flowing downpipe at high velocity. This can generally enable the
stormwater to be directed to any part of a development site, even to the highest
elevated areas. Rainwater tanks or other collection devices can then be used to harvest
and store the rainwater for later reuse. This stormwater can be utilised for many
different activities ranging from simple landscape irrigation to toilet flushing, vehicle
washing and other uses.
Other areas of the research to be undertaken at UniSA include the effects of air
entrainment on the performance of siphonic systems and investigation into the
transition area between siphonic downpipe outlets and underground drainage systems.
The design of siphonic outlets to minimise air entrainment will also be investigated.
CONCLUSIONS
References:
Arthur, S. and Swaffield, J.A. (2001). “Siphonic roof drainage: current understanding.” Urban
Water, 3, Taylor and Francis, pp. 43-52.
Arthur, S., Wright, G.B., and Swaffield, J.A. (2005). “Operational performance of siphonic
roof drainage systems.” Building and Environment 40, pp. 788-796.
Bramhall, M.A., and Saul, A.J. (1999). “Hydraulic performance of siphonic rainwater
outlets.” Proceedings of the 8th international conference on urban stormwater drainage.
Sydney. Australia
May, RWP, (1995), Design of conventional and siphonic roof drainage systems, Public
Health Services in Buildings - Water Supply, Quality and Drainage, IWEM Conference,
London.
Wright, G.B., Jack, L.B., and Swaffield, J.A. (2006). “Investigation and numerical modelling
of roof drainage systems under extreme events.” Building and Environment 41, pp.126-135.