performance of flat roof outlets - hr wallingford · tef: + 44 (0)1491 835381 fax:. + 44 (0)1491...

tr

Performance of Flat Roof Outlets

M EscarameiaRWPMay

Report SR 473August 1 996

R Wallingford

Address and Regbt€red Otlice: HR Wallingford Ltd. Howbery Park, Wallingford, Oxon OXIO 8BATef: + 44 (0)1491 835381 Fax:. + 44 (0)1491 832233

Red.tr.d h g|c|d|d tlo. 562090. HR Wdiro[dd i. a xtxit @d 61jril|4y ol HR WCirolo.d Grdp L6.

E"

sF473 301O8v96

tr

sR473 30rc996

trContract

This report describes work commissioned by the Department of the Environment(DOE) under Research Contract Cl39/5191 for which the nominated officers wsre MrP Woodhead for DOE and Dr W R White lor HR Wallingford. The HR iob numberwas RTS 0053. The report is published on behalf of the Department of theEnvironment, but any opinions expressed in this repod are not necessarily those ofthe funding Depailment. The work was carried out and managed by Ms MEscarameia under the superuision of Mr R W P May.

Prepared by rt^.r*I4 E'r'trk\4'ry\114- . . fl"*o""l"(name)

Approved by RnP %

3D A*A.^rt lg'4b" ' c J '

@ Crown Copyright 1996

Published by permission of the Controller of Her Majesty's Stationery Office

stu73 30108l"6

tr

IV sa473 0210919,6

trSummary

Pedormance of Flat Roof Outlets

M EscarameiaR W P M a y

Report SR 473August 1996

HR Wallingford was commissioned by the Construction Sponsorship Directorateof the Department of the Environment (DOE) to carry out a laboratory study onthe performance ol grated and plain outlets used for drainage of flat roofs. Thestudy involved the testing of four different conventional outlets and of foursiphonic systems (where the pipework connected to the outlets is designed toflow full and siphonic action is achieved by excluding or greatly reducing theamount ol air normally present in conventional syslems). Two of these latleroutlets were also tested with a non-siphonic pipework system. The test facilityconsisted basically of a2 m x 3 m tank, situated approximately 6 m abovedischarge levelto allow the study of siphonic systems. In the tests the flow wasincreased in small steps to lind the submergence limit of the outlet (ie thetransition between weir-type flow and orifice{ype flow), and to determine therating curves of the outlets.

Four main different types of outlet were identified based on the study of shapesand sizes of the outlets, gratings and leafguards and of their relative positions.Guidelines were produced to determine the effective diameter of an outlet foreach of these types. Design equations were developed for weir-type flow andorifice-type flow;the effect of gratings or leafguards on the capacity of the outletstested was also determined.

sPA73 U2t09196

tr

VI sil73 t2to0t6

trContents

Title pageContractSummaryContents

Page

iiiiv

vii

t n t r oduc t i on . . . . . . . . . 11 .1 Scopeo fs tudY . . . . . . - 11.2 Backgroundtheory . - - -2

Description of outlets and siphonic systems tested - . . . .32 .1 Gene ra l cons ide ra t i ons . . . . . . . . . . . - 32.2 Description of the conventional outlets tested - - . . . .42.9 Descr ip t ionof the s iphonic systemstested . . . . . . . .5

Test rig

Tests4.14.24.3

Tes tp rocedure . . .Conven t iona lou t l e t s . . . ' . . . . 6

Details of test conditions for conventional outletsResults of tests with Outlet AResults of tests with Outlet BResults ol tests with Outlet CResults of tests with Outlet DResults of tests with Outlet EResults of tests with Outlet FList of siphonic outlets testedPipe diameters in siphonic systems G, H and I

5.2

Siphonic systems

Analysis and discussion ot results " " " " 8

5 . 1 R a t i n g c u r v e s " " " " 85.1.1 Convent ionalout lets """"85.1.2 SiPhonic sYsfems " " 9

Development of design equations for flat roof outlets ' ' ' ' 10

5.2.1 Effective diameter 105.2.2 Weir f towequat ion . . " " ' 11

5.2.3 Oif iceflowequation-.-. 12

Conclusions and recommendations 13

Acknowledgements . . . . .

References

' t5

1 6

TablesTable 1Table 2Table 3Table 4Table 5Table 6Table 7Table BTable 9

v l l sPA73 U2lO9l96

trContents continued

Figures

Plates

Table 10Table 11Table 12Table 13

Figure 1Figure 2Figure 3Figure 4Figure 5Figure 6Figure 7Figure IFigure 9Figure 10Figure 11Figure 12Figure 13Figure 14Figure 15Figure 16Figure 17Figure 18Figure 19

Figure 20

Figure 21

Plate 1Plate 2Plate 3Plate 4Plate 5Plate 6Plate 7Plate 8Plate 9Plate 10

Results of tests with siphonic system GResults of tests with siphonic system HResults of tests with siphonic system IResults of tests with siphonic system J

Schematic diagrams of outlets A, B, C, D and ESchematic diagrams of outlets F, G, H, I and JSchematic diagram of test rigLocation of tapping pointsSchematic diagram of siphonic systems G, H and ISchematic diagram of siphonic system JRating curve of Outlet ARating curue of Outlet BRating curve of Outlet CRating curve of Outlet DRating curue of Outlet ERating curve of Outlet FRating curve of Outlet GRating curve of Outlet HRating curve of Outlet IRating curve of Outlet JRelationship between Q and hDefinition of effective diameterComparison of measured and predicted flows for weirflow conditionsComparison ol measured and predicted flows for orificeflow conditionsComparison of measured and predicted flows for orificellow conditions (inc. safety factor)

O u t l e t A ( Q = 3 1 / s )Outlet A without leafguard (Q = 6 l/s)Outlet B with domed grating (Q =1.5 l/s)Outlet B with flat grating (Q = 3.4 Us)Outlet C (Q = 6.11/s)Outlet D (O = 2.0 l/s)O u t l e t E ( Q = 2 . 1 V s )Outlet F (Q = 18.1 l/s)Outlet GO u t l e t J ( Q = 5 1 / s )

vi l l sR473 0210996

Elntroduction

1.1 Scope of studyThe present repod describes work to investigate the performance ol flat roofoutlets which was commissioned by the Construction Sponsorship Directorate ofthe Depailment of the Environment (DOE). This laboratory study was carried outbetween April 1995 and March 1996 and arose form a need to obtain specificinformation on outlets used for drainage of flat roofs. The current British StandardCode of Practice BS 6367: 1983 "Drainage of roofs and paved areas" containsdesign recommendations for outlets obtained lrom earlier experimental workcommissioned by HR Wallingford on plain outlets installed in gutters. Outlets forflat roofs are normally litted with leafguards or gratings, and the approachconditions of the flow are often different from those of outlets in gutters. Nosuitable test data on outlets for flat roofs were available when BS 6367 wasprepared, so an approximate (and unverified) method of allowing for the effect ofgratings was suggested. The primary objective of the present study, thereforewas to develop generaldesign equations for tlat roof outlets for use in British andEuropean Standards.

The scope of the study also included an investigation of the performance ofsiphonic outlets for flat roofs as well as conventional outlets; in the first case,attention had to be given not only to the relationship between the flow rate andhead of water on the roof (the rating curve of the outlet), but also to other aspectsthat are relevant only to siphonic systems. Amongst these are: the ability of thesystem to exclude air from the pipework and start to act siphonically, and thenoise associated with the priming action. Siphonic systems have specificfeatures and design requirements when compared with conventional roofdrainage, which are not described in detail in lhis report. For a detailed reviewof these systems and assessment of their performance when the outlets areinstalled in gutters, see May and Escarameia (1996).

As mentioned above, one of the obiectives of the study was to measure thedepths of water on the roof for various f low rates and hence determine the ratingcurves of the outlets. lt is currently recommended in BS 6367 (1989) that thewater depth on flat rools should not exceed 30 mm to reduce the risk of infiltrationol rainwater into the building. A draft European Standard 'Gravity DrainageSystems lnside Buildings: Paft 3 - Roof Drainage, Layout and Calculation" (prEN12056-9) has been developed, and it is currently expected that it will be publishedin 1997 and thus supersede BS 6367. lt includes a recommendation that themaximum design depth of water on flat roofs should be 35 mm (although this limitcan be raised according to the padicular circumstances). lt is therefore importantto know the discharge that an outlet can convey without exceeding a water depthof 30 or 35 mm; in fact outlets are normally rated according to such a criterion.It was also the objective of the study to produce design equations that would besuitable for a wide variety of outlets, which can vary signif icantly both in size andshape, and have gratings or leafguards with extremely diverse characteristics.The equations would later be introduced in relevant design documents such asthe British Standard.

One of the primary objectives of the "Partners in Technology" programme of theConstruction Sponsorship Directorate of DOE, which padly funded this proiect,is to develop the involvement of industry padners in research. Contacts were,therefore, made with some of the major manufactures of roof damage systemsin order to establish a 'club" of manufacturers who would supply and, in manycases, inslall their outlets/systems in HR's lest facilily. Five companies agreed

sR473 30/0€/96

Eto participate in the project and it was decided that the resulls of particular testsshould not be directly linked to particular products.

This was done to allow the research to be concentrated on the scientific aspectsand to avoid direct comparisons of pedormance between the various outletstested. Forthese reasons the outlets described in this repoft are denoted by theletters A, B, etc. The five manufacturers involved in the study were (not in orderA, B, etc): ACO Fulbora (ACO Technologies Plc), FC Frost Ltd, Fullflow SystemsLtd, Geberit UK, Glynwed Foundries and Sommerhein AB.

The first chapter of this repofl explains the need for the study and its objectives,and also gives some background information on the theory of flow through roofoutlets. Chapter 2 describes the outlets and systems tested, whereas the testfacility used is described in Chapter 3. ln Chapter 4 the test procedure is outlinedand the results are presented. The analysis of resutts is explained and discussedin Chapter 5, where design equations are also presented based on the testresults; Chapter 6 gives the conclusions and recommendations.

1.2 Background theoryIt has been found experimentally that the flow, Q, through an outlet such as thoseused for roofs, can be uniquely related to the head of water above the outlet, h.For small depths of water the outlet is likely to behave as a weir with a lengthapproximately equal to the perimeter of the outlet. ln this case, the type of flowthrough the outlet is called weir flow and the lollowing relationship between Q andh is observed:

Q = cr hl 'u (1)

where c, is a numericalcoefficient of proportionality that depends, amongst otherfactors, on the length of the weir.

For greater depths ol water, the outlet will act as an orifice and the relationshipbetween Q and h changes, as follows:

Q = ce hos (2\

where c. is another numerical coefficient that depends on the area of the orifice.

Both coeflicients c, and 9 need to be determined experimentally in eachsituation, as they are affected by the approach conditions of the flow into theoutlets. These are mainly determined by: the location of the outlet in relation tothe building walls or parapets, and by the presence of gratings or leafguards.When the outlet is positioned close to a wall or parapet, this can produce swirlingwith entry conditions which are lar from radial and therefore have to bespecifically investigated. Also, gratings or leafguards will cause someobstruction to the flow and, although they may have a positive etfect in creatingradial flow, the obstruction will normally reduce the capacity of the outlet. Dueto the small water depths recommended for flat roofs, it is likely that the flow willbe of weir-type for most ol the time. However, when the outlet begins to choke(which will lead to submergence ol the outlet if the flow rate continues toincrease), the flow will transition from weir-type to orifice-type. This transitionmay be quite sudden for some outlets, or more gradual for others, in which caseit is rather difficult to def ine. Once the outlet is submerged, a small increase inflow rate will produce a steep increase in the head required over the outlet.

sF473 30/08/96

trThe current British Standard, BS 6367 (1983), recommends the use of thelollowing equations to determine the capacity ol grated flat roof oullets when testdata are not available:

e=L* h t 'uweir flow

24000

and

Q = A h o ' s12000

where Q is the flow (in Us), h is the head (in mm) above the top of the outlet, \is the total length of the openings measured around the perimeter of the outlet (inmm) along which flow enters the outlet and A is the totalarea of the openhgs(hmm2) that are submerged by the flow.

Values of h should be calculated from Equations (3) and (4) and the higher of thetwo values should be adopted for design. As explained in Section 1.1, thisapproximate method of applying Equations (3) and (4) to gratings was not verifiedas test data were not available when BS 6367 was prepared.

2 Description of outlets and siphonic systemstested

2.1 General considerationsConventional drainage systems for flat roofs usually consist of outlets whichdischarge into vefiical rainwater pipes;the connection between the outlet and therainwater pipe is left unsealed in many practical applications to ensure that acontinuous air core is present along the whole length of the pipe. This is asafeguard against negative pressures, which the pipes may not be designed towithstand, and against the uncertainty of having a variable driving head which isdependent on the flow through the system. Recommendations on the design ofconventional systems given, for example in BS 6367, were based on ventilatedrainwaler pipes and therefore the present work on grated outlets was carried outin a compatible way. The conventional outlets tested in the present studydischarged freely into the rainwater pipe by means of a shott pipe section, asexplained later in Section 3. When assembled in this manner, the capacity ofconventional systems is determined by the size and shape of the outlet and bythe diameter of the tailpipe.

Siphonic systems used for drainage of llat roofs differ from conventional drainagein that the outlets have special geometries to reduce the amount of air flowinginto the system, and the pipework connected to the outlets is designed to flow fullunderdesign conditions. ln flat roofs, where the water depth allowed on the roofis only of the order of tens ol millimetres, the driving head is practically equivalentto the difference in level between the outlet and the discharge point. Comparedwith conventional systems of similar outlet size, (which have a driving headequivalent to the water depth on the roof), the head of siphonic systems isconsiderably higher; this means that the overall capacity of the system issubstantially higher too. However, due to their greater complexity, siphonicsystems require a more careful and accurate design in order to perform

orifice flow

(3)

(4)

sR473 3008/96

traccoding to expectations. lt can therefore be inlerred that the capacity of thesesystems does not depend solely on the outlet, but also on the pipewoft that fonnsthe system.For fudher information on these systems refer to May andEscarameia (1996).

The present study involved the testing of lour different conventional outlets;these were installed in the test rig without a direct attachment to the rest of thepipework so that siphoning action could not occur. The test programme alsoincluded the assessment of the performance of four different siphonic systems;two of these outlets were also tested wilh a non-siphonic pipework system so thatthey could be compared with conventional outlets.

At the start of the test programme, it was agreed that the results would bepublished in this repod without identifying particular products or makes. Theoutlets were therefore referred to by letters of the alphabet: eg Outlet A, Outlet B,and so on. For simplicity, the siphonic systems tested were also denoted as"Outlets" in spite of the inaccuracy of such a term, as explained in the aboveparagraphs. Figures 1 and 2 show schematic diagrams of allthe outlets tested.The dimensions in the figures as well as the shapes of the outlets are onlyapproximate and are given as indications of the sizes of the outlets.

2.2 Description of the conventional outlets testedOutlet AMade of heavy duty PVC, Outlet A was designed primarily for use in flat roofs(see Plates 1 and 2). Wiih an approximate diameter of 200 mm, it included a 95mm deep sump with a shott spigot pipe which allowed tailpipes of differentdiameters to be used. The leafguard was approximately 315 mm in diameter (atthe base) by 185 mm high and made of polypropylene. Tests were carried outwith two different pipe sizes: 50 mm and 75 mm (extemaldiameters).

Outlet BThe body of Outlet B had afunnelshape with a vertical spigot 75 mm in diameter(extemaldiameter) and was made of aluminium. With a top diameter ol 230 mm,the outlet also included a clamping ring to which the isolating roof membrane isusually attached. Two different types of leafguard were tested: a domed grating

which sat inside the outlet on the clamping ring (see Plate 3), and a flat grating

which was fixed by bolts to the ring (see Plate 4).

Outlet COutlet C was funnel-shaped with a top diameter of 260 mm and a 75 mmdiameter spigot (extemaldiameter). All the components of the outlet (ie the outletbody, the leafguard and clamping rings) were made of cast-iron. The leafguardhad a domed shape and a height ol about 100 mm (see Plate 5).

Outlet DWith a square cross-section at roof level (220 mm x 22O mm) which then taperedto a 75 mm diameter spigot, Outlet D and all its components were fabricated in

cast-iron. The lealguard was also square in plan and approximately 60 mm high'as can be'seen in Plate 6.

Outlet EUsually forming part of a siphonic system, Outlet E consisted of a 60 mm deepsump, approximately 180 mm in diameter. The sump was made of stainless steelwith a streamlined underside. The spigot was 56 mm in diameter (extemaldiameter) and the leafguard was 75 mm high and made of plastic material, asshown in Plate 7.

sR473 300&96

EOutlet FOutlet F was similar to Outlet E but bigger in size: the top diameter of the outletwas approximately 295 mm and the spigot extemal diameter was 75 mm indiameter. The sump depth was, however, equalto that of the smaller outlet, ie 60mm - see Plate 8.

2.3 Description of the siphonic systems testedOutlet GOutlet G was a siphonic outlet which included: a solid air baflle with altematelarge and small vanes to straighten the flow, a clamping ring with a 184 mmdiameter gravelguard and a dome-shaped leafguard. The outlet wasapproximately 150 mm in diameter and had a shallow 20 mm deep sump (seePlate 9).

Outlets H and IThese outlets were similar to Outlets E and F, respectively, but for the fact thatthey were used as part of siphonic syslems, ie the outlet tailpipe was sealed tothe rest of the pipewok . A solid plastic air baffle, with a streamlined underside,is normally part of the outlet to help exclude air and therefore promote siphoningaction.

Outlet JAlso used as pad of siphonic systems, Outlet J had a sump approximately 85 mmdeep , was 275 mm in diameter and had a 75 mm spigot (external diameter). Theair baffle consisted of an inverted bowl with pedorated sides and slots on the top.The leafguard was made of plastic and was 130 mm high, as shown in Plate 10.

The pipework used in these systems is described later in Section 4.3.

3 Test rig

The pedormance of flat roof outlets was investigated in an existing test rig whichwas adapted specifically for the study (see Figure 3). The test facility consistedbasically of a2 m x 3 m wooden tank situated approximately 6 m abovedischarge level to allow the study of siphonic systems as well as conventionaloutlets. Flow to the tank was introduced from two sides (as opposed to a singleentry point), in order to produce uniform flow distribution and smooth entryconditions into the outlets. Baffles were incorporated in the tank for the samepurpose, as can be seen in Figure 3.

The testfacilrty was designed as a re-circulating rig, with water being drawn froma sump at ground level and pumped to the tank by means of a pump with acapacity of 30 Us. The flow through the outlet was then discharged back to thesump by pipework which varied in size depending on whether the pipework wasintended to operate conventionally or siphonically.

In the case of conventional outlets, the length of the tailpipe was always keptequal to the hydraulic diameter Dn; the flow from the tailpipe was::made todischarge freely into a larger diameter pipe which was not sealed to the tailpipebut was concentric with it. For circular tailpipes the hydraulic diameter Dr, is equalto the intemal pipe diameter since, by delinition Dr, = 4R = 4nD2 | 4nD = D, whereR is the hydraulic radius (=cross-sectional wetted arealwetted perimeter), and Dis the pipe diameter. The reason lor limiting the tailpipe length to Dn was to avoidmeasuring an artificial increase in discharge caused by longer pipes which willoften not be representative of practical roof drainage installations. With tailpipes

sB473 30,0996

trshorter or equal to Dn, the water nappe remains detached from the walls of thepipe whereas in longer pipe seclions the attachment of the nappe can causenegative pressures and result in higher discharge coefficients. In the case of thesiphonic outlets the pipework formed part of the system and therefore needed tobe included in order to achieve realistic operating conditions. As can be seenin Figure 3, the roof outlets were installed flush with the floor of the tank.

The floor of the tank incorporated a removable square plate in the centre tofacilitate the installation and removal of the outlets. The outlets were installeddirectly on the floor of the tank without inclusion of any material to simulate roofmembranes. Two drains were also included in the rig and discharged directly intothe sump.

The flow rate was measured by a 200 mm diameter electromagnetic flowmeter(Altometer, type SC80AS) which has an accuracy of +0.8%of the actualflow. Avalve was used to control the flow rate to the rig.

A total of eight tapping points were drilled in the floor of the tank for themeasurement of water depths, as shown in Figure 4. Tappings B, C and D werelocated 1m from either side of the outlet centreline, whereas the two tappings Awere set 400 mm away. Preliminary tests showed that, for the range of flowsstudied, this latter position was sufficiently close to the outlet to giverepresentative measurements of the water depths approaching the outlet, withoutbeing so close as to measure the drawdown of the water sudace at the entry intothe outlet. Tappings A were therefore used in the majority of tests. Transparentllexible tubes were connected to each tapping point and the water depths wereread against a scale f ixed on a piezometer board.

4 Iesfs

4.1 Test procedureThe test procedure was basically the same for conventional outlets and forsiphonic systems: the flow was increased in small steps untilthe submergencelimit was reached (ie when a small increase in flow coresponded to a steepincrease in water depth); after submergence had occurred, the flow was fudherincreased to try to measure the performance of the outlet under such conditions.

Readings were taken of the water depth on the f lat roof for each of the f low rates,after allowing sufficient time for establishment of steady conditions in the test rig.The flow increments in the tests of conventional outlets were of the order of 0.5[/s, but were in generalhigher in the tests ol siphonic systems due to their overallhigher capacity. Once the outlet was submerged, even small increments in theflow rate caused the water levels to rise significantly. In theory this situationcorresponds to a ma*ed change in the discharge curve of the outlet. However,the transition between an essentially weir-type flow (before the outlet issubmerged) to an orifice-type flow (when the outlet becomes submerged), is notalways very sharp, pafticularly when the outlet and/or the leafguard havecomplex geometries. For this reason, the judgement of the conditions at whichthe change occurs is necessarily somewhat subjective. In the present tests anoutlet was considered to be submerged when an increase in flow caused a steeprise in water depth.

4.2 Conventional outletsA list of the conventional outlets tested is presented in Table 1 . lt can be seen inthe table that tests were carried out with and without a leafguard or grating

sR473 3ry0€v96

Ecovering the outlet. This was done in order to assess the etfect ol the leafguardon the depth of water on the roof and on the entry conditions into the outlet. Intests A4 and A5 the outlet tailpipe was connected and sealed to the dischargepipe. These tests were outside the main tests but are of interest because theyshowed a substantial increase in capacity when compared with the adoptedconfiguration. This willbe discussed later, in Section 5.

ln some of the tests summarized in Table 1 an air baffle was introduced in theoutlet (for example Tests E3 and E4 ). Air baffles are features of siphonic outletsand their objective is to help exclude air from such systems and thereforepromote priming. Although not normally used in conventional roof drainage, airbaffles were included in some tests to see whether or not they induced higherflow capacities than those obtained in their absence.

The results of the tests are presented in Tables 21o7, which also include someobservations on the flow conditions at the outlets. Examples ol tests withconventionaloutlets are shown in the following plates: Plate 1 (Test A3, at 3Us),Plate 2 (Test A5, at 6Vs), Plate 3 (Test 82, at 1.51/s), Plate 4 (Test 83, at 3.4Us),Plate 5 (Test C2, at 6.1 Us) and Plate 6 (Test D2, at2Vs).

4.3 Siphonic systemsAs mentioned in Section 2.3, the tests ol siphonic systems involved threemanufacturers, one ol which supplied two outlets ol different sizes for testing.Table 8 lists the outlets tested and it can be seen that, with the exception of testsH2 and H4, all the tests were carried out with air baffles, as these are regularfeatures of siphonic systems. In most tests the standard leafguards were alsofitted.

ln siphonic systems the design capacity depends not only on the sizes of theoutlet and tailpipe, but is also strongly dependent on the layout and size of thepipework that constitute the systems. Therefore it is important to design thesesystems accurately in order to satisfy the specific requirements of eachinstallation. In the present study, the manufacturers involved were asked toproduce a design which was suitable for the conditions of the flat roof rig. Theseconditions were: a vedicaltailpipe of about 1m length and an approximately 3.8m long horizontalsection, followed by a verticaldrop of about 5m to the dischargepoint.

The layout and sizes of the pipes that formed the drainage systems for OutletsG, H and I are shown in Figure 5 and Table 9. lt can be seen in the figure that avalve was inlroduced at the downstream end of the downpipe to controlthe flowat that point. This valve was part of the system installed by the suppliers of OutletG, and, as it was left in the rig, the suppliers of Outlets H and I were asked toaccommodate it in their design. Both tests with Outlet G presented here werecarried out with the same degree of valve closure. The valve was keptcompletely open in tests Hl , H2,11and 14 but was padially closed in the otherfourtests to reduce the capacity of the system to approximatety: 6l/s (in tests H3and H4), 21.51/s (in test 12) and 12lls ( in Test l3). In each of these cases thesetting ol the valve was carried out prior to the tests in the following manner: thevalve was gradually closed to cause submergence of the outlet at the requiredflow rate; once the correct valve position was achieved, no fudher adjustmentswere made and the tests proceeded as described in Section 4.1. The systemdesigned to incorporate Outlet J is presented diagrammatically in Figure 6,where it can be seen that the control valve was omitted and the pipe diameterwas constant along the whole system.

sR473 3Cv08196

trExamples of the tests are shown in: Plate 7 (Test H3, at 2.1Vs), Plate 8 (Test ll,at 18.1[/s) and Plate 1O (Test J1, at 5Us). Tables 10 to 13 summarize the resultsof tests carried out with siphonic systems.

5 Analysis and discussion of results

5.1 Rating curvesThe values of flow rate and water depth on the roof, which are presented inTables 2 to 13, were plotted in Figures 7 to 16 to illustrate the stage-dischargecurves of each outlet. In these curves lhe flow rate is given in Us and the waterdepth on the roof in mm; for most types of outlet the curyes include tests caniedout with and without leafguards and, in the case of some siphonic outlets, plotswere also produced fortests with and without the air baffles so that their effect onthe performance of the systems could be determined.

The limiting capacilies ol the outlets, which are also discussed in the followingsections, were taken as the flow rates for which the flow transitioned from weir-type to orifice-type (see Section 4.1). Mention will also be made in the nextsections of the flow rates that corresponded to measured water depths on theroof of 30 mm and 35 mm. These values are, as explained in Chapter 1, thelimiting depths recommended at present in BS 6367 and in the draft EuropeanStandard prEN 1 2056-3, respectively.

5.1.1 Conventional outletsThe stage-discharge curues obtained from tests A1, A2 and A3 of Outlet A arepresented in Figure 7. Curves for Tests 44 and A5 were not produced sincethese tests were not carried out with a ventilated tailpipe and would not becomparable with results for other outlets. However, it is interesting to notice inTable 2 that the ultimate capacity of the outlet was greatly increased when thetailpipe was sealed (an increase of more than 600%); this indicates that thesystem possibly developed a strong siphonic action in which the air wasremoved from the pipework and the pipes were therefore flowing full. Comparingthe curues for Test Al (with leafguard) and Test A2 (without leafguard), it can beseen that there is an increase ol about 20% in the capacity of the outlet when theleafguard is not present.

Figure 8 shows the results obtained for Outlet B: Test 81 (without leafguard), Test82 (with a domed leafguard) and Test 83 (with a flat leafguard). For llow ratesbelow approximately 5 l/s the three curves are quite similar; the differences areonly significant in terms of the limiting capacities. The domed leafguard reducesthis capacity by 3% whereas the flat leafguard reduces it by about 18% whencompared with the outlet without a leafguard. Outlet B with the domed leafguardproduces a water depth on the rool of 30 mm for a flow rate of approximately 8l/s; with the flat leafguard the outlet becomes submerged before that depth isachieved.

The rating curues for Outlet C are shown in Figure 9: Test Cl (without leafguard)and Test C2 (with leafguard). !t can be seen that, although for flows below 5 l/sthe two curves are very similar, the limiting capacities of the outlet in the two testsare substantially different, with an increase of 28Yo when the leafguard isremoved. With the leafguard, Outlet C produces 30 mm and 35 mm of waterdepth on the roof lor 7 lls and about 7 .4 Vs, respectively.

The curves produced lor Outlet D, which are shown in Figure 10, indicate asignificant increase in limiting capacity (around 53%) when the leafguard is not

sF473 3ryO&96

Epresent. lt is also apparent in the Figure that the shapes of the two curves arequite different, even for small flows. Water depths of 30 mm and 35 mm areachieved for flows of about 6 l/s and 6.5 Us, respectively, for the outlet withleafguard.

The results of the four tests canied out with Outlet E are plotted in Figure 1 1 (twowithout air baflle - Tests E1 and E2 - and two with the air baffle normally usedwhen the outlet is part of a siphonic system - Tests E3 and E4). lt can be seenthat the two curves without the lealguard - Tests E1 and E3 - are practicallycoincident (except for flows close to the limiting capacity), as are the other twocurues with the leafguard - Tests E2 and E4.

Figure 12 shows the rating curues for Outlet F. As obserued for Outlet E, the twocurues without leafguard ( Tests F1 and F3) are significantly different from thosewith leafguard (Tests F2 and F4). Contrary to what was observed in all the othertests, the presence of the leafguard does not appear to decrease the limitingcapacity of the outlets (compare Tests F1 with F2 and F3 with F4); in fact, a smallincrease can be seen with the leafguard in Tests Fl and F2. As for Outlet E, theair baffle appears to reduce the limiting capacity of the outlet, by between 10%and 15%.

5.1.2 Siphonic systemsThe rating curues forthe siphonic system incoporating Outlet G with and withoutits leafguard (Tests Gl and G2, respectively) are shown in Figure 13.These twocurves are very similar and the capacity of the system (25 Us) is little affected bythe presence or absence of the leafguard. This system was found to dischargeapproximately 4 Vs and 6 l/s with water depths on the roof of 30 mm and 35 mm,respectively.

All of the four tests of siphonic systems incorporating Outlet H were carried outwith the leafguard and are presented in Figure 14. In Tests H3 and H4 thecapacity of the system was limited to 6 l/s, as explained in Section 4.3, whereasin Tests Hl and H2 the valve at the downstream end of the pipework was keptcompletely open. The plot of Test Hl shows an irregularity in the rating curve atabout 5 to 6 Us, when the outlet was observed to become submerged and thewater depth increased significantly; however, after about 10 minutes the siphonicaction was recovered and the ultimate capacity achieved was about 23 l/s. Thisphenomena was not obserued during Test H2 which was carried out without theair baffle but the water depths measured were signif icantly higher than in Test H1for flows above 'l5 Vs. The presence of the air baff le caused a decrease in thelimiting capacity between 8"/" and 17"/" (compare Tests Hl and H2, and H3 andH4). ln general, Outlet H was found to discharge approximately 5.5 Us with 30mm of water on the roof and 6.5 l/s with 35 mm.

All the tests carried out with Outlet I included the leafguard and the results areshown in Figure 15. The differences between the results of the various testsreside mainly in the ultimate capacity of the system; this was determined byadjustment to valve positioned at the downstream end of the pipework (Tests 12and 13) and by changes in the tailpipe diameter (Test 14). As can be seen inFigure 15 and Table 12,the capacity of the test rig was reached before the outletbecame submerged for Test 14. Whatever the pipework in the system, in generalterms Outlet I was found to discharge about 8 Us with 30 mm of water depth onthe roof and about 11 Us with 35 mm of water.

The results ol tests wilh Outlet J (Test J1 with leafguard and Test J2 withoutleafguard) are plotted in Figure 16. Padicularly for flows above approximately 10

sF47a 3ry08r96

trl/s, the leafguard tends to increase the values of water depth on the roof. For awater depth of 30 mm, the system with leafguard was found to discharge about10 l/s and, for a water depth of 35 mm, approximately 12.5 l/s.

5.2 Development of design equations for flat roof outletsIt was mentioned in the Introduction of this repod that BS 6367 (Appendix B)recommends the use of weir flow and orifice flow equations for design of outletsinstalled in roof gutters. The development of these equations was based onexperimental measurements carried out in gutters but they did not cover thespecific case of grated outlets in flat roofs. The British Standard thereforerecommends that grated outlets should be tested to determine their flow capacity;however, if test data are not available, the capacity of grated outlets can beestimated approximately by applying Equations (3) and (4) which are those alsorecommended lor plain rectangular outlets in gutters.

The present tests were carried out with the ultimate oblective of developingsuitable equations for grated outlets in llat roof conditions. Based on previousexperience and on visual obseruation of the flow during the tests, it was decidedto investigate if the data for unsubmerged conditions could be well representedby a weir-type equation. Allthe data collected in the tests before the outletsbecame submerged were plotled in a log-log graph (Figure 17) where it can beseen that the power of h in Equation (1) is approximately 1.5. This findingconfirms the assumption of weir-type flow for most of the flow conditions in thetests.

5.2.1 Effective diameterThe design equations in BS 6367 for circular outlets depend, amongst othervariables, on the effective diameter of the outlet. These equations are of lhefollowing form:

for weir flow

and

for orifice flow

where ( and Iare numerical coefficients and D is the effective diameter of theoutlet. ln orderto develop design equations for flat roofs that are compatible withthe existing equations for outlets in gutters, it was necessary to deline what ismeant by 'effective diamete/'for the case of flat roof outlets. For this purpose,each outlet was carefully inspected and schematic diagrams were produced tohelp identify the main features and observe the most relevant details.

It was decided to study the shape of each outlet and the position of the leafguardin relation to the outlet (ie whether the leafguard is set inside the outlet bowl or ispositioned surrounding the outlet). These pararneters are impodant with weir'typeflow, fordetermining where the water passes through critical depth; the length ofthe outlet perimeter at this location has an impodant effect on its flow capacity.The complex shapes of some of the outlets tested, as well as the wide variationin the types of leafguard available, introduced some difficulties, as expected.Because of the variation in shapes of tlre outlets in the UK market, it is necessaryto study the shape of each individual outlst in order to estimate its etfectivediameter. lt is particularly important to notice the relative sizes of the leafguard

(s)Q = D h 1 ' 5

K1

(6)Q = D h o ' ut9

10 s8473 3ry08/96

Eand outlet and to observe il the outlet contains an intemal ring to fix the roofmembrane, in which case the point of critical flow may be located at the ring. lt isalso important to identify the position ol gravel guards as they can producecritir:alllow conditions and therefore determine the effective diameter of the outlet.Four major types were identified f rom all the outlets tested and are presented inFigure 18.

Forthe determination of the effective diameter, D, it is recommended to use the{ollowing guidelines:

1. Type | - lf a gravel guard is present, D = Dg,"*rsu",u

2. Type ll - lf 1.6 ) Db".*d/ Do,.r",> 1.2, D = Dh"tg*,u

3. Type lll - lf 0.8< Ds"re*,0 / Dod,, < 1.2, D = Dor."t

4. Type lV - lf Db"rsr"d / Do66, '. 0.8, D = Doutret rins (or D = Dod". if the outlet has noclamping ring)

In the above relationships Du"rg*,o corresponds to the basal diameter of theleafguard; if the outlet has a round edge, Dorr, is measured f rom the point wherethe flow will tend to drop into the outlet , ie from the stafi of the outlet curvature(see, for example, Outlet B in Figure 1). These guidelines resulted from analysisof the experimental data in terms of the value of the numerical coefficient Kt inEquation (5);this analysis is explained in detail in the next section.

5.2.2 Weir flow equationThe value of coefficient K' in Equation (5) was determined for all themeasurements with a leafguard before submergence occurred, taking intoaccount that the effective diameter, D, varies with the size and shape of theoutlet. This was done in the following manner: 1. an average value of K.' wascalculated for each outlet; 2. comparison of geometrical data from allthe outlets,including gratings, gravel guards and leafguards, led to the determination ofeffective diameters D (as described in Section 5.2.1) which produced anapproximately constant value of K, for all outlets (this value of K' was determinedbased on test data for convenlional outlets with leafguard); 3. the flow rates forallthe data including tests with leafguard and siphonic system were predictedusing Equation (5) with the new value of K' and were plotted against themeasured flows (see Figure 19 - data points represented by diamonds); 4.coefficient K, was then revised to incorporate a safety factor of about 1.25 interms of the flow rate into the design equation (see data points represented bysquares in Figure 19). The equation recommended for design of circular outletsin weir-type conditions is:

o= D h 1 ' 5

6000

where Q is in Vs, D is the effective diameter in mm as delined in Sections 5.2.1and h is in mm. This equation is valid for all the types of outlet and leafguardtested (see Figure 1); in the case of square outlets, D is the side length (also inmm).

The presence of the leafguard produced, in some tests, an increase in the flowdepths on the roof when compared with tests without the leafguard (see, forexample tests F2 and Fl in Figure 12); however, in several cases, the leafguard

(7)

1 1 s8473 3CV08v96

Ewas found to have only a marginal effect on water depths, padicularly lor low llowrates (compare, for example, Tests C'l and C2 in Figure 9). lt was also observedthat the limiting capacity of outlets can be increased if the leafguard is removed.This increase varied from practically zero to as much as 53% (for the squareOutlet D), but in the case ol Outlet J the capacity actually decreased slightly byabout 3% when the leafguard was not present. These very differing results areobviously due to the wide range of shapes, sizes and types of leafguard tested.It is possible, however, to determine an average value of the coefficient K' inEquation (5) for outlets without a leafguard. Adopting the same procedure usedfor determining Equation (7), the resulting equation (including a safety factor olabout 1.25) is given by:

o= D h 1 . 5

5000

where Q is in Vs and D and h are in mm. Compared with Equation (7), Equation(8) represents an increase in flow of 20"/". In spite of the likely benefits of theremoval of the leafguard in the capacity of the outlet, it should be noted that theleafguard is an important part of flat roof drainage systems as it prevents theentry of debris (and particularly ballast stone) into the pipework. Also theleafguard is vital for siphonic outlets in order to prevent the small openingsaround the air baffles becoming blocked.

5.2.3 Orifice flow equationThe amount of data available from the tests corresponding to the near-submergence of the outlets was substantially less than for weir-type flowconditions; it was also found that the scatter of the results was much greater.This meant that the agreement between measured flows and flows predicted witha design equation would not be as satisfactory as that obtained for weir-typeconditions. lt should also be noted that the results obtained for the siphonicsystems were not included in this analysis. Following the procedure describedin the previous section, the value of coefiicient lt in Equation (6) was determined,and the measured and predicted flows with the value of K, were plotted in Figure20. With a safety factor of 1.2 (see Figure 21), the recommended designequation for orifice-type flow with leafguard is as follows:

o= D2 h0 .5

66000

where Q is in t/s, D is the effective diameter in mm defined as in Section 5.2.1 andh is also in mm. This equation is not valid for siphonic syslems because the flowcapacity also depends on the size of the pipework and the overall height of thesystem. The water depth above which weir-type flow transitions to orifice flowwas found to be approximately given by:

(8)

(e)

h t 9 for circular outlets1 1

(10)

Equations (9) and (10) can only be applied with confidence to circular outlets dueto the insufficient data collected on square outlets under orifice'type flowconditions.

' t2sR473 3&08/96

tr

2.

6 Conclusions and recommendations

1. There has been a long-standing need for data on the f low capacity ol gratedoutlets used forflat roofs since the design formulae given by BS 6367: 1983'Drainage of roofs and paved areas" were based on data collected from

tests ol plain outlets installed in gutters. The effect of gratings or leafguadson the capacity of outlets used in flat roofs, where the water depths are limitedto a few tens of mm, is likely to be significant.

The objectives of the present study were:

(a) to measure the pedormance of grated flat roof outlets used inconventional and siphonic drainage systems;

(b) to produce lormulae for the capacrty of the outlets to be included in BS6367 (1983) or the new European Standard prEN 12056-3 which isdue to supersede the British Standard in 1997;

(c) to assist manulacturers to produce improved designs of outlets.

A comprehensive laboratory study was carried out to determine thepedormance of grated outlets used for flat roofs. lt involved the testing ofsix conventional outlets and four outlets used in siphonic systems under arange of flow conditions. The outlets and leafguards tested had a widevariety of shapes and sizes which are schematically presented in Figures1 and 2.

Rating curyes for each outlet were obtained and are shown in Figures 7 to16. For each type of outlet, these figures also allow estimates to be madeof the flow rates that correspond to the maximum water depths on the roofrecommended by BS 6367 (30 mm) and by the future European Standard(35 mm).

In the analysis of the test results two types of flow were considered: weirflow and orifice flow. This follows the approach adopted in BS 6367. Forlower water depths on the roof, the outlet acts as a weir and the flow iscontrolled by its perimeter; as the water depth increases, the outlet acts asan orifice and the flow is controlled by the cross-sectional area of the outletin plan.

Design equations, which incorporate a safety factor of approximately 1.25compared with the best-fit coefficient, were developed based on datacollected both for convenlional outlets and siphonic systems. The safetyfactor allows for some variability in this amount of blockage of the leafguardcaused by leaves and other debris.

The formulae presented are valid only within the range covered in the study,lor

2 . s < D o u t l " t . 4 . s

D.,,0,0"

and

3.

4.

5.

6.

te

g.5 . Dlearguaro

.1 .6Dour|et

where Do*o" is the internal diameter of the outlet pipe.

'13sR473 3ry@196

E

7.

For weir-type flow the capacity of an outlet can be estimated by Equation(7):

o= D h l . s

6000

where Q is the flow rate in l/s, D is the effective diameter (or side length ofa square outlet) in mm and h is the water depth on the roof , also in mm.

This equation was derived assuming that the outlet is installed with alealguard; if the leafguard is not present, the numerical coefficient in theabove equation should take the value of 5000. This represents an increaseof 20% in capacity. However, for impodant practical reasons (such as theneed to exclude debris from the drainage pipes), it is recommended toinclude the leafguard in flat roof installations.

For orifice-type flow, the recommended design equation (Equation (9),

which was developed for circular outlets due to insulficient data on squareoutlets) is as follows:

o= D2 ho.s66000

where Q is in l/s, D is in mm and h is also in mm.

The calculation of the effective diameter, D, used in the recommendeddesign equations depends on the type of outlet under consideration. Duringthis study il was possible to identify four different types (see Figure 18):

1. Type | - lf a gravel guard is present, D = Dsrawrsmrd

2. Type ll - ll 1.6 > Dr""rs*rd/ Do'r.r > 1.2, D = Dbaremd

3. Type lll - lf 0.8< Dburs,"d / Do*6, r 1.2, D = Dorr",

4. Type lV - lf Dburq*,6 / Do*6, ., 0.8, D = Doutret ring (or D = Dou&r if the outlethas no clamping ring)

ln the above relationships Do.re,",a corresponds to the basal diameter of theleafguard; if the outlet has a round edge, Do*", is measured from the point

where the llow willtend to drop into the outlet , ie from the stail of the outletcuruature (see, for example, Outlet B in Figure 1).

Partly due to the complex shape of the leafguards, it was found in the tests

that the transition belween weir and orif ice flow is not always well defined,and in some cases can be very gradual. ll was, however, possible todetermine an approximate relationship that defines the transition point lorcircular outlets. This relates the effective diameter of the outlet with thehead of water on the roof and is given by Equation (10):

h t 9 lor circular outtets.1 1

8.

1 4 sR473 3W0&96

tr9. The study also allowed some conclusions to be drawn regarding the

specific case ol siphonic systems used for drainage of flat roofs' Allthesystems tested were found to have capacities similar to those predicted bythe manufacturers, but it was apparent that careful design of the pipewokis necessary to guarantee adequate performance. Inadequate pipe sizecan impair the priming of the system and eventually give rise to surging andviolent expulsion of air. With adequate design, the siphonic systems testedshowed smooth operation conditions and low levels of noise.

Atthough the air baffle is an essentialfeature for outlets in siphonic systems,the tests provided some interesting information. A comparison of testscanied out with and without the air baffle, showed that the design capacityol the system was reduced by between 8 and 17"/" when the baffle waspresent. This effect was only noticeable on the value of the limitingcapacity and was not appreciable at lowerflow rates. The obstruction to thellow caused by the air baffle, which promotes the exclusion of air andlherefore the priming of the system is believed also to accelerate thetransition from weir to orifice flow and therefore the submergence of theoutlet.

For grated flat roofs outlets with shapes that dilfer significantly from thetypes tested, it is recommended to carry out laboratory measurements todetermine their capacities.

7 Acknowledgements

The collaboration from the following companies, who provided roof outlets anddrainage pipework for testing is gratefully acknowledged: ACO Fulbora (ACO

Technologies plc), FC Frost Ltd, Fullflow Systems Ltd, Geberit UK, GlynwedFoundries and Sommerhein AB. Some of the laboratory tests described in thisrepoil were carried out with the assistance of Ms C Hide.

1 0 .

1 1 .

1 5 sR473 30/0€/96

E8 References

1 BS 6367: 1983. 'Drainage of roofs and paved areas". British Standardslnstitution, 1983.

2 European Standard'Gravity Drainage Systems Inside Buildings: Pad 3 -

Roof Drainage, Layout and Calculation" prEN 12056-3. Draft version of20 .1 .95 .

3 May R W P and Escarameia M, 1996. "Perlormance of Siphonic DrainageSystems for Roof Gutters". HR Wallingford Repofl SR 463.

1 6 sR473 3ry08/96

tr

Tables

sR473 3ry0€r96

trTable 1 Details of test conditions for conventional

outlets

Outlet Tailpipeint.diameter(mm)

Leafguard Ventilatedtailpipe

Air baffle Testnumber

A 5050757s75

X

X

xx

X

X

X

X

X

A1A2A3A4A5

B 757575

X

/ domed/ tlal

X

xX

B18283

c 7575

X xX

c1C2

D 7575

X X

xD1D2

E 52525252

X

X

X

X

E1E2E3E4

F 70707070

x

x

X

xF1F2F3F4

sR473 3ry0€r96

trTable 2 Results of tests with Outlet A

Test a (Ys) Water depth (mm) Observations

A1 1 . 52.02.53.0

1 1 . 812.51 6 . 8

Levels notstabilized

Choking at outletOutlet submerged

A2 1 . 52.O2.53.1

9 .51 0 . 512.O


Outlet submerged

A3 1 . 52.O2.73.74.85.0

9.012.O16.021.024.4


Outlet submerged

A4 1 . 52.O2.53.03 .54 . 15.06.49.614.21 9 . 523.326.628.630.532.0

' t2.0

14.017.01 9 . 021.024.O27.033.04 1 . 052.060.072.O76.s79.084.587.5 Outlet still not submerged

A5 1 . 52.04.O6.01 0 . 114.2

7 .51 0 . 017.021.533.541.O

Choking at outlet

Outlet still not submerqed

sM73 390&96

trTable 3 Resulfs of tests with Outlet B

Test a (Ys) Water depth (mm) Obseruation

B1 1 . 62.02.83 .24.35.26.46.97.99.0

8 .010.01 3 . 014.O18.021.O25.O26.028.0



82 1 . 52.02.73.34.25.36.36.87.98.7

9.010.011 .514.5'17.o

19.023.025.030.0


Choking at outlet

Outlet submerged

B3 1 .52.02.73.44.25.26.26.87.6

8.51 1 . 01 3 . 51 5 . 518.523.024.O26.5


Choking at outlet

Outlet submerged

sR473 3ry0&96

trTable 4 Resu/fs of fesfs with Outlet C

Test o (us) Water depth (mm) Obseruations

c1 1 . 52.73.34 .25.36.26.97.88 .18.79.39.61 0 . 110.41 1 . 1

8.012.O14.017.51 9 . 521.523.524.525.527.530.033.056.067.5


Choking at outlet

Outlet submerged

c2 1 . 52.O2.73.34 .25 . 16 .16 .97.98.7

7.09.01 1 . 512.517.51 9 . 524.O30.546.5


Choking at outlet

Outlet submerged

sR473 3{v08r96

ETable 5 Resulfs of fesfs with Outlet D

Test o (Us) Water depth (mm) Obseruations

D1 1 . 51 . 92.83.34.35 .16.26 .97.78.39 .410.4

9.510.013.51 5 . 518.523.O25.O26.026.530.032.5


Choking at outlet

Outlet submerged

D2 1 . 62.02.83.34.35 .16.36.8

8.010.514.s16.523.025.533.0


Choking at outlet

Outlet submerged

s8473 3d0€v96

trTable 6 Besu/fs of tesfs with Outlet E

Tesl o (l/s) Water depth (mm) Obseruations

E1 1 . 42 .12.73.33.74.3

7.O1 1 . 014.01 6 . 020.0


Choking at outlet

Outlet submerged

E2 1 .52.O2.53 .13.73.9

1 1 . 01 5 . 017.520.523.O



E3 1 . 52.02.73.33.6

8.010.014.01 6 . 0



E4 1 .51 .92.63.33.5

12.01 4 . 517.524.5


Outlet submerged

sR473 3d08v96

trTable 7 Resu/fs of fesfs with Outlet F

Test a (Ys) Water depth (mm) Obseryations

F1 1 . 42.53.74.65.96 .97.5

6 .59.01 3 . 01 5 . 518.521.5


Choking at outlet

Outlet submerged

F2 1 . 42.83.34.65.56.87.8

10.014.016.01 9 . 521.527.O


Choking at outlet

Outlet submerged

F3 1 . 42.53 .54.85.76.8

7.O10.012.516.0'18.5


Outlet submerged

F4 1 . 72.53.34.65 .56.36.8

12.O13.016.020.022.026.5


Surging at outlet

Outlet submerged

sR473 3d0€/96

trTable 8 List of siphonic outlets tested

Outlet Outlet pipe sizeInt diameter (mm)

Lealguard Air baffle Test number

G 6969 X

G 1G2

H 52525252

X

x

H 1H2H3*H4*

69696969

t l12*"l3***140

J 6969 X

J 1J2

* System set up for capacity equal to 6 l/s** System set up for capacity equal to 2l.5 Vst*t System set up for capacity equal to 12 UsO 110 mm diameter tai lp ipe

ETable 9 Pipe diameters in siphonic sysfems G, H

and I

Note: See Figure 5

Pipe number Test number Ext. Diameter(mm)

lnt. Diameter(mm)

1 ,2 ,3 G1, G2, H1.H4,t1-14

1 1 0 104

4 G 1 , G 2 , H 1 - H 4 ,l 1 - 1 4

751 1 0

69104

5 G1, G2, H1-H4,l1 -13

1 4

75

1 1 0

69

104

ETable 10 Flesu/ts of tesfs with siphonic system G

a (ys) Water depth (mm) Obseruations

Test Gl1 . 73.67.01 0 . 113.61 6 . 12'1.623.825.2

2029

36.542.54954

6 1 . 565.5

Levels not stabilized Outlet submerged

Test G21 . 63.57.01 0 . 01 3 . 016.221.s25.0

1 827374345

5 1 . 560.5

Levels not stabilized Outlet submeroed

trTable 11 f,lesu/fs of fesfs with siphonic system H

o (l/s) Water depth (mm) Obseruations

Test Hl1 .42.O2.63.44.05.05.76.37.O8.29.09.711 .012.214.316.017.s19 .420.522.223.5

12.51 5

1 8 . 521.524.528.547.553

34.537

39.54 1 . 544.547

5 1 . 554.558.559.567.570.5

Levels not stabilized

lntermittent submergenceIntermittent submergence

Outlet submerqed

sR473 3ryOS/96

trTable 11 f,lesu/fs of fests with siphonic system H

continued

Test H21 . 42.54.O5 . 15.46 . 16.87.48.39.410.2' t0.9

12.614.416.317.318.21 9 . 120.421.622.823.724.525.3

1 21 9 . 52630

31.s3437394 144

45.547.552.556.56769

70.5737679

80.58 1 . 583.5


a (|/s) Water depth (mm) Observations

Test H31 .52.12.53.13.84.24.95.66.16.88.0

1 21 5 . 51 82124

25.52830

3 1 . 535.5

Levels not stabilized Outlet submerqed

sR473 3ryO8/96

trTable | 1 filesu/fs of fesfs with siphonic system H

continued

Test H41 . 52.22.83.34 .14.85.56.77.48.49.4

1217

20.5222528

31 .5353841

Levels not stabilized

s8473 3ry0&96

trTable 12 Results of tests with siphonic system I

o (us) Water depth (mm) Observations

Test ll1 . 52.73.65 .16.47.48.810.512.51 5 . 11 8 . 120.322.924.425.426.326.927.728.',|28.729.530.1

1 01 41 72 125262933364 14549

52.556

56.55859

59.562

62.570


Test 121 . 62.63.95.26 .17.38.79.010.212.513.715.316.717.719.019.720.2

1 01 4

18.5222528

30.53235

37.539

42.545464850

Levels not stabilized Outlet submerqed

sR473 30/08196

trTable 12 Resulfs of fesfs with siphonic system I

continued

Test 131 .42.53.85.98.310.2

I1 61 8

23.529.5


a (ils) Water depth (mm) Observations

Test 141 . 52.55.37.88 .110.012.81 5 . 117.61 9 . 922.625.027.330.032.535.135.936.2

10.51 6 . 523

29.53235404448

5 1 . 554.559

6 1 . 563

66.572.572.574 The capacity of the test rig

was reached

sR473 3cy0€v96

trTable 13 Results of tesfs with siphonic system J

a (us) Water depth (mm) Obseruations

Test J11 .63.05.07.29.3't't.2

12.012.713 .514.016.318.320.o

1 11 4212328323435374 1

42.546


Test J21 . 83.05.07.09.31 1 . 112.115.217.81 9 . 11 9 . 5

1 01 31 9222629

3 1 . 5364 1

44.5Levels not stabilized Outlet submeroed

sR473 gdOS/96

tr

Figures

sR473 2205196

tr

;fc o l-l

T:1

| 230m

Outlet B with flatleafguard

_f*l-ro l' l

Y

Outlet D (square) NOTTO SCALEApproximate dimensions in mm

Outlet A Outlet B with domed

Figure 1 Schematic diagrams of outlets A, B, C, D and E

r 295

ffi

-fl .o l*l

tY_T

o l' lY

:l*l-ro l' l

Y

| 275

ffi

Outlet J

NOTTO SCALEApproximate dimensions in mm

, 1 5 0 1

F-f

Io lo l- l

L *a(\l_T

baffle

Outlet H Outlet I

Figure 2 Schematic diagrams of outlets F, G, H, I and J

Sectional view

e(\l

Dimensions in metres

Plan

M91/1-96'/ l.line

Figure 3 Schematic of test rig

sR473 22/05/96

L

0.6 0.4 0.4 0.6

Plan

Dimensions in metres

BX

cX

DX

BX

cX

DX

AX

MEPttl-96il Llhe

Figure 4 Location of tapping points

tr

Outlet

II

Discharge

-1000

NOTTO SCALEDimensions in mm

Figure 5 Schematic diagrams of siphonic systems G, H and I

NOTTO SCALEDimensions in mm

Schematic diagram of siphonic system J

JT

Outlet A30

25

20e915-c

1 0

5

03

a (/s)

Figure 7 Rating curve of Ouflet A

Outlet B40

30

EE 2 0

1 0

0

a (rs)

Figure 8 Rating curve of Outlet B

Outlet C80

60

E9 4 0-c

20

06

a (rs)1 0 1 2

Figure 9 Rating curve of Ouilet C

Outlet D40

30

Es -20

1 0

06

a (/s)1 0 1 2

,"

Figure 10 Rating curve of Outlet D

7I

Outlet E30

25

^ 2 0Eg1s-c

1 0

5

0

//'

./ --,/ -/

/ '- '-

2a (Ys)

Figure 11 Rating curve of Ouilet E

Outlet F

EE-c

30

20

1 0

0

Figure 12 Rating curve of Outtet F

Figure 13 Rating curve of Ouilet G

Figure 14 Rating curve of Ouilet H

Outlet I80

60

E9 4 0-c

20

01 0 20

a (Ys)40

Figure 15 Rating curve of Outlet I

Outlet J

Figure 16 Rating curve of Outlet J

I

ci

qo

ooci

ooIo

E-

o oo

ooo

ooeo

(sleu) o

Figure 17 Weir flow data

tr

1 . 6>Dlgafgusrd/Doutlet>1 .2

D = Dgravelguard

Type I

D = Dleafguard

Type II

, Doutlet ,r-?--r , Doutler Ir------1Leafguard a)

fguard b)

0.BcDlsafgugrd/Doulet<1 .2 DteatguardDoulet<0.B

D = Doutlet

Type III

D = Doutlet ring (or Doutlet)

Type IV

NOT TO SCALEApproximate dimensions in mm

Leafguard

Figure 18 Definition of effective diameter

sR473 2A05196

I

rifoo

(f)

qo

o(f)

EN O9 oo o

U'6o

eo

sqo

(f)

qo

oteo

(sleu) O palctpard

qo

T

I

I@

oI

I

tT

'1R.t@f

T@

. @vA(\@

I

Figure 19 Comparison of measured and predicted flows for weir flowconditions

E

Orifice flow1 0

8o=,T ) 6o.9'tf,E 4o-a

2

0

I

II

r l r

I

4 6Q measured (l/s)

1 0

Figure 20 Gomparison of measured and predicted flows for orificeflow conditions

Orifice flowlncluding safety factor

1 0

8a\E 6o.9EO Ao

o2

T

r !

r t t r I

I

4Q measured (l/s)

1 0

Figure 21 Comparison of measured and predicted flows for orificeflow conditions (inc. safety factor)

sR473 2205196

tr

Plates

sA473 22lo5l

tr

Plate 1 Outlet A (O = 3 l/s)

Plate 2 Outlet A without leafguard (Q = 6 l/s)

tr

Plate 3 Outlet B with domed grating (Q = 1.5 l/s)

Plate 4 Outlet B with flat grating (Q - 3.4 l/s)

tr

' -+. . . . =a

Plate 5 Outlet C (O = 6.1 l/s)

Plate 6 Outlet D (O = 2.0 t/s)

tr

Plate 7 Outlet E (O = 2.1 l/s)

Plate 8 Out let F (O = 18.1 l /s)

tr

Plate 9 Outlet G

Plate 10 Outlet J (O = 5 l/s)

performance of flat roof outlets - hr wallingford · tef: + 44 (0)1491 835381 fax:. + 44 (0)1491...

Documents