You are on page 1of 8

Journal of Building Engineering 2 (2015) 9–16

Contents lists available at ScienceDirect

Journal of Building Engineering


journal homepage: www.elsevier.com/locate/jobe

Alternative low-cost overflows for siphonic roof drainage systems:


Proof of concept
Terry Lucke a,n, Simon Beecham b
a
Stormwater Research Group, University of the Sunshine Coast, QLD, Australia
b
School of Natural and Built Environments, University of South Australia, SA, Australia

art ic l e i nf o a b s t r a c t

Article history: This paper describes an experimental investigation of alternative overflow systems for roof structures
Received 26 November 2014 that are drained siphonically. Many buildings that have siphonic roof drainage systems currently in-
Received in revised form corporate a separate secondary overflow system, which is quite often also siphonic. An alternative low-
1 March 2015
cost technique that is explored in this study is to connect each overflow outlet to a single, vertical
Accepted 20 March 2015
downpipe. Seven different overflow configurations, each with five different downpipe lengths, are in-
Available online 21 March 2015
vestigated in terms of maximum flowrate and corresponding water depth in the gutter. The results of this
Keywords: study are significant in that they do not support the common theoretical assumption that there is a
Siphonic roof drainage system limiting length of a siphonic downpipe over which pipe-full flow may occur before gravity causes the
Secondary overflow system
water to accelerate enough for it to no longer occupy the full cross-sectional area. Instead, this study has
Siphonic outlets
found that once a downpipe is flowing full, it is possible for pipe-full flow conditions to be maintained
Full pipe flow
over the entire pipe length.
& 2015 Elsevier Ltd. All rights reserved.

1. Introduction conventional roof drainage systems and also requires extensive


underground pipework systems (Fig. 1). This has significant cost
Siphonic roof drainage systems (SRDS) are a highly efficient implications for the broader construction industry.
type of drainage system that are particularly suitable for buildings Unlike conventional drainage systems, the pipework of a SRDS
with large roof areas that need to be drained quickly. SRDS were is designed to flow full at its design capacity [18]. Through the use
first developed in the late 1960s by Ebeling and Sommerhein in of specially designed gutter outlets and pipework, air is purged
Scandinavia [20] and they have much appeal for architects and from the system and the pipes quickly fill with water. Once the air
designers due to the many advantages they offer over conventional is purged from the pipes, they then operate under sub-atmo-
roof drainage systems. These include a significant reduction in the
spheric pressures. The driving head for the system is the effective
number of downpipes required, the possibility of relocating
difference in levels between the water surface in the gutter and
downpipes to areas that are esthetically less sensitive, and the ease
the discharge point, which is usually near ground level. This causes
of harvesting all of a building's roofwater for later reuse.
significant increases in both flow velocity and volumetric flowrate
Conventional roof drainage systems typically include box, eave
or valley gutters that collect the runoff from the roof and channel compared to traditional systems [20,23]. May [20] explained that
this rainwater into outlets connected to vertical downpipes located these increases can cause siphonic outlets to have up to 10 times
in the soles of the gutters. The volume of water that can enter the the capacity of conventional outlets.
gutter outlets depends primarily on the depth of water in the One of the major advantages of SRDS is that the roof runoff
gutter, and on the cross sectional area of the outlet. This volume from each siphonic system is usually directed into a single
can be estimated using standard weir and orifice equations [20]. downpipe, so the excessive number of vertical downpipes, and the
However, the volume is relatively limited because up to two-thirds extensive underground drainage pipe system typically associated
of the downpipe volume can be taken up by an air-filled core with conventional systems can be virtually eliminated (Fig. 1). A
[1,19,26]. This necessitates the installation of many downpipes in single downpipe also makes it much easier to harvest all the
roofwater from a building. However, some building designs, par-
n
Corresponding author.
ticularly those with roof areas at different levels, may incorporate
E-mail address: tlucke@usc.edu.au (T. Lucke). numerous siphonic systems and downpipes.

http://dx.doi.org/10.1016/j.jobe.2015.03.006
2352-7102/& 2015 Elsevier Ltd. All rights reserved.
10 T. Lucke, S. Beecham / Journal of Building Engineering 2 (2015) 9–16

Fig. 1. Conventional and siphonic roof drainage system layout.

1.1. Siphonic design undertaken to try to better understand the priming process
[3,16,23] it is still not fully understood.
The theory of siphonic action has long been understood and is The design of a SRDS for a commercial building can be a highly
broadly based on simple energy principles, as expressed by Ber- complex procedure. It generally involves an iterative design pro-
noulli's energy equation [20]. The current steady (peak) flow de- cess of adjusting pipe lengths, pipe diameters and the inclusion or
sign of SRDS generally uses a version of this equation that esti- exclusion of pipe fittings, in order to accurately balance both sides
mates the difference in energy between two points (1 and 2) by of Eq. (1) for a particular design storm. This is usually only possible
summing the pressure, kinetic and potential energies at those by using a computer program [20]. Apart from ensuring that a
points (Eq. 1). This energy is then balanced against the pipe sys- SRDS has the capacity to cope with the roof runoff from a parti-
tem's friction (Hf) and form losses (HL). cular design storm, the other principle design objectives are to
⎛ P1 Q 12 ⎞ ⎛ P2 Q 22 ⎞ ensure that the outlets are balanced during operation and that the
⎜⎜ + + z1⎟⎟ − ⎜⎜ + + z2 ⎟⎟ pipe pressures do not become too negative [20].
2 2
⎝ ρg 2gA1 ⎠Point 1 ⎝ ρg 2gA2 ⎠Point 2
Previous studies [2,3,17,20] and design manuals [4,25] suggest
= ∑ H1,2 that the expected internal pipe pressures should be limited to a
minimum pressure of 90 kN/m2 below atmospheric pressure. This
= ∑ Hf + ∑ HL (1) pressure limit is currently recommended for two reasons. The first
where: P is pressure, Q is flowrate, ρ is density, g is gravity, A is is to ensure that the generation of negative pressure transients
cross-sectional area and z is height above datum. does not lead to system failure due to pipe wall collapse [2]. The
Although SRDS are designed to operate under pipe-full flow second reason is to reduce the likelihood of cavitation, which
conditions, they can also operate efficiently when the pipes are could lead to serious erosion damage on the inside of the pipes
only partially full. The transition between these two flow states [17,20].
involves priming or unpriming of the system. Priming is the term As siphonic drainage systems are designed for pipe-full flow
used to describe the process where resistance to flow is sufficient conditions, the design usually assumes that there is no air in the
to cause the pipe system to purge air from the pipework and be- system. To achieve this, most siphonic outlets are specially de-
come full of water [2,20]. It is the friction and form losses present signed to reduce the amount of air entering the system. This is
in every pipe flow situation that resists the movement of the water often achieved by including some type of baffle plate, or similar
and assists in the development of pipe-full flow conditions [19]. configuration, in the outlet's design (Fig. 2a). These baffle plates
The priming action that occurs in SRDS is an extremely complex help restrict the formation of a vortex above the outlet that would
process. Although there have been a number of studies otherwise draw air into the system. Air drawn in through a vortex

Fig. 2. Normal and overflow siphonic outlets.


T. Lucke, S. Beecham / Journal of Building Engineering 2 (2015) 9–16 11

can break the siphonic action required to drive the system. water level and the overflow outlets only become operational if
The particular storm selected for the SRDS design is highly the water level in the gutter rises to the height of the weir [22].
dependent on its location. In Australia for example, SRDS are The height of the outlet weir is dependent on the environmental
generally designed with a capacity to accommodate the peak conditions at the building's location as well as the individual de-
flowrate resulting from a 100 year average recurrence interval, sign requirements of each SRDS. Generally, the weir height must
5 min duration design rainfall intensity at the building location be determined on a case-by-case basis, however, it is typically
(referred to as 100I5). This design standard is based on procedures between 70 and 120 mm above the base of the outlet. Some pro-
outlined in Australian Rainfall and Runoff, [14]. In New Zealand, a prietary systems include a baffle plate above the weir (Fig. 2b),
50 year average recurrence interval, 10 min duration (50I10) design while others simply use a normal siphonic outlet surrounded by a
rainfall intensity at the building location is used. However, in the weir which is open to the atmosphere [9].
UK for example, SRDS are designed according to a derived statis- Many SRDS designs also generally include a complete second-
tical range of rainfall intensities, r (L/s/m2), and are often based on ary system to allow for potential overflow situations. For example,
a critical storm duration of 2 min resulting from a 50 year average the Australian Standard AS3500.3:2003 [5] states that (Clause
recurrence interval (50I2) [12]. 3.7.5.1) “The hydraulic capacity of an overflow device shall be not
Irrespective of the procedure used to select the appropriate less than the design flow for the associated gutter outlet.” This
design storm for the SRDS location, it is clear that the use of peak requirement for secondary systems is evident in other interna-
flow design method means that there will only be one solution for tional standards. For example, the British Siphonic Standard BS
Eq. (1) for each individual SRDS. This means that every other 8490:2007 [11] simply states that for overflow arrangements, each
rainfall intensity that the SRDS experiences is either below or roof area has to be fitted with… “at least one emergency overflow
above the design capacity of the SRDS. While this is generally not a pipe discharging in a conspicuous location outside the building.”
problem for SRDS experiencing rainfall events below the design Similarly, the equivalent American standard (ASTM F2021:2006)
storm intensity, it could potentially result in operational problems states that the gutter design should… “incorporate a built-in
such as gutter overtopping and building flood damage for rainfall overflow facility and shall be designed to hold water up to this
intensities that occur above the design storm [6,22]. This could level should the design rainfall be exceeded or unexpected
now be of particular concern as climate change is potentially blockages occur.” Probably the most specific standard is the Ger-
creating situations where normal ranges of rainfall intensity are man VDI 3806:2000, which requires that… “the drainage capacity
being frequently exceeded [13]. of the emergency overflows is set to the difference between the
100 year rainfall and the maximum drainage capacity…” of the
1.2. Gutter overflow configurations SRDS.
The requirement for a complete secondary overflow system
To reduce the risk of gutter overtopping and flood damage due significantly increases the overall design, installation and material
to unexpected rainfall events, or due to outlet or pipework costs associated with SRDS [10,21]. In SRDS designs where the
blockages, the design of SRDS generally also includes a secondary secondary system is required to cope with the entire runoff vo-
overflow system [8,9,23]. The design of conventional roof drainage lume in case of primary system failure, the secondary system costs
systems also often incudes provisions for gutter overflow devices. can be similar to the primary system costs. Fig. 3 shows an ex-
However, these are generally not specifically designed for siphonic ample of a SRDS in Australia where both the primary and sec-
action, and may not be suitable due to potential limitations in ondary systems were designed to manage the total roof runoff
operating pressures, airtightness and outlet designs. Notwith- volume independently. It is clear that the secondary system would
standing this, previous research has shown that siphonic action have considerably added to the cost of this building.
can develop in conventional downpipes [23]. Many structures have been specifically designed with a SRDS
May [22] explained that in SRDS design, there is an increasing because building esthetics are a leading priority and unsightly
tendency to install two parallel systems (“primary” and “second- downpipes are undesirable, or because designers have explicit
ary” systems) due to the uncertainty of SRDS operating sa- requirements or restrictions regarding the location of downpipes.
tisfactorily during high-intensity storms. The outlets of secondary This is particularly the case in sports stadia where traditional
SRDS are typically surrounded by some type of raised weir ar- downpipes connected to downward sloping roofs would be highly
rangement around their perimeter (Fig. 2b). The weir is set to obtrusive. Other structures, such as large factory buildings with
some predetermined height above the normal operational gutter valley roofs and box gutters, may opt for a SRDS to reduce the

Fig. 3. Example of primary and secondary SRDS.


12 T. Lucke, S. Beecham / Journal of Building Engineering 2 (2015) 9–16

number of downpipes (and internal supporting columns) required and was measured using a magnetic flow meter (Type ABB Mag-
in order to maximize open floor space. For these types of struc- master CL16) that was installed in the 150 mm diameter supply
tures, the installation of a complete secondary overflow SRDS is pipework.
probably the best option. Four different prototype overflow devices were tested during
However, for structures that are using SRDS for other reasons, this study. These included: one 90 mm diameter circular stainless
such as to minimize underground pipework requirements, or to steel outlet (Fig. 5a); one 80 mm diameter circular polyvinyl
harvest rainwater from the roof (Fig. 3), a complete secondary chloride (PVC) outlet (Fig. 5b); one 300 mm square, proprietary
overflow SRDS may not be necessary. For more functional struc- steel siphonic outlet with a 110 mm diameter outlet in its base
tures where esthetics are not the main priority, there may be al- (Fig. 5c); and one 300 mm square, raised-weir type (similar to the
ternatives methods available to manage potential overflows. If a configuration shown in Fig. 2b) proprietary PVC siphonic overflow
more cost effective overflow system could be developed for these outlet with a 110 mm diameter outlet in its base (Fig. 5d). Each of
types of structures, this may make SRDS a more attractive option the overflow outlets was installed at the single outlet location in
for building designers and owners. This could significantly in- the sole of the box gutter shown in Fig. 4 and tested separately. All
crease the appeal of SRDS and potentially lead to a substantial of the test rig's normal operational siphonic outlets were sealed
increase in the number of SRDS installed globally. during the study and this enabled accurate measurement of the
This study investigated the possibility of connecting SRDS flowrates through the new single outlet.
overflow outlets to single, vertical downpipes as a potential low- The four selected overflow types represent the two main ca-
cost alternative to the requirement for a complete secondary SRDS tegories of overflows used in siphonic systems. The first two types
to manage gutter overflow situations. The weir and orifice equa- (Fig. 5a and b) are two different sizes of gravity-fed, non-siphonic
tions typically used to predict flowrates through gutter outlets are overflows. The remaining two types (Fig. 5c and d) are two dif-
based on the assumption that the system is operating under at- ferent siphonic overflows, one level with the gutter floor and one
mospheric pressure. Hence, maximum flowrates predicted using with a raised weir outlet. Together these four types of overflow
these equations for typical outlet and downpipe sizes are generally represent the range of overflows commonly used in siphonic
relatively low. However, by purposely inducing siphonic action systems. It should be noted that because most gutters in siphonic
(and sub-atmospheric pressures) in the vertical downpipes, the systems are internal box gutters, the use of gutter wall cut-out
flowrates through the outlets could be significantly increased. overflows are very rare.
Normal weir and orifice equations are no longer valid under these A variety of vertical downpipe materials, diameters and lengths
flow conditions [20]. This paper describes an experimental in- were investigated for each of the seven overflow configurations,
vestigation of the performance of seven different overflow con- involving four different overflow types. For each configuration, five
figurations that can be used in SRDS. different downpipe lengths were investigated, namely 1, 2, 3,
4 and 5 m. These were attached to each of the overflow config-
urations and the maximum possible flowrate through each over-
2. Methodology flow was recorded (Fig. 6). The discharge from the downpipes was
collected and returned to the water supply system using a col-
2.1. Experimental design lection pipe system with adjustable heights to allow for the dif-
ferent downpipe lengths (Fig. 6c and d).
A full-scale experimental siphonic drainage rig (Fig. 4) installed Fig. 6a shows a 75 mm diameter high density polyethylene
at the University of South Australia was used in this study to in- (HDPE) downpipe connected to a 90 mm diameter outlet (Fig. 5a)
vestigate options for a low-cost alternative overflow system for and Fig. 6b shows an 80 mm diameter acrylic downpipe connected
SRDS. The experimental rig had a 600 mm wide  300 mm high to an 80 mm diameter PVC outlet (Fig. 5b). Fig. 6c shows a 5 m
box gutter that was 32 m long. The maximum potential driving length of 75 mm diameter HDPE downpipe discharging to the
head of the rig was approximately 6 m. The gutter inflow rate for collection pipe system below. Fig. 6d shows a 4 m long, 80 mm
the experimental rig was controlled by a number of gate valves diameter acrylic downpipe discharging into the opening of the
collection pipe system.

2.2. Maximum flowrate measurements

The primary objective of this investigation was to measure the


maximum flowrate through each of 35 configurations shown in
Table 1.
The following test procedure was followed for each of the 35
tests listed in Table 1:

a. The flowrate into the test rig was first set to 5 L/s.
b. The water level in the gutter was then allowed to stabilize for at
least 5 min.
c. The water level depth directly next to the outlet was measured
and monitored for a further 10 min. Depending on the inflow
rate, the water level in the gutter often cycled in a range be-
tween distinctive maximum and minimum levels.
d. If the water level did not rise above the maximum reached in
Step c) above after 10 min, the water inflow rate into the rig
was increased by a further 5 L/s.
e. Steps (b)–(d) were repeated until such time as the water level
failed to stabilize. The flowrate of the previous increment was
Fig. 4. Full-scale experimental study rig. then recorded as the maximum flowrate for the particular test.
T. Lucke, S. Beecham / Journal of Building Engineering 2 (2015) 9–16 13

Fig. 5. Four different overflow types.

[15,20,24], which have reported a degree of siphonic action oc-


curring in vertical downpipes of both conventional and siphonic
2.3. Gutter water level measurements systems. May [20] concluded that the effective head acting on the
outlet is increased by the distance that the pipe is flowing full and
A secondary objective of this investigation was to measure the this can significantly increase flowrates through the outlet. The
average water levels in the gutter above the outlet as the different results shown in Fig. 7 concur with this finding. The previous
downpipe configurations were experiencing the maximum flow- studies cited above concluded that the siphonic action was limited
rates described above. by the length of pipe-full flow that occurred in the downpipes. For
example, May [20] explained that a “pipe may flow full for a cer-
tain distance until gravity causes the water to accelerate enough
3. Results for it to no longer occupy the full cross-sectional area”.
However, this study found that there was no limiting length for
The maximum flowrates measured through each of the outlets which downpipes flowed full. These results were in contrast to
with the different test configurations are shown in Fig. 7. The those reported by May [20] who concluded that there was a limit
testing for Configuration 2 was discontinued after the initial re- to the length of full pipe flow. The experiments demonstrated that
sults showed that the maximum flowrate for the outlet with the once a pipe was flowing full, it would continue to flow full for its
5 m long downpipe was only 20 L/s and further testing of the entire length (Fig. 9). The results verified that increasing the length
outlet with a 75 mm downpipe was considered to be unnecessary. of the downpipe increased the flowrate through the outlets.
However, the other six configurations were tested with all five However, full-pipe flow conditions would only occur if the depth
downpipe lengths. of water in the gutter was sufficient to ensure continuous flow and
Fig. 8 shows the average water levels in the gutter above the did not allow air to be drawn into the downpipe. Otherwise, air
outlet for the six different configurations and the five different would be drawn into the outlet and the flowrate would decrease
downpipe lengths. [6]. Lucke and Beecham [17] also reported full-pipe flow in single,
vertical downpipes up to 16 m in length. However, a cavitation
zone was found to develop at the top of the downpipe when pipes
4. Discussion longer than approximately 12 m were flowing full. Increasing the
length of the downpipe further was found to increase the length of
It is clear from the results in Fig. 7 that the flowrates through the cavitation zone and not the length of pipe-full flow.
outlets connected to single vertical downpipes can be significant From Fig. 8 it can be seen that the dominant effect on water
when the depth of water above the outlet, and the volume of level was the presence of a baffle plate as the water levels were
water in the gutter are sufficient to maintain full-pipe flow in the considerably lower for configurations 4 and 6 than for configura-
downpipe. This is consistent with previous research results tions 1–3. This is consistent with experimental data from a multi-
14 T. Lucke, S. Beecham / Journal of Building Engineering 2 (2015) 9–16

Fig. 6. Different downpipe configurations and materials used in the study.

outlet siphonic rig [7]. Obviously the water levels were increased depths.
by the raised weir used in configurations 5 and 7 but again, be- As previously mentioned, some buildings are specifically de-
cause of the presence of baffle plates, the water levels did not rise signed with both a primary and secondary SRDS because of
above those for configurations 1–3, which had no baffle plates. building esthetics or due to restrictions on the location of down-
Interestingly, the length of downpipe seemed to affect the pipes. However, for more functional structures, a primary SRDS
water levels in configurations 4–7 more than an increase in outlet may be used in conjunction with a number of strategically placed
diameter (from 110 mm to 125 mm). Overall this experimental overflow outlets each connected to a single, vertical downpipe.
investigation has demonstrated that single, vertical downpipes This may provide a more cost-effective alternative overflow sys-
may potentially be utilized as a low-cost alternative to a complete tem to a complete secondary SRDS. The alternative overflow out-
secondary SRDS to meet box gutter overflow requirements. The lets could be incorporated into the roof drainage design without
study found that, if gutter water conditions are suitable (i.e. the the need for an expensive secondary SRDS system, or for an ex-
depth of water above the outlet, and the volume of water in the pensive additional underground drainage system. The downpipes
gutter are sufficient to maintain full-pipe flow in the downpipe), could either be turned out through the wall a short distance below
siphonic action can occur in vertical downpipes connected to the gutter but still relatively high off the ground (High Level Dis-
overflow outlets to produce flowrates greatly exceeding the charge PipeFig. 10b), or taken down to the base of the building and
maximum allowable 16 L/s per outlet specified in AS3500.3:2003. directed out through the wall to the overland drainage system
However, the water depths required within the gutter to produce (Low-level Discharge PipeFig. 10c). In these cases, the average
these increased flowrates (Fig. 8) may necessitate the use of dee- frequency of once every 100 years for overflow discharges exiting
per box gutters in order to provide the same freeboard. Further the building may be considered acceptable.
research would be required to quantify these increased gutter Use of the alternative low-cost overflow solutions suggested

Table 1
Outlet and downpipe configurations.

Configuration number Outlet shape and diameter Downpipe diameter and material Downpipe lengths

1 Circular 90 mm Steel Outlet, no Baffle Plate (Fig. 5a) 125 mm HDPE 1–5 m
2 Circular 90 mm Steel Outlet, no Baffle Plate (Fig. 5a) 75 mm HDPE 1–5 m
3 Circular 80 mm PVC Outlet, no Baffle Plate (Fig. 5b) 80 mm Acrylic 1–5 m
4 300 Mm Square Steel, 110 mm Outlet, with Baffle Plate (Fig. 5c) 110 mm HDPE 1–5 m
5 300 mm Square PVC Raised Weir, 110 mm Outlet, with Baffle Plate (Fig. 5d) 110 mm HDPE 1–5 m
6 300 mm Square Steel, 110 mm Outlet Weir, with Baffle Plate (Fig. 5c) 125 mm HDPE 1–5 m
7 300 mm Square PVC Raised Weir, 110 mm Outlet, with Baffle Plate (Fig. 5d) 125 mm HDPE 1–5 m
T. Lucke, S. Beecham / Journal of Building Engineering 2 (2015) 9–16 15

60

1m 2m 3m 4m 5m
50

40
Maximum Flowrate (L/s)

30

20

10

0
1 2 3 4 5 6 7
Outlet Configuration Number

Fig. 7. Maximum flowrates for each configuration and downpipe length (1–5 m).

above may make SRDS a more attractive option for building de-
signers and owners, which in turn could significantly increase
their rates of adoption in modern buildings.

5. Conclusions

This study has investigated seven alternative siphonic drainage


overflow system configurations. The experimental results show
that, for these configurations, increasing the length of the down-
pipe increases the flowrate through the outlets providing there is a
sufficient depth of water in the gutter to maintain pipe-full flow
conditions in the downpipe. In contrast, an increase in outlet
diameter did not lead to increased flowrates, although only four
outlet diameters were investigated in this study and more data is
required to verify this finding. Overall, it is concluded that de- Fig. 9. Pressurized pipe flow conditions in the vertical downpipe (Configuration 3,
2 m long downpipe, 25 L/s).
signers may be able to consider lower-cost alternatives to ex-
pensive full secondary siphonic overflow systems.

250

1m 2m 3m 4m 5m

200
Avearge Gutter Water Level (mm)

150

100

50

0
1 2 3 4 5 6 7
Outlet Configuration Number

Fig. 8. Average gutter water levels at maximum flowrates for downpipe lengths of 1–5 m.
16 T. Lucke, S. Beecham / Journal of Building Engineering 2 (2015) 9–16

Fig. 10. Possible conceptual designs for overflow outlets.

Acknowledgements [11] British Standard (BS 8490:2007), Guide to Siphonic Roof Drainage Systems, BSI
Group, Chiswick, London, UK, 2007.
[12] British Standard (BS EN 12506-3:2000), Gravity Drainage Systems Inside
The authors would like to thank the members of Standards Buildings. Roof drainage, layout and Calculation, BSI Group, Chiswick, London,
Australia Committee WS-014-03-02 – Siphonic Systems for their UK, 2000.
support and advice throughout this research project. The re- [13] D.P. Campbell, Novel modulated flow, self configuring, siphonic roof drainage
system, J. Build. Serv. Eng. Res. Technol. 35 (4) (2014) 349–361.
searchers also gratefully acknowledge the funding provided by [14] Engineers Australia, Australian Rainfall and Runoff: A Guide to Flood Estima-
Syfon Systems Ltd. Finally, the authors would like to thank Mr. Tim tion, Engineers Australia, Canberra, Australia, 1999.
Golding and Mr. Doug Coleman from the University of South [15] A.A. Kalinske, Hydraulics of Vertical Drain and Overflow Pipes, Iowa Institute
of Hydraulics Research, Iowa State University, Iowa, USA, 1940.
Australia for the technical advice and support provided to this [16] T. Lucke, S. Beecham, Y.Y. Qu, Estimating flowrates through individual outlets
project. of siphonic roof drainage systems, Build. Res. Inf. (2015), http://dx.doi.org/
10.1080/09613218.2014.987030.
[17] T. Lucke, S. Beecham, Cavitation, aeration and negative pressures in siphonic
roof drainage systems, Build. Serv. Eng. Res. Technol. 30 (2) (2009) 103–119.
References [18] T. Lucke, S. Beecham, Capacity loss in siphonic roof drainage systems due to
aeration, Build. Res. Inf. 38 (2) (2010) 206–217.
[19] T. Lucke, S. Beecham, Aeration and gutter water levels in siphonic roof drai-
[1] S. Arthur, J.A. Swaffield, Siphonic roof drainage system analysis utilising un-
nage systems, Build. Res. Inf. 38 (6) (2010) 670–685.
steady flow theory, Build. Environ. 36 (2001) 939–948.
[20] R.W.P. May, Design of conventional and siphonic roof drainage systems, in:
[2] S. Arthur, J.A. Swaffield, Siphonic roof drainage: current understanding, Urban
Proceedings of Public Health Services in Buildings – Water Supply, Quality &
Water 3 (1) (2001) 43–52.
Drainage, IWEM Conference, London,1995.
[3] S. Arthur, G.B. Wright, J.A. Swaffield, Operational performance of siphonic roof
[21] R.W.P. May, Manual for the hydraulic design of roof drainage systems: a guide
drainage systems, Build. Environ. 40 (2005) 788–796.
to the use of British Standard BS 6367:1983, H.R. Wallingford, UK, 1996, Report
[4] ASTM International (ASTM F2021:2006), Standard Guide for Design and In-
No. SR 485.
stallation of Plastic Siphonic Roof Drainage Systems, ASTM International,
[22] R.W.P. May, Manual for the design of roof drainage systems: a guide to the use
Pennsylvania, USA, 2006.
of European Standard BS EN 12056-3:2000, H.R. Wallingford, UK, 2003, Report
[5] Australian/New Zealand Standard™ (AS/NZS 3500.3:2003), Plumbing and
No. SR 620.
Drainage-stormwater Drainage, Standards Australia, North Sydney, Australia,
[23] R.W.P. May, M. Escarameia, Performance of siphonic drainage systems for roof
2003.
gutters, H.R. Wallingford, UK, 1996, Report No. SR463.
[6] S. Beecham, T. Lucke, Air water flows in building drainage systems, Urban
[24] J. Swaffield, S. Arthur, RW.P. May, Priming of a siphonic rainwater drainage
Water J. (2013), http://dx.doi.org/10.1080/1573062X.2013.820335.
system, in: Proceedings of Water Supply and Drainage for Buildings, CIB W62,
[7] S. Beecham, T. Lucke, Influence of channel geometry on water levels above
Rotterdam, Netherlands, 1998.
siphonic roof outlets, J. Build. Serv. Eng. Res. Technol. 35 (1) (2014) 83–98.
[25] Verein Deutscher Ingenieur (VDI) (3806:2000-04), Dachentwässerung mit
[8] R. Bowler, S. Arthur, Siphonic roof rainwater drainage – design considerations,
Druckströmung (German Institute of Engineers Siphonic Roof Design Manual),
in: Proceedings of Water Supply and Drainage for Buildings, CIB W62, Edin-
Verein Deutscher Ingenieure e.V., VDI Standards Departement, Düsseldorf,
burgh, UK, 1999.
Germany, 2004.
[9] M.A. Bramhall, A.J. Saul, Examination of the performance of siphonic rainwater
[26] G.B. Wright, S. Arthur, J.A. Swaffield, Numerical simulation of the dynamic
outlets, in: Proceedings of Water Supply and Drainage for Buildings, CIB W62,
operation of multi-outlet siphonic roof drainage systems, Build. Environ. 41
Rotterdam, Netherlands, 1998.
(2006) 1279–1290.
[10] M.A. Bramhall, A.J. Saul, The hydraulic performance of siphonic rainwater
outlets relative to their location within a gutter, in: Proceedings of Water
Supply and Drainage for Buildings, CIB W62, Edinburgh, UK,1999.

You might also like