epa water treatment manual filtration 1

8/6/2019 Epa Water Treatment Manual Filtration 1

http://slidepdf.com/reader/full/epa-water-treatment-manual-filtration-1 1/80

4

C)

An GhnIomhaireacht urn ChaornhnUComhshaoil

WATER TREATMENT

MANUALS

FILTRATION

This document

contains 80 pages

;f3Lf8

ENVIRONMENTAL PROTECTIONAGENCY



©Environmental Protection Agency 1995

Parts of hispublication maybereproduced without urtherpermission, provided the

source isacknowledged.

WATER TREATMENTMANUALS

FILTRATION

Publishedbythe Environmental Protection Agency, Ireland.

The Agency personnel involved in the production and preparation of this manual

were Mr. Noel Bourke,Mr. Gerry Carty, Dr. MattCrowe and Ms. MarionLambert

(word processing).

IFC.



WATER TREATMENTMANUALS

FILTRATION

Environmental ProtectionAgencyArdcavan, Wexford, Ireland.

Telephone:+353-53-47120 Fax : +353-53-47119

Co 3S—5'3



CONTENTS 1

CONTENTS

CONTENTS 1

LISTOFFIGURES 4

LIST OF TABLES 4

PREFACE 5

ACKNOWLEDGEMENTS 6

LIST OFABBREVIATIONS 7

1 INTRODUCTION 9

1.1 PROCESS DESCRIPTION 9

2 FILTRATIONMECHAN1MS 11

N3 SLOWSANDFILTRATION 13

3.1 SLOWSAND FILTER MEDIA 13

3.2 FILTER LAYOUTANDOPERATION 13

3.3 SLOWSAND FILTER CONTROL 15

3.4 SLOWSAND FILTER CLEANING ANDRESANDING 15

3.4.1 WORKER SAFETY AND HYGIENE PRACTICE 15

3.4.2RECOMMENDEDPROCEDURE OR CLEANING ASLOW SAND FILTER 16

3.4.3 RECOMMENDEDPROCEDUREFOR RESANDINGASLOW SANDFILTER 16

4 RAPIDGRAVITY ANDPRESSURE FILTRATION 19

4.1 RAPIDGRAVITY FILTRATION 19

4.2 PRESSURE FILTRATION 19

5. PROCESS CONSIDERATIONS 23

5.1 RAPID GRAVITY FILTER MEDIA 23

5.2OPERATIONAL CRITERIA 24

5.2.1FILTER LAYOUT 24

5.2.2 FILTER PRODUCTIONAND FILTRATION RATE 24

5.2.3FILTRATION EFFICIENCY 24



2 FILTRATION

5.3FILTER OPERATION 25

5.3.1 RAPID GRAVITY FILTRATION 25

5.3.2BACKWASHING25

5.4 FILTER CONTROL SYSTEMS 26

6 ACTIVATED CARBONFILTERS 29

7 INTERACTION WITH OTHER TREATMENT PROCESSES 31

7.1 PREFILTRATION TREATMENT 31

7.2 IN-LINE FILTRATION 31

8 PROCESS MONITORING ANDCONTROL 35

9 OPERATING PROCEDURESASSOCIATED WITHNORMAL PROCESS CONDITIONS 37

9.1 INDICATORS OFNORMAL OPERATING CONDITIONS 37

9.2 PROCESS ACTIONS 37

9.3 PROCESS CALCULATIONS 40

9.3.1 FILTRATION RATE 40

9.4 RECORDKEEPINGANDQUALITY CONTROL40

9.5 FILTER MONITORING INSTRUMENTATION 41

10 OPERATING PROCEDURES ASSOCIATED WITH ABNORMALPROCESS CONDITIONS 43

10.1 INDICATORS OF ABNORMAL CONDITIONS 43

10.2 PROCESS ACTIONS 43

10.3AIRBINDING 44

10.4 EXCESSIVE HEAD LOSS44

11 STARTUP ANDSHUTDOWN PROCEDURES 47

11.1 CONDITIONS REQUIRING IMPLEMENTATIONOFSTARTUP ANDSHUTDOWN PROCEDURES 47

11.2 IMPLEMENTATION OF STARTUP/SHUTDOWN PROCEDURES 47

11.3 FILTER CHECKING PROCEDURES 47

11.4 BACKWASH PROCEDURES 47

11.5 FILTER STARTUP PROCEDURES48

11.6 FILTER SHUTDOWN PROCEDURES 48

12 PROCESS ANDSUPPORT EQUIPMENT OPERATION ANDMAINTENANCE



CONTENTS 3

12.1TYPESOFEQUIPMENT 51

12.2EQUIPMENT OPERATION 51

12.3PREVENTIVE MAINTENANCE PROCEDURES 51

12.4 SAFETY CONSIDERATIONS 52

GLOSSARY 53

APPENDIX A: TYPICAL FILTERS OPERATING RECORD SHEET 57

APPENDIX B: CONVERTING TO SI (METRIC) UNITS 62

APPENDIX C: MUDBALLEVALUATION PROCEDURE 65

APPENDIX D: HYDRAULIC CALCULATIONS 67

REFERENCESAND READING LIST 71

USER COMMENT FORM 73

RECENT ENVIRONMENTAL PROTECTIONAGENCYPUBLICATIONS 75



4 FILTRATION

List of Figures

FIGURE]:TYPICAL WATER TREATME)VTPROCESSES 10

FIGURE 2:COMBINATION OF TWO TYPES OF STOCKSAND 14

FIGURE 3: SLOWSANDFILTERAND PREFILTRATIONCONTROL CHAMBER 15

FIGURE 4: RESANDINGA SLOW SAND FILTER USINGTHE TRENCHINGMETHOD 18

FIGURE 5: CONVENTIONALANDDIRECTFILTRATION 20

FIGURE 6: RAPIDGRAvifYFiLTER 20

FIGURE 7: GRAVifY FiLTERMEDiACONFIGURATIONS 21

FIGURE 8:PRESSURE FILTER 22

FIGURE 9:MEASUREMENTOFHEADLOSS 39

FIGURE 10: TYPICAL O[TTFLOWTURBIDifY DATA 39

FIGURE 11: FILTRATION PROCESSMONiTORING -PARAMETERS ND LOCATiONS 41

FIGURE 12: FILTRATIONMODE OF OPERATION 50

FIGURE 13:BACKWASHMODE OFOPERATION 50

FIGURE 14:MUDBALL SAMPLER 66

List of Tables

TABLE 1: TYPICAL MEDIA FILTERCHARACTERISTICS 24

TABLE2:SUMMARYOFROUTiNE ILTRATION PROCESS ACTIONS 32

TABLE3:FILTRATIONPROCESSTROUBLESHOOTING 33

TABLE 4:READINGATURBIDITYROD 37

TABLE 5: PROCEDUREFOR BACKwASHINGA FILTERUNDER MANUALCONTROL 49

TABLE6: POTENTIALHAZARDS NFILTRATION 52



PREFACE 5

PREFACE

The EnvironmentalProtectionAgencywasestablished in 1993 to licence, regulate and controlactivities forthe purposesof environmental protection. In the Environmental ProtectionAgency Act, 1992, it is statedthat "theAgency may, andshall ifso directedbytheMinister, specify andpublishcriteria andprocedures,which in the opinion of the Agency are reasonable and desirablefor the purposes of environmentalprotection". Thesecriteriaandprocedures n respectofwater reatment arebeingpublished by the Agencyinanumberofmanuals under hegeneral headingofWaterTreatmentManuals.

This manual on Filtrationsets out thegeneralprinciples and practices which shouldbe followedby thoseinvolved in the production of drinking water. Next year, the Agency intends to prepare and publishadditional manuals on Disinfection, Coagulation, Flocculation, Clarification and Fluoridation. Wherecriteria and procedures are published by theAgency,a sanitary authorityshall, in the performance of itsfunctions, haveregardtosuch criteriaandprocedures.

This manual includes nformation on many aspects of the filtration process. The Agencyhopes that it willprovide practical guidance to those involved in plant operation, use, management, maintenance andsupervision. TheAgency welcomes anysuggestions which usersof the manualwishtomake. Theseshouldbe returnedto theEnvironmental Management and PlanningDivision at the Agency headquarters on theenclosedUserComment Form. -



6 FILTRATION

ACKNOWLEDGEMENTS

The Agencywishes oacknowledge thosewhocontributed toand reviewed this manual. A review panelwas

established by the Agency to assist in the finalisation of the manual and we acknowledge below those

persons who took the time to offer valuable information, advice and in many cases comments and

constructive criticism on the draft manual. We gratefully acknowledge the assistance offered by the

following persons:

John Anderson, Department of he Environment

MartinBeirne, Environmental HealthOfficers Association

AideenBurke, Jones Environmental (Ireland) Ltd.

ProfessorTomCasey, University College,Dublin.

Ned Eeming,DublinCorporation

TomLoftus,DublinCorporation

Eamon Mansfield, Waterford County Council

DavidMcBratney,M.C. O'Sullivan& Co. Ltd.

ColumMcGaughey, Dublin Corporation

JohnO'Flynn,Waterford County Council(representingtheCounty and CityEngineers Association)

M.C.O'Sullivan,M.C. O'Sullivan& Co.Ltd.

Paul Ridge, Galway County Council

John Walsh, E.G. Pettit&CompanyLtd.

The Agency also wishes to acknowledge the assistance of the Sanitary Services sub-committee of the

Regional Laboratory, Kilkenny, whocommented onthedraftmanual.



LIST OF ABBREVIATIONS

• EBCT: Empty Bed ContactTime

ES: Effective Size

GAC: GranularActivated Carbon

LOH: LossofHead

rn/h: m3 perm2 perhour

mm: millimetres

NTU: Nephelometric TurbidityUnit

p.s.i.: pounds persquare inch

THMs: Trihalomethanes

TON: Threshold OdourNumber

TWL: TopWaterLevel

UC: Uniformity Coefficient

UFRV: UnitFilter RunVolume

VOC: Volatile Organic Carbon

LISTOFABBREVIATIONS 7



8 FILTRATION



1. INTRODUCTION

1.1 PROCESS DESCRIPTION

Filtration is theprocess ofpassingwater throughmaterial to remove particulate and other

impurities, including floc, from the water beingtreated. These impurities consist of suspendedparticles (fine silts and clays), biological matter

(bacteria, plankton, spores, cystsor othermatter)and floc. The material used in filters for publicwater supply is normallya bed of sand, coal,orothergranular ubstance. Filtration processes can

generally be classified as being either slow or

rapid.

Slow sand filters are the original form offiltration. Thefirst one was builtin 1804 byJohnGibb of Paisley, Scotland to treat water for his

bleachery, with the surplus treated water sold to

thepublic.2 Slow sand filters were first used inLondon in 1820 to treat water from the River

Thames. From about the 1930s water treatment

by coagulationand rapid gravity filtration or

pressure filtration tended to replace slow sand

filtrationin new plantsand, in some cases, slow

sand filters were replaced by rapid gravity iltersfollowing ntroduction ofacoagulation stage.The

slow sand filtration process has come back intofavour in recentyearsdueto its superior ability,

compared to rapid gravity filtration, to remove

pathogenic micro-organisms such as Giardialamblia andCryptosporidium.

Typical water treatment processes are shown

schematically in Figure i"4, with the relative

position of iltraticfn llustrated.

-4 9

I INTRODUCTION



TREATEDWATER

EFFECTON WATER

Excludes fishand emus es leases, sticks and other

largedebris.

Breaks down colloidal stability. Adjusts pH for optimum

coagulation.

Mixes chemicalso th raw water,containing ine particles thato ll not readilysettleor filter oat of he water.

Gathers together ine, lightparticles to form arger clumps flocto aid the sedimentationiflotation nd iltrationprrcesses.

Sedimentation settlesour large suspended particles.Plotation floats oat the particles with dissolvedair.

Rapidgravits filtration iltersor remoses

remaining uspended particles.

Slow sand filtrarioralso involvesbiologicalaction.

Kills / nactisares disease-causing organisms. Pros des chlorine

residual fordistribution system. where chlorine s used.

Helpscontrol corrosivepropertiesof water.

Helpscontrol dental caries n children andyoungadults.

FIGURE1: TYPICAL WATERTREATMENT PROCESSES

10 FILTRATION

RAW TREATMENT

0

0

z

DISINFECTION

]_STERILIZATION

pH CORRECTION

Stores waterprior o discharge to service esers ems.



2 FILTRATION MECHANISMS

2 FILTRATIONMECHANISMS

Filtration is essentiallya physical and chemical The relative importance of these removal

process and, in the case of slow sand filtration, mechanisms will depend largelyon thenature ofbiological as well. The actual removal the water being treated, choice of filtrationmechanisms are interrelated and rather complex, system, degree of pre-treatment, and filterbut removal of colour and turbidity is based on characteristics.thefollowing factors:

• chemical characteristics of the water beingtreated(particularly source water quality);

• nature of suspension (physical and chemical

characteristics ofparticulates suspended in thewater);

• typeanddegreeofpre-treatment (coagulation,flocculation, andclarification); and

• filtertypeandoperation.

A popular misconception is that particles areremoved in the filtration process mainly byphysical straining. Straining is a term used to

describe the removal of particles from a liquid(water) by passing the liquid through a filter orfabric sieve whose pores are smaller than the

particles to be removed. While the strainingmechanism doesplayarolein theoverall emoval

process, especially in the removal of largeparticles, it is important to realize that mostof the

particles removed during filtration are

considerably smaller than thepore spaces in the

media. This is particularly true at thebeginningof the filtrationcycle when the pore spaces areclean (that is, not clogged by particulatesremovedduring filtration).

Thus, a number of interrelated removalmechanisms within the filter media itself arerelied uponto achieve high removal efficiencies.Theseremoval mechanisms include the following

processes:

• sedimentation on media(sieveeffect);

• adsorption;

• absorption;

• biological action; and

• straining.



12 FILTRATION



31 SLOW SAND FILTRATION

3 SLOW SAND FILTRATION

3.1 SLOW SAND FILTERMEDIA

Allen Hazen was one of the pioneers in the

scientific investigation ofslowsand filtration and

introduced, as long ago as 1892, the concept of

effective size (ordiameter).

The EFFECTIVE SIZE (ES) is defined as thesize ofa sieve opening through which 10percent

(by weight)of the particles (sand)will just passand is given thesymbol d10. Inasimilarway,the

size ofasieve opening through which60percent

(by weight)of the particles (sand) will just passis given the symbol d60. The UNIFORMITY

COEFFICIENT, (UC) which is ameasureof the

gradingof the material, is the ratiod d10

It is normal to employ ungraded sand as

excavated from natural deposits in slow sand

filters. The sand should have a uniformitycoefficient of less than 3, but there is little

advantage in having a U.C. of less than 1.5 ifadditional cost is thereby incurred. Pit-run sand

maybe washed to remove admixedclay , loam ororganic matter and this will remove the finest

grains thus lowering the U.C. and raising the

average particlediameter.

Ideally the effective size should be just small

enough to ensure a good quality outflow and

prevent penetration of clogging matter to such

depth that it cannot be removed by surface

scraping. This is usually in the range 0.15 - 0.35

mm and is determined byexperiment. Both finer

and coarser materials have been found to work

satisfactorily in practice, and the final selectionwillusually depend on the available materials. It

is possible tocombine wo ormoretypesofstocksand to bring the effective size of the mixture

closer to the ideal, as shown in Figure 22. The

broken line in the Figure indicates the filteringmaterial obtainedby mixing 3 parts of sand A

with one part of sand B to give an effective

diameterofabout 0.25mm.

Desirable characteristics forall filter media are as

follows:

goodhydraulic characteristics (permeable);

• does not react with substances in the water(inertandeasy toclean);

hard anddurable;

• freeof mpurities; and

• insoluble inwater.

Gravel is used to support the filter sand and

shouldhave similarcharacteristics.

3.2 FILTER LAYOUTAND OPERATION

The ilter media is usually containedin concrete

filter tanks which are all the same size, but the

size will vary widely from works to works. In

general, the minimumnumberof iltersis three o

allow for one filter to be out of service and yethave sufficient capacity available to meet the

average demand. Atypicalsectionofaslowsand

filter and prefiltration control chamberis shown

inFigure3.

Water, admitted to a slow sand filter, properly"conditioned", flows downward through the

media. In the slow sand filtration process,

particles are removed primarily by straining,

adsorption, and microbiological action. Themajor

partof heparticulate material isremoved ator inthe top layer of sand, which becomes covered

with a thin slimy layer, called the

"schmutzdecke", built up by micro-organisms as

the filter 'ripens'.

Slow sand filtration does not materially reducethe true colour of a water. True colour is the

colour attributable to light-absorbing organicmolecules which are dissolved in the water as

distinct from that attributable to light-scatteringcolloidalparticles such as clay. Truecolour 5,13 is

measured ollowing the removal ofsuchcolloidal

matter by passing a sample of water through a

membrane filter of aperture approximately 0.5

tm. With a standard of 20 mg/i platinum/cobaltfor colour as the maximum allowable

concentration in Ireland, slow sand filtration is

unlikely to be suitable forraw waterwith colourinexcessofabout25 mg/i platinum/cobalt, unless

colour soxidised in the filtratebyozonisation.



C)

C

C

C)

>

=C)

C)

14 FILTRATION

99

90

80

70

60

50

40

30

20

0.!

Grain size,d, mm

0.2 0.3 0.4 0.5 0.7 I .0

FIGURE2:C0MBINATI0N F TwoTYPES OF STOCK SAND



SLOW SAND FILTRATION

Filtration rates areextremely low, typically 0.1 to0.2 mperhour,depending on whether owland orupland water is being filtered and the treatment

regimeapplied. The higherfiltration ratesapplyto upland waters, with which the classicschmutzdecke does not form, or to lowlandwaters which are pretreated by microstraining orroughing filters, as in the case of the Belfast

supply rom CastorBay, Lough Neagh, where thewater is first treated by rapid gravity filtration.Chemicals such as chlorine arenot addedbeforeslow sand filtration as this would inhibit

microbiological activity.

The entire top layer of the filter, including theschmutzdecke, must be physically removed whenthe filter becomes clogged, or else the filtercleaned in-situ. No works within-situmechanical

cleaning exists in Ireland as far as is known.

Disadvantages toslowsand

filtration are the argeland area required and the manpower, orinvestment in plant, required for cleaning thefilters.

3.3 SLOWSANDFILTER CONTROL

Slow sand ilters are normally supplied throughan inlet chamberhousinga control valve and ameasuring weir (see Figure 3). A single inletchamberusually feedsa groupof filtersor even

an entire small works and controls the level ofwater in the filters. The head on the filters insmaller works is normally controlled by anadjustable bellmouth in the outlet chamber ofeach filter,sometimes fitted withascale o enablethe flow over the lip to be gauged. Whenafilterhas been cleaned and put into service, the

headloss through it is only a few cm and thebellmouth is set near the top of its travel. Theheadloss increases as the filter ages and thebellmouth is lowered to maintain thedesired flow

through the filter. When the belimouth hasreached the bottom of its travel,the filter hasreached theendof tsrun and needs to be cleaned

again.The

adjustableoutlet bellmouth

ensuresthat negative pressures cannot occur within thebed. Alternative forms of adjustable outlets tothat describedaboveareavailable. Controlof hefiltration rate on larger works may be achieved

automatically by measuring the outflow fromeach filter usinga flow rate controllerand usingthis signal to raise and lower the bellmouth orotherdevice.

3.4 SLOWSANDFILTER CLEANING ANDRESANDING

3.4.1 WORKER SAFETY AND HYGIENEPRACTICE

During filter cleaning orresanding operations allworkers should wear rubber boots, which aredisinfected in a tray of bleaching solution before

entering the filters. The highest standards ofpersonal hygiene should be observed by allworkers and no worker with symptoms that might

be attributable to waterborne or parasitic diseaseshouldbepermitted tocome nto director indirectcontact with the filtermedium.

FIGuI3: SLOWSANDFILTERAND PREFILTRATIONCONTROL CHAMBER



16 FILTRATION

3.4.2 RECOMMENDED PROCEDURE FOR

CLEANING A SLOW SAND FILTER

1. Draining he filter andpreparing tor

cleaning

Whena filter has reached the end of its run and

arrangements for itscleaning have beenmade, the

raw water inlet valve is closed, to initiate the

cleaning operation, thus allowing the filter to

continue to discharge at a reducing rate to the

clear water tankfor as long as possible (usually

overnight).

The outflow valve is then closed (e.g. next

morning) and the remaining water abovethe sandisrun to waste.

The water withinthe bed is then lowered about

100mmbelow the surface of the bed by openingthe drainvalve.

2. Cleaning the filter

Cleaning should start as soon as the

schmutzdecke is dry enough to handle. If the

filter is left too long, it is likely to attractscavenging birds that will not only pollute the

filter surface but disturb the sand to a greater

depth than will be removed by scraping. If, as is

normal, mechanised methods are not available,

workers using squarebladed shovels should stripoff the schmutzdecke and the surface sand

adhering to it and stack it in ridgesor heaps for

removal.

Great care should be taken to minimise

disturbance of the upperlayersof filter media so

that the biomass is protected. For this reason,dumpers or other machines used for removal

shouldbe designedto operate with low pressure

tyres and barrows or handcarts should always be

run on protective planks.

Cleaning is a simple matter when the

schmutzdecke consists largely of filamentous

algaeformingan interwovenmat. The knackof

curlingback this mat in reasonably largesections

is quickly acquired, provided that the operation is

timed so that the material is neitherwaterlogged

nor so driedout that it isbrittle.

Cleaningwill be

less easy if the schmutzdecke consists largelyof

non-filamentous algae and greater care will be

necessary to control the depth ofscraping, which

shouldbe between 15 and 30mm.

Afterremoval of the scrapings the bed should be

smoothed to a level surface. The quicker the

filter-bed is cleaned the less will be the

disturbance to the biomass and consequently the

shorter the period of re-ripening.The micro-

organisms immediately below the surface will

quickly recover, provided they have not been

completely dried out, and will adjust to their

position relative to the new bed level. A day or

two should be sufficient for re-ripening in this

event.

Before refilling the filter, hewallsbelow normal

top water level should be swabbed down to

discourage the growth of slimesandalgae.

3. Refilling the filter

The filter is refilled by closing the drain valve,

reopening the outflow valve and allowing the

filtered waterto backflow through the underdrain

system until thereis sufficient depth on the sand

surface to prevent disturbance of the bed by

inflowing raw water. When the filter is

sufficiently chargedthe outflow valve is closed.

The raw water inlet valve is then gradually

opened and the filter is filled to normal operatinglevel.

The bellmouth ofthe telescopic outletis raisedto

near its maximum and the filter outflow is run to

wasteata gradually increasing rate, (or if there is

a suitable pump, recycled to theworks inlet), for

a day or two until the filter is re-ripened and

analysis shows that the outflow satisfies the

requiredquality tandards.

4. Disposalofscrapings

The materialremoved rom the filter, dependingon the size and available equipment at the works.

may be washed for reuse or disposed of on land

by burial or used in agriculture. The workforce

whoscrape the filters or wash the sand should be

instructed innecessary hygiene practice.

3.4.3 RECOMMENDEDPROCEDURE FOR

RESANDINGA SLOW SAND FILTER

Each cleaning of the filter removes between 10

and 15 mm of filter sand, so that after twenty orthirty scrapings the thicknessof the sand bedwill

have been reduced to its minimum designthickness, usually about 300mm and resanding is

then necessary. The following procedure is



SLOW SAND FILTRATION

17

recommended for re-sanding a slow sand filter.The filter can either be re-sanded by the'trenching' method which makes use of theresidual

sand,or

by refilling with newsand.

Generally, re-ripening of the filter is quicker ifthe renching method isemployed.

1. Preparation

The bed is cleaned, as described n section3.4.2above, and the water level lowered to thebottomofthe sand layer.

2a.Resanding - he 'trenching' method

Mostof theresidualsand is removed from a stripofthe filter along one wall thus forming a trench,taking care not to disturb the underlying gravellayer by leaving 100 to 150 mm of residual sandaboveit. The sandis placed adjacent to the filterfor laterre-use.

Fresh sand (either new or washed sand fromfilter

cleaning) is placed in the trench to a thicknesswhich, withtheresidual sand, equals thedepthof

sand in thefilterpriortore-sanding seeFigure 4)or as determined in consultation with thedesigner.

Residual sand from theadjacent strip is "thrownover" on topofthe freshlyplaced sand inthe first

strip, fresh sand is placed in the second trench

justformed andsand from the nextadjoining stripis"thrownover" as shown inFigure 4.

Thisprocess is continued untiltheentirefilter hasbeenresandedand theresidualsand fromthefirststrip is placed on top of the last strip filled. Inthis way, the residualsand forms the upper layerof he re-sanded filter and thenew sand the lowerlayer.

and replaced during the resanding operations, asthe carbon will have become saturated with theimpurities and/or nactivated.

2b. Resanding - cleansandmethod

Analternative to the trenching method describedaboveis to remove all theold sand from the beddown to the support gravel and refill with cleansand.

3. Refilling he filter

The bed should be smoothed to a level surfaceand refilled with filtered water through the

underdrainage system. The raw water inlet isthenopened as described in section 3.4.2.

4. Re-ripening the filter

The ripeningof a filter with clean sand will take

longer than the time for re-ripening a filterresandedusingthe trenching method. Caremustbe taken to ensure that a satisfactory quality

outflow is being produced beforethe filteris putback into service.

The filter-bed mayhavea layerofcrushed shells

incorporated near the bottom, to correct the pH,where theraw water is naturallyaggressive. Theresidue of the shell layer must normally beremoved and replaced during re-sandingoperations. Occasionally, a layer of activated

carbon about 100mm deep is placed near thebottom of the filter-bedto absorb traces of tasteand odour-producing substances that havepassedthrough hefilters. This layer shouldbe removed



18 FILTRATION

FIGURE4: REASANDING ASLOWSANDFILTERUSINGTHE TRENCHINGMETHOD

level th ifiter



4RAPIDGRAVITYANDPRESSURE FILTRATION

4 RAPID GRAVITY AND PRESSURE FILTRATION

Filtrationpreceded by coagulation, flocculationand clarification is commonly referred to asconventional filtration. In the direct filtration

process, although coagulation and flocculation

may be used, the clarification step is omitted.Typical treatment processes for these twofiltration methods are shown in Figure 5. Theconventional filtration (treatment) process is usedin mostmunicipal reatment plants. Thisprocessincludes "complete" pretreatment (coagulation,

flocculation,and

clarification/flotation). Thissystemprovidesflexibility and reliability in plantoperation, especially when source waterqualityisvariable or the water is high in colour and

suspended solids.

Directfiltrationis an alternative to conventionalfiltration, particularly whensource watersare owin turbidity, colour, plankton, and coliform

organisms. Direct filtration can be definedas atreatment system in which filtration is notprecededbyclarification orflotation.

4.1 RAPIDGRAVITYFILTRATION

In all gravity filtration systems thewater evel or

pressure (head) abovethemediaforces the water

through the filtermediaasshown inFigure6.Therate at which water passes through the granularfiltermedia(the filtration rate) may vary widely,dependingon the purposefor which the water isrequired. However, for public water supply in

Ireland, 5 m/hourmaybe regarded as the standard

rate and most authorities limit the maximumfiltration ate to between5 and7.5 rn/hour. Ratesin excess of this may be used where specialtechnology is employed. The rate of water flow

through the filter is referredto as the hydraulicloading or the filtration rate. The filtration rate

depends on the raw water quality and the type offilter media. Various filter media configurationsused are illustrated nFigure7.

They are:

• multi or mixed media (sand, anthracite,garnet).

Activatedcarbon (in granular form)can also beused in association with theseconfigurations forthe removal of tastes, odours, and organicsubstances.

In rapid gravity filtration the particulateimpurities are removed in or on the media, thus

causingthe filterto clogafter aperiod. Cloggedfilters are cleaned by backwashing. Gravityfiltration is the most widely used form of watertreatment nthiscountry.

4.2 PRESSURE FILTRATION

Apressurefilteris similarto a gravity sandfilterexcept that the filter is completely enclosedin apressure vessel such as a steel tank, and isoperatedunder pressure, as shown in Figure 8.Pressure filters have been used in public water

supplies, and have limited applicability, forinstance, in the removal of iron and manganesefromgroundwaters. Their use in Irish practice smainly confined to the treatment of water forindustrial purposes, but they have been installed

byanumberof local authorities.

Pressure filters have been found to offer lowerinstallationand operation costsin small filtration

plants. However, they are generally somewhatless

reliable than gravity filters. Maximumfiltration rates forpressurefilters are in the5 to7.5 rn/hrrange.

• singlemedia(sand);

• dual media(sandandanthracite); and



c..—t,

Fl1t,Uou

FIGURE5: CONVENTIONALAND DIRECTFILTRATION

FIGURE6:RAPID GRAvITY FILTER

20 FILTRATION

PtocoitionAid

Co.gd.LloAFioccul.tio Clrlflc.to Filton

Coiiveaboia1Filtratloil

IAiiorI—

Direct Filtretlon



(14 Mono Media

toflow

TWLV7

RAPIDGRAVITYAND PRESSURE FILTRATION

coal

FIGuRE 7: GRAVITY FILTERMEDIACONFIGURATIONS

1

FUteeOutflow

Inflow

Inflow

IWL T

coal

sand

I. grav&

1500 o 2000 m

800mm

t11terOutflow

Underdnino

IC)Multi-Media

brmixed)

I sand

1500 o2000mm

tt800mm



22 FILTRATION

FIGURE8: PRESSUREFILTER



5. PROCESS CONSIDERATIONS

5 PROCESS CONSIDERATIONS

23

5.1 RAPIDGRAVITY FILTER MEDIA

The most common filtering material in rapidgravity (and pressure) filters is sand3. The dualmedia filter is a refinement of the rapid gravitymonosand filter, in which an upper layer ofanthracite or similar material above the sand

provides ncreased void space to store impuritiesremoved from the incoming water. The mixed

media filter is a further refinement, in which a

layerof garnetorsimilardensematerial isplacedbelow the sand. Other nert materials mayalso be

used including activated carbon. Gravel iscommonly used to support the filter media.

Desirable haracteristics for all filter mediaareasfollows

• goodhydraulic characteristics (permeable);

• does not react with substances in the water

(inertandeasytoclean);

. hardanddurable;

• freeof impurities; and

• insoluble in water.

Gravel is used to support the filter sand and

shouldalso have he abovecharacteristics.

Ithas been thecustom to designate sandforrapid

gravity filters in terms of its effective size and

uniformity coefficient. Othermaterials for rapid

gravity filters are referenced by means of the

supplierscode reference

(e.g. No 2Anthracite

)5A U.K. standard for the specification, approvalandtestingofmaterials for rapid gravity filtrationhas recently been published by the British

Effluentand WaterAssociation6. Thestandard isintended to cover all granular materials used in

rapid gravity iltration, in terms of theparameters

necessary for the successful operation of thefiltration system including backwash. Effectivesize and uniformity coefficient are omitted and anew parameter, hydraulic size, is defined. This

parameter can be used in calculating the

fluidization hreshold, the point in backwashingwhere the hydraulic (pressure) loss through thefilter sandequalsthe dead(submerged) weightofthe material. Other important parameters in this

calculation are thedensity of thematerial and its

voidage, or porosity as it is more commonly

known. The porosity is dependant on theshapeof the grains. Rounded grains tend to have alower porosity and can be washed at a lowerbackwash rate; however, as a filtering material ithas less space within it to hold the flocs andsolids removed fromthewater and alsoproducesa higher headloss. Given the correct backwashrate in the first place, material with higher

porosity has much to commend it. Anthracite,whichis acrushed material, is agoodexample ofahighporosity medium which has been inuse for

many years.

Two factors are very important in making

judgementsabout media selection:

• the time required for turbidity to break

through hefilterbed;and

• the time required for the filter to reach

limitinghead oss.

With a properly selected media, these times are

aboutthesame.

If the limitinghead loss is frequently a problemand turbiditybreakthrough rarely occurs, then a

largermedia sizemaybe considered. If turbidity

breakthrough is frequentlyaproblemand limitinghead loss is rarely encountered, then a smallermediasizemaybeconsidered.

Ifboth head loss andturbiditybreakthrough area

problem, while the filter is operating within itsrated capacity, a deeper filter bed with a largersand size maybe required. The optimumdepthneededtoobtain agivenqualityand length ofrunvaries with the size of the sand. However,

increasing the media depth is notalwayspossiblewithout modification of the filter. Adequateclearance mustbe allowedbetween he topof themedia and theweir cill ofthe washwater channel.

Otherwise, filter media will be carried over into

the washwater channel during backwash, whenthebedexpands.

The relationship between turbiditybreakthroughand

limitinghead ossis also

stronglyaffected

bythe efficiency of chemical pre-treatment. Poorchemical pre-treatment can often result in earlyturbiditybreakthrough, rapidhead lossbuildupor

bindingof he filtersand.



24 FILTRATION

It must also be remembered that backwashing is

possibly a more critical part of the filtration

processthan the forward filtration phase. If thebackwash rate is

slightlytoo

high.filter material

can be lost rapidly. If it is slightly too low, the

cleaning efficiency will fall abruptly. the filterwill cease to be cleaned properly and its

performance will quickly deteriorate. Thebackwash sequence must be selected takingaccount of the material used and the temperatureof the backwash water. Existing backwash

procedures will require adjustment if a new

filtering material has a differentsize, density orvoidsratio.

Selection of anappropriate

media forrapidgravity filtration depends on the source water

quality, filter design and anticipated filtrationrate. Generally the more uniform the media theslower the head loss buildup. Media with

uniformity coefficients ofless than 1.5 are readilyavailable. Mediawith uniformity coefficients ofless than 1.3 are only available at a high cost.

Typical filter media characteristics are given in

Table1.

Table 1: Typical Media FilterCharacteristics

Material SizeRange(mm)

Specific

GravityConventional 0.5-0.6 2.6

5.2.1 FILTER LAYOUT

In rapid gravity filtration, the filter media isusually contained in a numberof concrete filter

tanks(or cells)which arc all the same size. Steel

or aluminium tanks are sometimes used.

However, the size will varywidely fromplant to

plant. In general, the minimum numberoffilters

is three. Thisal!oss for one filter to be taken outof service leaving sufficient filter capacityaailable to meet the demand. Typical filtersections are shown in

Figure7.

5.2.2 FILTER PRODUCTION AND

JLTRATJONPATE

For public watersupply in this country. thestandard filtration rate for rapid gravity filters is

usually 5 rn/hour and most authorities limit the

maximum filtration rate to between 5 and 7.5rn/hour. The minimum number of filters should

be three for the smallest works as noted above.but the number and size should be

keptreasonable in relation to the total works

throughput with the size based on a 20 hour

working day. In larger water treatment plants.filter capacities range up to about 25.000 m3/dayfor the largest filter units. Filtration plants

operatedat rates in excess of thosequoted above

generally require additional supervision.

5.2.3 FILTRATION EFFICIENCY

Rapidgravity filtration efficiency is roughly

measured by overall plant reduction in turbidity.although it should be noted that up to 90% of thereduction may take place in the pretreatment

stages. Overall reductions of over 99.5 percentcan be achievedunderoptimum conditions, while

a poorly operated filter and inadequate pre-treatment (coagulation. flocculation, and

clarification) can result in turbidity removals ofless than 50 percent. The bestwayto assure highfiltration efficiency is to select an outflow

turbidity target and stay below the target value

[such as 0.5 NTU (Nephelometric TurbidityUnits)1.

Solids removal efficiency depends largelyon: the

quality of the water being treated: theeffectiveness of the pre-treatment processes in

conditioning the suspended particles for removal

by clarification and filtration: and, filter

operation.

Filter unit design and filter media typeand depthalso play a role in determining solids removal

efficiency. but are less important than water

quality and pre-treatment considerations. Rapidgravity sand filters usually produce a filtered

water urbidity comparable to thatofadual-media

filter if the applied water quality is similar.

However, the operational differences betweensand and dual-media filters are significant.

Sand

CoarseSand 0.7-3.0 2.6

Anthracite/Coal 1.0-3.0 1.5-1.8

Gravel 1.0-50 2.6

5.2 OPERATIONAL CRITERIA



r 25

0 PROCESS CONSIDERATIONS

Because of their smaller media grain size,

typically 0.8 to 1.0 mm,sand filters tendto clogwithsuspended matterand floc morequickly than

dual-media filters. This means hatmore frequent

backwashing will be required to keep the sandfilteroperating efficiently. Sand ilters have fine,

light grains on the top which retain all floc and

particulates at or near the surface of the filter.

Dual-media filters have lighter, larger diameter

grainsin the top layerof the media which retain

the larger particles; the smaller particles are

usually retained deeperin the filter. The larger

grainsizeof theanthracite layer (upto 1.5 mm) n

the top portion of a dual-media filter permits

greater depth penetration of solids into the

anthracite

layerand

largersolidsstorage volume

in the filter. The sand layerbelow the anthracite

is used as a protective barrier against

breakthrough. These characteristics generally

producefilter runs which are longer than those

achievedbysand filters. Tapsatvarious depths in

the filter may be used to observe the depth ofsolids penetration.

Multi-media filters (sand, activated carbon and

anthracite) are alsoused to extend filter run times,and these filters generally perform in a manner

similar o sandfilters,except that the filtermediais enclosed ina pressure vessel. With the filter

media fullyenclosed, t is impossible to assess themedia condition by simple visualobservation. Inaddition, excessive pressure in the vessel will

force solids as well as water through the filtermedia. Obviously, this will result in the

deterioration of filtered water quality. Pressureshould therefore be maintained withinthe range

provided orbythedesigner.

Particular care should be taken when a filter is

being backwashed to ensure that the remainingfilters are notoverloaded. Backwashing, togetherwith routine and emergency maintenance, should

be completed withoutoverloading the remainingfilters.

5.3 FILTEROPERATION

5.3.1 RAPID GRAVITY FILTRATION

In the filtration mode of operation, watercontaining suspended solids is applied to the

surface of the filter media. Depending on theamount of suspended solids in the water beingtreated and the filtration rate, the filter will

exhibit head loss and "clog" after a given time

period (varies fromseveral hours toseveral days).

Clogging may be defined as a buildupof head

loss (pressure drop)across the filter media until itreaches some predetermined design limit. Total

design head loss ingravity filters generally rangesfrom about 1.8 to 3.0 m depending on the depthof the water over the media. Cloggingof the

filter leads to breakthrough, a condition in which

solids are no longer removed bythe already over-

loaded filter. The solids pass into the filtered

waterwhere hey appearasincreased turbidity.

A filter is usually operated until just before

clogging or breakthrough occurs, or a specifiedtimeperiodhas passed, generally24 to 40 hours

andis related to the efficiency of the clarification

process. In order to save money, energy and

water by maximizing production before

backwashing, filters are sometimes run until

clogging orbreakthrough occurs. This is a poor

practice and should be discouraged. When

breakthrough occurs, therewill be an increase in

filtered water turbidity and hence a decrease in

water quality with the risk that pathogenic

organismsmay pass through thefilter.

5.3.2 BACKWASHING

Afterafilterclogs (reaches maximum head loss),or break-through occurs, or a specified time

period haspassed, the filtration process is stoppedand the filter is taken out ofservicefor cleaningor backwashing. Detailed procedures for

backwashing tend to be particular o each plantand specific instructions are provided by plantcontractors.

Backwashing is theprocess of reversing the flow

of water through the filter media to remove theentrapped solids. Backwashing may comprisethe application of water alone, air and water

seperately and sequentially, or air and water

simultaneously. The latterprocedure is generally

acknowledged tobe the most efficient, but filters

must, in general, be designed for an air/waterbackwash as changing existingfilters to use this

systemis fraughtwith difficulty. With all three

typesofwash, the process conditions are related

to the (minimum) fluidization threshold as areference point,even withcombined air andwater

washes. This threshold is the point where thehydraulic (pressure) loss through the filter bed

equals the dead (submerged) weight of thematerial. The maximum backwash water flowrate shouldnot exceed 20 rn/hr. as higher flow

rates will result in excessive media loss. Duringthebackwash cycle the bed should be expanded



26 FILTRATION

bya minimum of 10% and amaximum of20% to

ensure adequate leaning.

As mentioned previously, backwashing is

possibly a more critical part of the filtration

process than the forward filtration phase. If the

backwash rate is slightly toohigh, material can be

rapidly lost. If it is slightly too low, the cleaning

efficiency will fall abruptly, the filter will cease

to be cleaned properly and its performance will

quickly deteriorate. The backwash sequence andduration must be selected taking account of the

filtering materials in use and the temperature ofthe backwash water. It may be necessary in

some cases to adjust the backwash rates

seasonally in order to maintain optimum

conditions without osingmedia.

Filtered water is almost invariably used for

backwashing and is usually supplied by a

backwash pump. Anelevated storage tank mayalso be used to store water to backwash filters.The backwashing process will use about two to

four percent of the filter output, the lower

percentages being associated with combined

air/waterwashing.

It is very important to observe the surface of the

filter during the wholebackwash operation. Anynon-uniform surface turbulence during air scour

mayindicate a problem with the air distribution

arrangements or the occurrence of "mudballs" in

themedia. Mudballs occur whenbackwashing ofafilter is inadequate. The surface 'crust', formed

by the filtered solids and sand at the top of the

filter bed becomes cemented into a compact crust.

This cracks and pieces sink into the expandedsand bed during backwashing, without beingbroken up. These pieces of crust graduallyincrease in size during the subsequent operation

of the filter. If mudballing is suspected themudball evaluation procedure detailed in

Appendix C should be used. The simplestmethod of improvement of a filter affected by

mud-balling is to agitate the expandedbed with a

"drag"or long-tined rake.

The used washwater should be settled and thesettled washwater either discharged to waste or

recycled through the treatment process.Recycling of washwater (without balancing)

directlyto the head of theworks has been found

to cause operational difficulties with chemical

dosingequipmentand control of chemical dosingplantandhence it s notrecommended.

The settled solids should be treated with the

sludge from the clarification stage. It should be

noted that recycling of washwater can result in a

concentration of viruses, cysts and other

undesirable particles, as well as polyelectrolyes,on the filters which, in turn, increases the

potential of breakthrough into the water supply.Undernormal operating conditions where the raw

water is uncontaminated this should not pose a

problem. However, it is important that plants

develop contingency plans which avoid the

recycling ofwashwater when a pollution incidentis known to have contaminated the raw water

supply. Under normal conditions, the settled

washwater should be recycled, if possible, to the

balancing tank or reservoirsupplying theworks.or mixed with the raw water, ahead of the

flashmixer.

Any polyelectrolytes used in the solids handlingphasesshouldbe suitable for use in potable water,ifsupernatant /otherliquorsare to berecycled.

5.4 FILTERCONTROL SYSTEMS

The filter control system regulates the flow rate

through the filter by maintaining an adequatehead above the media surface. This head

(submergence) forces water through a gravity

filter. The flow through afiltermust be as stableas possible and any changes in flow rate,whenever operating conditions at the plant

change. should be controlled inorder for thefilter

to yield the optimum outflow quality. The bestcontrolsystem herefore is one with simple, safe

and reliable controllers that controls filtrationwithout hunting, and includes sensors that

monitor the, largest possible water surface areasso that changes in set-point values are

representative.

Without aneffective filter control system, sudden

flow increases or surges could dislodge solids

trapped on the filter media. If these solids were

discharged, they would seriously degrade water

quality. An adequate depth of water above themedia surface is essential to ensure that the

inflow does not disturb(scour) the media. In this

way, the energy of the inflow is absorbed before

it reaches the media, thuspreventing scouring.

An essential element in the control systemfor

rapid gravity filtration is a slow start controller,which restricts the output from a filter for a

period after backwashing while the filter is

ripening'. In new works, slow start controls are

generally incorporated in the plant design, whilein older or smaller works hydraulic/mechanicalcontrols are used. The latter depend on the flow



PROCESS CONSIDERATIONS

27

ofwater into a ank, through a submerged orifice

with increasing depth of submergence, to opentheoutletvalvefromthefilter.

Rapid gravity filter control systems can beclassified into three ypes

constant rate,withacontroller;

constant rate,variable head ype; and

declining rate (orvariable flow rate).

The constant rate typewith acontroller, maybe

operatedeither on the ConstantLevel system orwith Flow Measurement. With ConstantLevel

control the inflow to the plant is distributed

equallybetween he filters,each receiving a flow

equal to the incoming flow rate divided by the

number of operating filters. Each filter is

equipped with a controller which detects the

upstream level, which it keeps constant by

adjusting the outflow controller. Because the

upstream level is kept constant, the outflow is

equal to the inflow and clogging is compensated

for until it reaches a limit, whichdepends on theavailable head. When a filter is shut down for

backwashing or maintenance, the inflow is

automatically distributed over the filters that arestillin servicewiththeexception offilters usingasurface flush of settled water. Equal distribution

of inflow is achieved simply andreliablyby staticdevices ( orifice plates,weirs, etc ). Useof this

control system also eliminates the discrepanciesbetween total filtered flow and incoming flow

thatcanoccur with control systemsbased onflow

ratemeasurements.

In the Constant Rate control with Flow

Measurement system each filter outlet has aflowmeter linked toa controller whichcomparesthemetered flow from the filter to the flow rate

setpointand adjuststhe outflow valve until theycoincide. This system has no means of

maintaining a specific water level above he filter

media, so an additional central controller isneeded. Normally the nflow rate to the filters ismeasured and the central controller

adjuststhe

individual set-point rateof the filters accordingly.If the nflow rate increases, the level upstream ofthe filters rises and thecentralcontroller adjuststhe set-point rate for the filters until the upstreamlevel stabilizes and plant inflow and outfloware

in balance. The central controller may

alternatively adjust he ndividual set-point rate ofthe filters by reference to the water level in theclearwater tank. Another central controller

detects the water level in the inflow channel and

adjusts the inflow control valve to provide the

filters with a flow to correspond with their set-

point rate. Thechange n waterlevel in the filters

canbeasmuch as 300 mm with his system.

Constant ate, variable head ilters operate with afiltered water outlet control structure (weir) to

control theminimum water level just above thatof the media in the filters. The total inflow to the

plant is distributed equally between all the

operating filters. When the filter is clean the

media is just coveredby water and at maximum

clogging thewaterreaches the level in the inflow

channel.

The self-backwash(orStreicher design) system, isa variation of the constant rate, variable head

filter. The inflow to the plant is distributed

equallybetweenalltheoperating filters,byaweir

arrangement. The water surface level in each

filter variesaccording to head loss, but the flow

rate remains constant or each filter. This systemreduces the amount of mechanical equipment

required for operation and backwashing, such as

washwater pumps, and also requires an outflow

control structure and adeeperfilter. Thesefiltershave a use where neither electricity nor

compressed air is available and the water to be

treated has low to moderate levels of suspendedsolids.

In declining-rate filters, flow rate varies with

head loss. Each filter operates at the same, but

variable, water surface level. This system is

relatively simple, but again requires an outflow

control structure (weir) to provide adequate media

submergence.



28 FILTRATION



6 ACTIVATED CARBON FILTERS7

ACTIVATED CARBON FILTERS

29

Theprimary purpose

of filtration is to remove

suspended particles and floc from the water beingtreated. Another dimension is added to thefiltration process by the use of activated carbon

(granular form) as a filter media. The high

adsorptive capacity ofactivated carbonenables ittoremove taste andodour-causing compounds, aswell as other trace organics from the water.

However, not all organic compounds areremoved

withthesame degree ofefficiency.

While activatedcarbon filtration is very effective

in removing taste and odour-causing compounds,the construction, carbonhandling equipment and

operating costs are generally quite high.Activated carbon can be addedto existing filters

provided here is sufficient depthavailable orcanbe incorporated as a separate process. Provision

shouldbemade for regeneration orreactivation of

"spent" carbon (carbon which has lost its

adsorptive capacity) eitheron oroff-site.

Theeffectiveness ofactivated carbon inremovingtaste and odourproducing organic compounds is

well known. The chlorination of water

containing certain natural organic substances,such as humic or fulvic acids or substancesderived from algae may give rise to the

production of trihalomethanes (THMs) orhaloforms. The concern regarding these

substances is reflected in the Drinking Water

Regulations8which state that "Haloformconcentrations must be kept as low as possible".The use of granulated activated carbon (GAC)filters to remove theorganic substances that arethe 'precursors' of THMs is becoming more

widespread.

GAC filters employed for organics removal are

similar to sand filters of the rapid gravity or

pressure type with generally similar design,filtration rates and backwashing arrangements,with the flow rate adjusted to suit the lower

density ofGAC. The GAC is placed in the filter

over the support gravel. Media depth is afunctionof theempty bed contact time (EBCT).GAC characteristics will vary according to thematerial from which it is made whether wood,coconutshell orpeat.

EBCT is normally in the range 5 to3O minutesand will vary for differentmicropollutants. For

pesticides removal an EBCTof 15 to 30 minutes

is used while 10 minutesis considered

adequatefor THMs and VOCs (volatile organic

compounds).

Although GAC removes most micropollutants

efficiently, its adsorption capacity towards someis low, so that frequent regeneration becomes

necessary, withconsequent costs. Breakthroughwill not normally occur for 2 to 3 years, when

usingan EBCTofabout 10 minutes, ifonly taste

and odour removal is required. Most pesticidesmayshow breakthrough in6 to 24 months usingan EBCTof 10 to 30 minutes; THMs in 6 to 12

monthsandVOCs in3 to 9months.



30 FILTRATION



32 FILTRATION

TABLE2: SUlMARYOF ROUTINE FILTRATION PROCESS ACTIONS

Actions Location Frequency Possible OperatorActionsMonitorProcessPefoi-nianceandEvaluateWaterQua//tv Conditions

MgaavteTurbidits Outfiou/lnfioss Intloss at least onceper

• increase sampling frequency shen

day ssaterqualitysvariable. PerformjarOutflow every 2hours tests f equired

• Makenecessary process changesMeasure Colour Infloss/Outflow At eastonce eseo 2hours a) Adjustcoagulanllcoagulant id

dosage

b) Adjust flash mixer/flocculator

mixing intensity

c Change filtration rateDetermineHeadLoss Filter At east twiceper da d) Backwash filter

• Verify esponses to process changes atMeasure Metal ion residuals Outflow Daily -

appropriate imes

OperateFiltersandBackss'ash

Put filter nto ervice Filter Depends on process • See operating procedures in Section

conditions 11.5

Change filtration rate

Remove filter from service

Backwash filter

Change backwash rate

CheckFilterMediaConditionMediaDepthevaluation Filter At east monthly • Replace losi iltermedia

Media cleanliness • Change backwash procedure

Cracksorshnnkage • Send sampleofmedia or esting

Make Visual ObservationsofBacksi'ash OperationCheckformediaboils Filter Ai eastonce per dayor • Change backash rate

wheneserbackwashing

Observe mediaexpansion occurs when ess frequent • Change backuash cycle ime

Check for media carryover into • Adjusi wash rate, air scour

washwater channel orcycle time

Obserse clanty ofwashwater • inspect iltermedia for isturbance

Check Filtration ProcessandBacksi'ash Equipment Condition

Noise/Vibration/Leakage Various Onceperday • Correctminorproblems

Overheating• Notif maintenance/supervisors

personnel ofmajor problems

Accuracyof losss Infloss/outfioss DailyAirscour Monthl • Recalibrate meters orauesBackwash Monthly

Inspect Facilities

Checkphysical facilities Various Once per day • Report abnormal conditions o

Daily maintenance/supervisory personnel

Checkfor lgae buildupon filter

walls and channels • Remove debris rom filtermedia

surface

• Perform outinecleaning/maintenance



INTERACTION WITH OTHER TREATMENT PROCESSES

TABLE3: FILTRATIONPROCESSTROUBLESHOOTING

33

Trigger ibjatorctionsSourceWaterQualityChanges

Possible Process Changes

Changes in:• Turbidity• Colour• pH• Temperature• Alkalinity• Chlorine demand

• Perform necessary analysis todetermine • Adjust coagulant osageextentofproblem • Adjustflashmixer/flocculatormixing• Assessoverallprocessperformance. intensity

• Performjar ests if ndicated • Change frequency of ludge emoval• Make appropriate process changes (see (increase ordecrease)

nght-hand olumn, 'Possible rocess • Change filtration rate(add ordeleteChanges') filters)

• Increase frequencyofprocess monitoring • Adjustbackwashcycle(rate,duration)• Verify response toprocess changes at

appropriate time(besure toallowsufficient timeforchange o akeeffect)

• Add ime orcausticsoda if lkalinity is

low

ClarificationProcessQuality Changes- • .• .• .- - - - • :

Changes in: • Assessoverallprocess performance Samesuggestions as orsourcewater- • Turbidity or loc carryover • Performjar estsif ndicated quality changes• Metal on residual arryover • Check hatselected chemical dosesare

- being applied ---------—-- - - - -- -

• Makeappropriate processchanges• Verify response toprocess changes at

appropriate time

FiltrationProcesschanges/Problems

Changes resulting in: • Assess overallprocessperformance • Adjust oagulant dosage• Head oss increase - • Performjar estsif ndicated • Adjust lashmixer/flocculator mixing• Short filter uns • Make appropriateprocess changes intensity• Media surface sealing • Verify responsetoprocess hanges at • Change requency ofsludge removal• Mudballs appropriate time • Decrease filtration rate• Filter media racks, shrinkage • Manually remove mudballs • Adjustbackwash cycle(rate,duration)• Filterwill notclean • Replenish lost media• Mediaboils • Clearunderdrain openingsofmedia,• Media oss corrosion orchemical deposits when• Excessive head loss filter is outofactive service

• Checkhead loss indicator for orrect

operation

FilterOutflow_Quality_Changes

Changes resulting in: - • Assessoverallprocess performance • Adjust oagulant dosage• Turbidity breakthrough - • Performjar estsif ndicated • Adjust lashmixer/flocculator mixing• Colour • Verifyprocessperformance: intensity• pH (a)Coagulation-flocculation process • Change frequency of ludge removal.• Chlonnedemand (b)Clarificationprocess • Decrease filtrationrate (add more(c)Filtration process filters)

• Makeappropriate process changes. • Changechlorine dosage• Verify response toprocess changes at

appropriate time

-



34 FILTRATION



835

PROCESS MONITORINGANDCONTROL

8 PROCESS MONITORING AND CONTROL

The quality of treated water for human

consumption must,afterdisinfection, meet legallydefined standards as set out in the EuropeanCommunities (Quality of Water Intended forHuman Consumption) Regulations, 1988g.

Parameters tobe monitored under he regulationsare divided into five groups -

organoleptic,physico-chemical, undesirable in excessive

amounts, toxicandmicrobiological.

It has been found in practice that control ofturbidity is, except in rare instances, perfectlyadequate formonitoring process performance. In

avery shortperiod, thecorrelation between ilterinflow turbidity or colour, filter performance(head loss builduprate and filter run time) andfilter outflow turbidity or colour becomes

apparentso that any departure from the normcanbe detected by the use of a turbidityrod andalmost by eye. It is of course necessary toconfirm the visual observation by instrument

readings for recordpurposes. Wheremetals fromtheprimarycoagulant have been detectedin thefinal water,atsignificant evels in relationto thestandards (e.g. aluminium), the level ofdissolvedmetals in the filter outflow shouldbe monitoredat intervals, particularly if any change in thechemical dosing regimehas occurred. In some

plants it may be possible to establish a

relationship between turbidity (or colour) andmetal ion residuals nd this maybe used toreducethefrequencyofmetal ionanalysis.

The recyclingof washwaters has been found tocauseprocesscontrol problems, where therecyclerate is variableand forms a significant proportionof the plant inflow. The optimum means ofhandling washwaters from a process control

viewpoint is to recycle the supernatent waterto abalancing tank or reservoirsupplying the worksto even out theeffects, both in terms of qualityand quantity, on theworks inflow. Where this isnot possible,it shouldbe fed to the inflow, at asuniformarateas possible, upstream of the flashmixer.



36 FILTRATION



OPERATING PROCEDURES ASSOCIATED WITH NORMAL PROCESS CONDITIONS

9 OPERATING PROCEDURES ASSOCIATED WITHNORMAL PROCESS CONDITIONS

37

.9.1 INDICATORS OF NORMALOPERATING CONDITIONS

Filtration is the final step in the solids removal

process. From a water quality standpoint, filteroutflow turbidity will give a good indication ofoverall process performance. However, theoperatormust also monitor the performance ofeach of the preceding treatment processes(coagulation, flocculation, and clarification

/flotation), in which up to 90% of the solidsremoval occurs as well as filter outflow water

quality, in order to anticipate water quality or

process performance changes which mightnecessitate changesin the treatment process suchasadjustmentofthechemical dosage.

Filter inflow turbidity levels can be checked on a

periodic basisby securing agrab sample eitheratthe filter or from the laboratory sample tap (ifsuch facilities are provided). Filter outflow

turbidity maybe monitored

andrecorded

on acoiitinuous basis by an on-line turbidimeter. Ifthe turbidimeter is provided with an alarmfeature, virtually instantaneous response to

process failures can be achieved. As noted

previously the level of dissolved metals in thefilter outflow shouldbe monitored ifmetals fromthe primary coagulanthave been detectedin thefinal water at more than trace levels. In largerworks, an on-line detector for dissolvedaluminium is desirable for the purpose ofdemonstrating thattheresidual s keptatall times

withinthe standard set in the Regulations. It isessential to have the services of an instrumenttechnician available on demand, where on-linemonitors areinstalled.

Forsmallerworks a turbidity rod, madebyfixingtwobrightplatinumwires, one 1mmdiameterandtheother 1.5mm diameter, at right angles to thebottomofarodmarked incentimetres, isauseful

practical guide to turbidity. The depth ofimmersion atwhichone wire disappears while theother remainsvisible whenviewed from above isrelatedto urbidity3asshown inTable4.

Other indicators that can be monitored todetermine if the filter is performing normally

include head loss build up and filter outflowcolour.

DEPTH OFIMMERSION

*(cm)

TURBIDITY

(mgfl SiOScale)

NOTES

210

1000100

Filterclogs

quickly

15 65Filters

operate with

difficulty

30

45

3018

Special careinoperationrequired

80 10Maximum

desirablelimit

A written set of process guidelines should be

developedto assistin evaluating normal processconditions and in recognizing abnormalconditions. These guidelines should be

developed based on water quality, designconsiderations, water qualitystandards and, most

importantly, experience inoperation of heplant.

9.2 PROCESS ACTIONS

In the ormal operation of the filtration process,a

varietyof functions areperformed with emphasison maintaining ahigh qualityfiltered water. Forall practical purposes, the quality of the filteroutflow constitutes the final productquality thatwill be distributed to consumers (subject to

disinfection).

The recommendations of the plant designer andsupplier in regard to process actions should befollowed in all cases. Typicalfunctionsto be

performed inthenormal operation of thefiltration

process include thefollowing:

TABLE4: READINGATURBIDITY ROD



38 FILTRATION

monitorprocess performance;

evaluate water quality conditions (colour and

turbidity) and make appropriate process

changes;

check and adjust process equipment (checkoutputsof chemical dosingplants);

backwash filters;

evaluate filter media condition (media loss,mudballs, cracking); and

visually nspect facilities.

Monitoring process performance is an ongoingactivity. It is essential hat seasonal water qualityvariations and the necessary chemical dose rate

adjustments are planned and actions taken toensure that the finished water quality ismaintained.

Measurement of head loss buildup (Figure9) inthe filter media will give a good indication ofhow well theprefiltration solids removal processis performing. The total design head loss fromthefilterinflow to theoutflow inagravity filter is

usually less than 3 metres at maximum. Theactual head loss from a point above the filtermediatoa reference point in the outflow can bemonitored as "loss-in-head". For example,suppose that a gravity filter isdesigned ora otal

potential head loss of2.5m. Ifatthebeginning ofthe filtrationcyclethe actual measured head lossdue to clean media and other hydraulic losses isO.5m; hiswould permitan additional head ossof2.Om due to solids accumulation in the filter. Inthis example, a working limit might beestablished at an additional l.8m of head loss

(totalof2.3m forbackwashing purposes).

The rateofhead oss buildup is also an importantindicator of process performance. Suddenincreases in head loss might be an indication ofsurface sealing of the filter media(lackofdepthpenetration). Early detection of this condition

may permit the making of appropriate processchanges such as adjustment of the chemical feedrateor the filtration rate.

Monitoring of filter outflowturbidity

on acontinuous basis withanon-line urbidimeter will

provide continuous feedback on the performanceof the filtration process. In most instances it isdesirable to cut off (terminate) filteroperation atapredetermined outflow turbidity evel. Preset the

filter cutoff control at a point where experienceand tests show breakthrough will soon occur

(Figure 10). Results of metal ion residualdeterminations should be checked for correlation

with turbiditybreakthrough. Residual metal iondeterminations should be carried out on a dailybasisifanaccuratecorrelation isnotestablished.

In the normal operation of the filter process, it isbest to calculate when the filtration cyclewill be

completed on the basis of the followingguidelines:

head oss;

outflow turbidity level; and

elapsed run time.

A predetermined value is established for each

guideline as a cutoff point for filter operation.Whenanyone of these levels isreached, hefilterisremoved rom service andbackwashed.

Filter performance from season to season, filterbed to filter bed and from plantto plant maybe

compared by reference to the filter run length in

hours. The reason for this is that as thefilter run

gets shorter, the amount of water used inbackwash becomes increasingly important when

compared tothe amount ofwaterproduced duringthe filter run. Percent backwash water statisticsshouldbe collected.

Although of some use, filter run length is not a

satisfactory basis for comparing filter runswithout considering the filtration rate as well.For example, at a filtration rateof7.5 rn/hour , a27 hour filter run is quiteadequate, whereas, at afiltration rateof4 rn/hour a 27 hour filterrun is

notsatisfactory.

The best way to compare filter runs is by usingthe Unit Filter Run Volume (UFRV) technique.The UFRV is the volume of waterproducedbythe filter during the course of the filter rundivided by the surface area of the filter. This isusually expressed in m3 per m2. UFRVsof200m3 per m2 orgreaterare satisfactory, and UFRVs

greaterthan 300 m3 per m2 are desirable. In the

examples in theparagraph above the UFRV forthe filter operating at 7.5 m/hour would be 202.5

m3/m2, and for the filter operating at 4 rn/hourwould be 108m3/m2.



1.0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.0

OPERATING PROCEDURES ASSOCIATED WITI-I NORMAL PROCESS CONDITIONS

FIGURE10: TYPICAL OUTFLOWTURBIDITY DATA

39

FIGuRE9: MEASUREMENT OF HEADLOSS

E0

—Ripe ingPeriod

0

- ITuzbfdlty

akthrou

0 4

0

-I0

n__________00

—

0:

I0 0

T

0

calPerfo

0 0

unce

0 00 0 0 0

0

0

0

0

,— — deal Performane

•. kI I I I • • • • • I I I I I • I

0 4 8 12 16 20

FILTRATION TIME (Bonn)

24 28 32



40 FILTRATION

Water quality indicators used to assess process

performance includeturbidity and colour (Figure11). Basedon an assessment ofoverall processperformance, changes in the coagulation-

flocculation process or inthe clarification processmayberequired.

At least onceayear:

examine the filter media and evaluate itsoverall condition;

measure the filter media depth for anindication of media loss during the

backwashing process; if noticeable sendmedia sample for analysis and/or adjust

backwash rates;and

measure mudball accumulation in the filter

media, as detailedin Appendix C, to evaluatethe effectiveness of the overall backwashing

operation.

In dailyoperations:

observe the backwash process to qualitativelyassess process performance. Watch for mediaboils

(unevenflow

distribution) duringbackwashing, media carryover into the

washwater channel, and clarity of thewashwater near the end of the backwash

cycle;

upon completion of the backwash cycle,observe thecondition ofthe media surface and

check for filter sidewall or media surface

cracks; and

inspect physical facilities and equipment as

part of good housekeeping and maintenancepractice. Correct or report abnormal

equipment conditions to the appropriatesupervisor ormaintenancepersonnel.

9.3 PROCESS CALCULATIONS

In the routine operation ofthe filtration process, a

variety of process calculations have to be

performed, related to filter operation (flow rate,filtration rate), backwashing (backwash rate,

surface wash rate), water production, and percentof water production used to backwash filters.

Typicalcalculations are shown inAppendix D.

9.3.1 FILTRATION RATE

Filtrationrates aremeasured inm3/m2/hourorper

day, abbreviated to rn/hr or rn/day. For public

water supply in this country, the standardfiltration rate for rapid gravity filters is usually 5

rn/hr and most authorities limit the maximum

filtration rate to between 5 and 7.5 rn/hr.

Problems candevelopifdesignfiltration ratesare

exceeded.

9.4 RECORDKEEPINGAND QUALITYCONTROL

Accurate records ofthefollowing

items shouldbemaintained:

process water quality (turbidity,

temperature, conductivity andcolour);

pH,

process operation ( filters in service, filtration

rates, loss of head, length of filter runs,

frequency of backwash, backwash rates and

UFRV);

process water production (water processed,

amount of backwash water used, andchemicals used);

percent ofwaterproduction used to backwash

filters;

process equipment performance (types of

equipment in operation, equipment

adjustments, maintenance proceduresperformed, andequipment calibration); and

media condition(e.g. congealing

of sand

grains which maybe caused by carry-over of

polyelectrolyte).

Entries in logs should be neat and legible, should

reflect the date and timeof an event, and shouldbe initialled bythe operator making theentry.

An example ofa typical dailyoperatingrecordforwater treatment plant filters is presented in

Appendix A. Irrespective ofthe size of plantornumberoffilters,asimilar ecordsheet should be

kept.



OPERATING PROCEDURES ASSOCIATED WITH NORMAL PROCESS CONDITIONS

41

9.5 FILTERMONITORINGINSTRUMENTATION

To evaluate filtration process efficiency,familiarity with the measurement of turbidity,colour, pH, temperature and conductivity is

essential. Turbidity may be quickly estimated

using the turbidity rod described in section 9.1

and moreexactly, for recordpurposes usinga

turbidimeter. In addition, on-line or continuouswater quality monitors, such as turbidirneters,

soluble aluminium and pH monitors will giveearly warning of process failure and will aid in

making a rapid assessment of process

performance.

Familiarity with methods used to measure filter

media loss and to determine the presence ofmudballs in thefiltermedia will alsobeneeded.

FIGURE 11: FILTRATION PROCESSMONITORING-PARAMETERS AND LOCATIONS

FilterInflow

Metalion esiduals,pH, Temperature,

Turbidity COlOUr

Filter Outflow



42 FILTRATION



1 0 OPERATING PROCEDURES ASSOCIATED WITHABNORMAL PROCESS CONDITIONS

10 OPERATING PROCEDURES ASSOCIATEDWITHABNORMAL PROCESS CONDITIONS

43

10.1 INDICATORS OFABNORMALCONDITIONS

Abrupt changes in water quality indicators suchas turbidity, pH, alkalinity, threshold odour

number (TON), temperature, chlorine demand

(sourcewater), chlorimi residual (in-process), orcolour are signals that the performance of thefiltration process, as well as pretreatmentprocesses (coagulation, flocculation, and

clarification/flotation) should be immediately

reviewed.

During a normal filter run, watch for both rapidchangesin turbidity and head loss buildupin thefilter. Significant changes in either of these

parametersmayindicateanupsetorfailureinthefiltration processorpretreatment rocesses.

Other indicators of abnormal conditions are asfollows:

• mudballs infiltermedia;

• mediacrackingorshrinkage;

• mediaboils during backwash;

• excessivemedia oss orvisibledisturbance;

• shortfilterruns;

• filters that will not come clean duringbackwash;

• algaeonwalls andmedia;and

• congealing of sand grainspolyelectrolyte carryover).

10.2 PROCESS ACTIONS

(indicates

A summary ofroutine iltration process actions is

presented in Table 2. Significant changes insource water turbidity levels, either increases ordecreases, require immediate verification of theeffectiveness of hefiltrationprocess in removingsuspended solids and floc. Aquick determination

of filtrationremoval efficiency can be made bycomparing filter inflow and outflow turbiditylevels with thoseofrecentrecord.

In theevent hat filterturbidity emovalefficiencyis decreasing, look firstattheperformance ofthe

coagulation and flocculation processes todetermineif the coagulant dosage is correct forcurrent conditions. This may require the

performance of jar tests in the laboratory toproperlyassess treatment conditions.

Increasesin source water turbidityand resultantincreases in coagulant feed rates may impose a

greater load on the filters if the majority of

suspended solids and floc are notremoved in theclarification tanks. This condition may requireadecrease in filtration rates (put additional filtersinto service) or more frequent backwashing offilters.

Changesin sourcewaterqualitysuchasalkalinityand pH may also affect filtration performance

throughdecreased

coagulation-flocculationprocessperformance. This isparticularly evident

when source water quality changes result from

precipitationand runoff, or algal blooms in asource water reservoir. It would appear to beAmerican practice to use filter aid chemicals,such as nonionic polymers,, to improve filter

performance nsuch cases. Current Irishpracticewould avoid the use of polymers, or other filteraid chemicals, in such circumstances and wouldconcentrateon the adjustment of the prefiltrationtreatmentprocesses.

Theneed for addition of polyelectrolytes duringperiods when water temperatures areabove12°Cshould be examined. Above this temperature,coagulants work more effectively and at some

plants it may be possible to reduce or omit

polyelectrolytes during theperiod from April to

early October, if the raw water quality is of areasonable tandard. The upflow ratesadopted nthe clarification stageshouldnot exceed1.0 m/hr,unless specialised technologies are employed, ifproblemswith filtration and floc carryover are to

be avoided. Coagulant and polyelectrolyte doserate adjustments shouldonlybe carried out by orunderthe control of a competent personand theeffectsofalladjustments shouldbemonitored.



44 FILTRATION

Increases in filter outflow turbidity may alsoresult from floc carryover from theclarification/flotation process. The optimum flocsize developed in the flocculation process rangesfrom about 0.1 to 3.0 mm. In conventional

filtration, the optimum floc size is closer to 3.0mmforsettling purposes. However, in thedirectfiltration process (no clarification stage) the

optimum floe size is closer to 0.1 mm to permitdepth penetration of the filter media. Whenflotation is not used and larger floe is notremoved in clarification (too light), it will becarried over into the filters, rapidly clogging themedia surface. Hydraulic forces in the filter willshear weakfloes, furthercontributing to turbiditybreakthrough. Re-evaluation of coagulation-flocculation and clarification performance may

be required if floe carryover into the filtersreducesfiltration efficiency. The size of floe canbe estimated by observation, as it is seldom

necessary to make an accurate measurement offloe size.

If backwash problems such as mediaboils,media

loss, or failureof the filter to come cleanduringthe backwash process are encountered, correctiveactions should be taken immediately. Generally,these problems can be solved by adjusting airscourand backwash flow rates, surface washflow

rateorduration, or adjusting the sequence and/orduration of the backwash cycle. In filters with

nozzle-type underdrains, boils are often the resultof nozzle failure. In this situation the filtershouldbe taken out of service and the nozzles

replaced. Possible corrective actions aresummarized inTable3whichgivesa summary offiltration process problems, how to identify thecauses of problems, and also how to correct the

problems.

Problems within the filter itself, such as mudball

formation or filter cracks and shrinkage, resultfrom ineffective or improper filter backwashing.Correction of these conditions will requireevaluation and modification of the backwash

procedures.

If filters are not thoroughly washed, materialfiltered from thewater is retained on the surfaceofthe filter. Thismaterial is sufficiently adhesiveto form minute balls. In time these balls ofmaterial come together in clumps to formmudballs. Usually as time goes on, filter media

becomes mixed in to give it additional weight.When the mass becomes greatenough, it causesthe mudballs to sink into the filter bed. These

mudballs, if allowed o remain, will clogareas in

the filter. Generally. regular and controlled

backwashing willprevent mudball formation.

10.3AIR BINDING

Shortened filter runs can occur because of aitbound filters. This is caused by the release ofdissolved air in saturated water due to a decreasein pressureor air trapped in the filter media oncompletion ofthebackwash cycle. Air isreleasedfrom the water when passing through the filterbed by differences in pressure produced byfriction through the media. Subsequently thereleased air is entrapped in the filter media. Air

binding will occur more frequently when largehead losses are allowed to developin the filter.Whenevera filterisoperated oahead loss whichexceeds the headof water on the media, air willbe released. Air bound filters are objectionablebecause the air prevents water from passingthrough the filter and causes shortened filter runs.When an air bound filter is backwashed, thereleased air can damage the filter media. Whenair is released during backwashing, the mediabecomes suspended in the washwater and iscarriedoutof the filter. Significant media loss

can occur,particularlywhenlightermediasuch asactivatedcarbonisused.

10.4 EXCESSIVE HEAD LOSS

Shortfilter runs mayresult from increased solids

loading, excessively high filtration rates,excessivemudball formation in the filter media,or clogging of the filter underdrain system.Possible corrective actions are summarized inTable3.

If excessive head loss persists in a filter after

backwashing, a epresentative sample of the filtermedia should be sent for analysis, to include thedetermination of:

hydraulic sizeofmedia;

organic content;

particle size distribution;

uniformity co-efficient;

grain shape actor;

grainsphericity; and



45

OPERATING PROCEDURES ASSOCIATED WITHABNORMAL PROCESS CONDITIONS

grain porosity.

The degradation of thefilter media over timecan

be determined by calculating the 'resistancetofiltration'and the 'efficiency of iltration'. If he

natureof thefilter media does not account for the

headloss, the filter underdrain system and the

head loss measurement equipment should bechecked for malfunctioning. High head lossescanbecausedby reduction in thesize andnumber

ofunderdrain openings. The underdrain openingscan be reduced in size or clogged by media,corrosion orchemical deposits. Pipeworkshould

also be checked for blockages; deposits, for

example,have been asociated with the use of

magnetic flowmeters and sodaash.



46 FILTRATION



11 STARTUPANDSHUTDOWN PROCEDURES

1.1 STARTUP AND SHUTDOWN PROCEDURES

47

11.1 CONDITIONS REQUIRING

IMPLEMENTATION OF STARTUP ANDSHUTDOWN PROCEDURES

Startup and shutdown of filtration is a routine

procedure in most water treatment plants, unlikethe coagulation, flocculation and clarification

processes. This is true even if the treatment plantis operated on a continuous basis, since it isconmon practice for a filter to be brought intoservice or taken off line for backwashing. Aclean filter may be put into service whena dirty

filter is removed for backwashing, when it isnecessary to increase filtration rates, or when

plantproduction needsto be increased as aresultof increased demand for water. However, most

plants keep all filters on line except for

backwashing and in service except formaintenance. Filtersare routinely taken off linefor backwashing when the media becomes

cloggedwith particulates, turbidity breakthroughoccurs,ordemands forwater arereduced.

11.2 IMPLEMENTATION OFSTARTUP/SHUTDOWN PROCEDURES

Typical actions performed in the startup andshutdown of the gravity filtration process areoutlined below. Theseprocedures also generallyapplytopressure ilters.

Figures 12 and 13 illustrate sectional views oftypical gravity filters, Thefigures show the valve

position and flow patterns in the filtration andbackwash mode of ilteroperation.

11.3FILTERCHECKING PROCEDURES

The followingactions should be taken to checktheoperational tatus ofafilter:

• check that filter flow control equipment isoperational;

• check that filter media and washwaterchannels areclearofall debris such as leaves,twigsand tools;

• ensure allaccess covers andwalkway gratings

are inplace;

• ensureprocess monitoring equipment such ashead loss and turbidity systems are

operational; and

• check source of backwash water to ensurethere is sufficient volume available. Thiscould be an elevated reservoir, washwater

storage tankorothersource.

11.4BACKWASHPROCEDURES

Filters shouldbewashed priortoplacingthem inservice. There is avarietyofdifferentbackwash

arrangements for filtration plants and these will

vary with individual plant designs. Usually filter

washing is divided into three distinct operations

Operation DetailsAir Scouring consisting of low pressure /

high volume compressed airONLYappliedup through thefilterbed.

Washing(low rate)

consisting ofa ow rate of low

pressure washwater ONLY

applied up through the filter

bed, with wash water drainedtowaste.

Washing(high ate)

consisting of a high rate oflow pressure washwaterONLY applied up through thefilter bed, with washwaterdrained owaste.

asfollows:

A furthercombination maybe the nclusion ofanair scourand low rate wash together for aperiodoftime. Thismaysupercede the necessity for alow rate wash.

Generally wash cycles are automaticallycontrolled,withpre-set wash rates andadjustabletime durations. The duration ofeach part of thewashcycle can therefore be adjustedin order toensure efficient ilter washing.



48 ALTRATION

If filters are to be washed automatically, checkthat the length of cycle times set for air scour,backwash and surface wash cycles are "correct".

"Correct" imesvaryfromplanttoplant andwith

timeofyear. These settingsshould be based on

physical observations of actual time required to

clean hefilter.

Iffilters are usually washed automatically, it is

good practiceto occasionally use a manual wash

procedure to ensure efficient cleaning of themedia during thewashcycle. This is carriedout

asfollows.

Each filter to be washed is drained down to the

level required to commence washing. This isachieved by closure of the filter inlet, whilst

maintaining thefilter outlet flow. Thisallows thefilter to drain down evenly, without any unduestress to, or compaction of, the filter bedprior to

washing.

When the water level has drained down to the

required level,a fewcentimetres above hemedia,the backwash cycleis started. The surface of themedia shouldshowanevenspread ofburstingair

bubbles coming through. Any unevenness in thedistribution shouldberegarded as an ndication of

potential difficulty, itscauseinvestigated asfaras

possible andtheobservation logged.

If ufficient time has been allowed orcleaning ofthe filter media the backwash water coming upthrough themediabecomes clear. This generallytakes from threeto eightminutes. If looding ofwashwater channels orcarryover of ilter media sa problem, the backwash rate must be reduced.Thismaybe accomplished by adjusting thebackwash control valve, thereby throttling heamount

ofwashwaterused.

Inmanywatertreatment plantsbackwash water isallowed to settle in a tank, and then the

supematant (clear, top portion of water) is

pumped back, if possible to the reservoir or

balancing tankfeedingthe works, to be recycledthrough the plant. Usually it isbest to graduallyaddthebackwash water to theheadworks of thetreatment plant(ahead of theflash mixer). This

is becausea sudden return requires changes inchemical dosagesdue to the additional flow andincreased turbidity.

Procedures forbackwashing afilter undermanualcontrol are providedoverleaf: (Referto Figures12 and 13).

11.5 FILTERSTARTUPPROCEDURES

The initial few hours after a filter is placed in

service isa time whenturbiditybreakthrough canpose aproblem. For hisreason,filters shouldbeeased into service o avoid hydraulic shock loads.

Afterwashing, filters shouldbe broughtback online gradually. With automatic equipmenthis is

generally donebyagradual openingof hefilter'soutflow valve (usually an actuated butterflyvalve). Manual operations require a gradualincrease of the amount of water treated by the

filter,usually achievedby anautomatic slowstartcontrol device. Many plants haveprovisions to

waste some of the initial filtered water (byopening V-6). This provision can bevery helpfulif turbidity breakthrough occurs. Turbidity

analyses of filtered water should be carriedoutand processadjustments madeasnecessary.

11.6FILTERSHUTDOWN PROCEDURES

Remove thefilter fromservice by:

•closing

inflow valve(V-l);

• allowing filterto drainto correctbackwashinglevel; and

• closing outflow valve (V-5).

Backwash the filter, as describedin precedingsection. If the filter is to be out ofservice for a

prolonged period, drain water from filterto avoid

algae growth. Note status of the filter in

operating record. Backwash againbeforeplacingfilter on-line',asnotedpreviously.



STARTUP ANDSHUTDOWN PROCEDURES

TABLE5:PROCEDURE FOR BACKWASHINGA FILTERUNDERMANUALCONTROL

Step ActionLog length offilterrun since lastbackwash

2 Logfilterloss ofhead(L.O.H.)

3 Closefilterinflow valve(V-1)

4 Allow filter to drain tocorrectbackwashing level

5 Closefilteroutflow valve (V-5)

6 Opendrainvalve(V-4)

7 Open air scourvalve(V-7)

8 Start air blower(s)

9 Runairscour orpre-determined period

10 Open backwash valve(V-3)

11 Startwash water pump(s) at owrate

12 Run low rate wash forpre-determined period

13 Stopair blower(s)

14 Closeair scourvalve(V-7)

15 Opensurface washvalve(V-2) (optional)

16 Increase wash rate toHighRateWash

17 RunHighRateand surface wash (optional) forpre-determined period

18 When washwater going o drain becomes clear:• Closedrain valve (V-4)• Allow filtertorefill

19 Once filter is refilled, put nto service by:• Stopping wash waterpump(s)• Closing backwash valve(V-3)• Closing surface wash valve (V-2) (optional)• Opening ilter inletvalve(V-i)

• Opening ilteroutletvalve (V-5)

20 Logdurationofeach partofwash and watervolumes used

21 Llossofheadatcommencement ofservice run

49



50 FILTRATION

FIGu1u13: BACKWASHMODEOF OPERATION

FIGuRE 12: FILTRATIONMODEOF OPERATION



PROCESSANDSUPPORT EQUIPMENT OPERATION ANDMAINTENANCE

12 PROCESS AND SUPPORTEQUIPMENT OPERATIONAND MAINTENANCE

51

12.1 TYPES OF EQUIPMENT

To runa iltration process he operating personnelmust be competent and familiar with the

operation and minor(preventive) maintenance of

a varietyof mechanical, electrical, andelectronic

equipment including:

• filter control valves;

• backwash and surface washpumps;

• air (dryer)compressor ndblowers;

• flowmeters and leveL/pressuregauges;

• water qualitymonitorssuch as colorimeters,

thermometers, pH and conductivity meters,turbidimeters and dissolved aluminium

detector/recorders;

•process monitors(head

loss and waterlevel);and

• electrical filtercontrol systems.

Since awide variety ofmechanical, electrical, and

electronic equipment is used in the filtration

process, the operating personnel should be

broadly familiar with the operation and

maintenance instructions for each specific

equipment temorcontrol system intheplant.

12.2EQUIPMENT OPERATION

Beforestartinga piece ofmechanical equipment,such as a backwash pump, be sure that the unithas been serviced on schedule and itsoperationalstatus spositively known.

After startup, always check for excessive noiseand vibration, overheating, and leakage (water,

lubricants). When in doubt about the

performance ofa pieceofequipment, refer tothe

manufacturer's nstructions.

Much of the equipmentused in the filtration

process may be automated and only requireslimited attentionby operating personnel during

normal operation. However, periodic calibration

and maintenance of this equipment is necessary,and this usually involves special procedures.Detailed operating, repair, and calibration

procedures are usually described in the

manufacturer's orsuppliers literature.

12.3PREVENTIVE MAINTENANCEPROCEDURES

Preventive maintenanceprogrammes aredesignedto ensure the continued satisfactory operation oftreatment plant facilities by reducing the

frequencyof breakdown. This is accomplished

by performing scheduled or routinemaintenance

on valves, pumps, and other electrical and

mechanical equipmenttems.

In the normal operation of the filtration process,routinemaintenance functions mustbe performedas part of an overall preventive maintenance

programme. Typical unctions include:

•keepingelectric motors free ofdirt, moisture,and pests (spiders, flies, larvae, rodents and

birds);

• ensuring good ventilation (aircirculation) n

equipmentworkareas;

• checkingpumps and motorfor leaks, unusual

noise,vibrations oroverheating;

•maintaining proper lubrication and oil levels;

• checking bearings for overheating and proper

lubrication;

• checkingfor proper valve operation, leakageor amming;

• checkingautomatic control systems orproper

operation;

• checking air/vacuum relief systems for proper

functioning, dirt andmoisture;

• checking chemical delivery lines for leakage

[chemical delivery lines should be colour



52 FILTRATION

coded, placed in ducts and lengths

minimisedi;

verifying correct operation of filtration and

backwashcycles byobservation; and

inspecting filter media condition (look for

algae and mudballs and examine gravel and

media forpropergradation).

Accurate recordkeeping is the most importantelementofany successful preventive maintenance

programme. Theserecords provide operation and

maintenance personnel with clues fordeterminingthe causes ofequipment failures. Theyfrequently

can be usedto forecast

impendingfailures thus

avoiding costly repairs.

12.4 SAFETY CONSIDERATIONS9'10'11

The filtration process does not normally involve

the useof chemicals. There are howeverhazards

involved for theplantoperators, which shouldbeidentified in the Safety Statement prepared for

each treatment works, as required by the Safety,Health and Welfare at Work Act, 1989 and

Regulations made under the Act. Reference

should be made to this Safety Statemenl by all

persons involved in the operation and

maintenance of he works.

Some of the potential hazards, which might be

encountered in differentareas and operations in

the filtration process are listed in Table6.

TABLE 6:POTENTIALHAZARDS N FILTRATION

Area Uperation

Electical Equipment

Mechanical Equipment

OpenWater Surfaces

Confined Space

PotentialHazard

• earthing of ools• locking out and tagging of switches and panels when

servicing equipment• electric shock due to lying waterorgrounding onpipes

• removal ofguards from rotating equipment• locking out and tagging of switches and panels when

servicing pumps, automatic valves andotherequipment• slippery surfaces due to ubricant spills• wearing of ooseclothing in the vicinity of rotating

equipment

• damage tohandrails or ailure toclose safety chains• slippery surfaces on stairways or adders due to spillages or

use ofunsuitable footwearS

• Hazardous atmospheres (toxic or explosive gases, lack of

oxygen)• Presence ofdust



GLOSSARY 53

GLOSSARY

ABSORPTIONThe taking inorsoaking up ofonesubstance intothebodyofanotherbymolecular orchemical action.

ACTIVATEDCARBON

Adsorptive particles or granules ofcarbonusually obtainedby heatingcarbon. Theseparticles orgranules

haveahighcapacity oselectively remove certain trace and solublematerials from water. -

ADSORPTIONThegathering ofagas, liquid,ordissolved substanceon the surface or nterface zoneofanother material.

AIRBINDING(LOCKING)

The clogging ofafilter,pipe or pumpdue to the presence ofair releasedfrom water. Air contained in the

filter media is harmful to the filtrationprocess. Air can preventthe passage of water during the filtration

process and can cause the oss offilter mediaduring thebackwash process.

ALGAE

Algaeare primitive organisms whichareusually classified as plants. There are hundreds ofdifferent ypes,

many of themmicroscopic, whichmaybecome visibleby multiplication. Whenpresent to excess theycause

troubleby blocking ilters. Outbreaks varywith theregionand theseason.

BACKWASHINGThe processof reversing the flow of water, either aloneor in association with air, back through the filter

mediatoremove theentrapped solids.

BASE METALAmetal (suchas iron) whichreacts with dilutehydrochloric acidtoformhydrogen.

BREAKTHROUGHA crackor break in a filter bedallowing thepassageofflocor particulatematter hrough afilter. This will

cause an ncrease in filteroutflow turbidity. Abreakthroughcanoccurwhen a filterisfirst placed inservice,

when theoutflow valve suddenly opensor closes, and during periodsofexcessive head loss through the

filter.

COLLOIDS

Verysmall, finely divided solids (particles that do notdissolve) hat remain dispersed ina liquidfor a long

time due to their small size and electrical charge. When most of the particles in water have a negativeelectrical charge, theytend to repeleachother. Thisrepulsionprevents theparticles from becoming heavier

and settlingout.

COLOUR

Manywaters havea distinct colour, normally due to the presence of complex organic molecules derived

fromvegetablematter(suchas peat,leaves,branchesetc.,), evenafterall turbidity has been removed. This

is expressed in erms of theplatinum-cobalt scale Hazen units). Exceptionally, naturalcolourmaybe due to

thepresenceofcolloidal ironand/or manganese inawater.

CONVENTIONALFILTRATIONAmethod of treating water which consistsof headditionofcoagulantchemicals, flashmixing, coagulation-

flocculation, clarification bysedimentation/ lotation and filtration.



54 FILTRATION

CRYPTOSPORJDIUM

The general descriptive term for the parasite Crvptosporidiuin Parvuni (C. Parvuni). C. Parvum is the only

species of cryptosporidium knowm to cause disease in man. The environmentally resistant transmittable

form ofcryptosporidium excreted in the faeces ofan infected host is calledanOocyte.

DIRECT FILTRATIONA method of treatingwaterwhich consists ofthe addition ofcoagulant chemicals, flash mixing, coagulation.minimal flocculation, and filtration. The flocculation facilities are occasionally omitted, but the physical-chemical reactions will occur to some extent. Theclarification/flotation process stage is omitted.

EFFECTIVE SIZE (ES)The diamter of the particles in a granular sample (filtermedia) for which 10 percent of the total grainsare

smallerand 90 percent larger on a weight basis. Effective size is obtained by passing granularmaterial

through sieves withvaryingdimensions ofmeshandweighing the material retained by each sieve.

FLASHMIXERA chamber in which

coagulantsare stirred into the raw water with considerable turbulence, inducedeither

hydraulically ormechanically.

FLOCFloe is the fine cloud of spongey particles that form in water to which a coagulant has been added. The

particles are basically hydroxides. commonly of aluminium or iron. They accelerate the settlement of

suspended particles byadhering tothe particles and neutralizing such negative charges asmaybe present.

FLOCCULATIONFlocculation is the practiceof gently stirring water in which floc has formed to induce the particles to

coalesce and 8row and thus settle more rapidly.

FLUIDIZATIONTHRESHOLDThe fluidization thresholdis the point during backwashing where the hydraulic pressure loss through the

filterbedequals the dead (submerged) weight of the filtermaterial.

FLUIDIZEDA massof solid particles that is made toflow like liquid by injection of water or gas is said to have been

fluidized. In water treatment, a bed of filter media is fluidizedby pumping backwash water andlor air

through thefilter.

GARNETA group of hard, reddish, glassy,mineral sands made up of silicates ofbase metals (calcium, magnesium,ironand manganese). Garnet hasahigherdensity thansand.

GIARDIA JNTESTINALJS (Giardialamblia)Aprotozoan parasite capable of infectingmanand causingdiarrhoea.

HEADLOSS

The head, pressureor energy lost by waterflowing in a pipeorchannel as a result of turbulence caused bythe velocity of the flowing water and the roughness of the pipe, channel walls or restrictions caused byfittings. Waterflowing in a pipeloseshead,pressure or energy as a esult offrictionlosses. The head loss

througha filter isdue to friction losses caused by material building up on the surface or in the interstices ofthe filtermedia.

HYRAULICSIZE

The hydraulic size of filter material is that grain size which gives the same surface area, and therefore thesame hydraulic behaviour,as the mixture ofsizes in the filtermaterial.



GLOSSARY 55

IN-LINE FILTRATIONThe addition ofchemical coagulants directly o the filter inlet pipe. Thechemicals aremixed bytheflowingwater. Flocculation and sedimentation facilities are eliminated. This pre-treatment method is commonlyused in pressure filterinstallations.

INTERFACEThe common boundary layerbetween two substances such as water and a solid (metal); or between twofluids such as waterand gas(air);orbetweena iquid (water) an another liquid (oil).

NTU

Nephelometric Turbidity Unit, numerically equivalent toJacksonTurbidity Unit.

PERMEABILITYThe property of a material or soil that permits considerable movement of water through it when it issaturated.

POREAverysmallopenspace in a rockorgranularmaterial.

POROSITYThe ratio (normally expressed as a percentage ) of thevolume of the space between grains (voids) to theoverall volume of the granular material. This will vary depending on howthematerial has beenhandled,whether tipped, backwashed orpacked. Alternatively calledvoidage.

SLURRYA watery mixture or suspension of insoluble (notdissolved) matter; a thin watery mudor any substances

resemblingit(such as a

gritslurryora ime

slurry)

SPECIFIC GRAVITY

(1) Weight of a particle, substance, or chemical solution in relation to theweightof an equal volume ofwater. Water hasa specific gravity of 1.000at4°C (or39°F). Particulates in raw water mayhaveaspecificgravity of 1.005 to2.5.

(2) Weightofaparticulargas inrelation o an equal volume ofair at thesame emperature andpressure airhasaspecific gravity of 1.0). Chlorine forexample hasaspecific gravity of2.5 as agas.

SUBMERGENCEThedistance between thewatersurface and the mediasurface inafilter.

TURBIDIMETERAn instrument for measuring and comparing the turbidityof liquids by passing light through them and

determining howmuch light is reflected bytheparticles in the liquid. The normal measuring range is 0 to100andisexpressed as Nephelometric Turbidity Units(NTUs).

UNIFORMITYCOEFFICIENT(U.C.)The ratio ofthediameter ofa grain ofa size that is barelytoo largeto pass through a sieve that allows60

percent ofthematerial (byweight) to pass through,to thediameter of a grain of a size that is barely too

large to pass through a sieve that allows 10 percent of the material (by weight) to pass through. The

resulting ratio is ameasure of thedegreeofuniformity inagranularmaterial.

Uniformity Coefficient = ParticleDiameter 0% /ParticleDiameter 0%

VISCOSITYThe property ofa substance to offer internal resistance to flow. Specifically, the ratioof the shear stress tothe shearstrain.



56 FILTRATION



APPENDIX A 57

APPENDIX A: TYPICAL FILTERS OPERATING RECORD

SHEET



58 FILTRATION



APPENDIX A 59

Sheet No.: 3/F

NAMID AIJFHOR1TY

'a€4de9 WATFR TRFATMRT PLANT

FILTERS

Daily Operating Record Date: 7/11/94

Filter Tine&Date Hours Operated Head Loss Backwash

No. Stail Stop TodajPrevious Total Start(m) Stop(m) Minutes OutityjpNotesoncondition of ilters

&problems inope_

1 04/11 07/?! 11:30

0S30 /1.30

2 05/11 '4:00

17:00

63:30 75:00 0.15 F.15 6 10! owJtatdt12

amalt'caeemed 2.30

31 55 0.155

3 03/!! 24:00 54 71 0.165

/100 :

4 03/11 07/1/ /1:00

14:00 /1.00

12:00 100:00 0.15 1,1 5 14 Stt F130.

a(eemed 9.00

5 05/!! 24:00

09:15

31:45 62:45 0.145

6 04/11 24:00

17:20

30:40 54:40 0.16

Shift I Operator

No.of ilters washed 2 Average Filtration Rate -ni/hr 4.97

'I9€31

AverageRun-Hours -TotalWash Water-m3

Percent ofWaterFiltered

17.5 MaximumFilter Rate -m3fhr 720

115 TotalWaterFiltered-n

1.01 No.ofFiltersOperating

/7160

6

AV.TiIIr ofWash -Minutes 5.5 Filters OutperWash -Mins 30

WaterQualitySampling Integrator Readings •

Location Turbidity Colour

Inflow

Temp (C) pH Timu Location Units Read'g/Tin Difference from Previous

Outflow I

Inflow Production Report Percentage &Reniirks

Outflow Total Inflow Oj.m 100%

Inflow To SupplyOutflow To Site Storage

To Filter Backwash

Other --



60 FILTRATION



APPENDIX A 61

NAMED AUTHORiTY

WATFR TRFATMENT PLANT

FLLTFRS

Sheet No.: Daily Operating Record Date:

Filter Tirre&Date Hours Operated Head Loss Backwash Notesoncondition of ilters

Start Stop Today Previous Tota1Start(m) Stop (rnj_ Minutes[Quantity m3 &problems in operation

1

2'I - . -

— -.--.. ——--....— I -- —5

6

Inflow

Outflow- Production_Report

Totalthflow'

Cu.mPercentage &Remarks

100%

Inflow ToSupply

9 tfiow To SiteStorage[ToFilterBackwashj

.

______ ______ _______ _______ ______ __________ Shift I Operator

[No.of ilters washed - ____ AveragçiltrationRate -mhr

AverageRun-Hours -. - MamarnFilter Rate -m3/hr ________

TotalWash Water- n3 TotalWaterFiltered m3

Percent ofWaterFiltered _______- No.ofFiltersQperating _______

TiofWas_TNinutes. [Filters OutperWash -Mins ________

WaterQualitySampling ____ _____ IntegratorReadings -

Lthca1L Tuthidy Colour_Temp C) pH Tint Location Units Readg/Tme Difference fromPrevious

Inflow

Outflow [

- - __________ _______________________

Other - . _______ ____



62 FILTRATION



APPENDIX B 63

APPENDIX B: CONVERTING TO SI (METRIC) UNITS

CONVERSION TO SI(METRIC)

UNITS FROM IMPERIAL

(FPS )UNITS'4"5

The SI ( orMetric)base units ofmost interest forwatertreatment are thoseof time, massand length. Theunitoftime is the second(symbol 's'),whichis the same in bothImperial and SI systems. The unitofmassis thekilogram symbol 'kg') (1 kg = 2.20462 lbs), which for allpracticalpurposes is themassof 1 litre of

pure waterat20°C. Theunit of length is themetre (symbol 'm') (1 m= 39.37008 inches). The other unitsused in water treatment are derived from these base units. The prefixkilo (k) means 1000 times the unit

following - kilogram (kg) is 1000 grams, kilometre km) is 1000 metres and the prefix mega (M) means

1,000,000 times theunit following but megagram is customarily referred to as a metric tonne. The term

megalitre (Ml) is frequently used for 1000 m3. Conversely, theprefixmilli(m) means one thousandth partof

the unit following - milligram (mg) is 1 / 1000of a gram,millimetre (mm) is 1/1000 ofa metre, millilitre(ml) is 1/1000 ofa litre; the prefixmicro (t - he Greek letter mu ) meansone millionth part of theunit

following - microgram(.tg) is 1/1 000 000 of a gram, while the prefix nano (n) means one thousandmillionth partof the unit following - nanogram (ng)is 1/1 000 000 000 ofagram.

The conversion actors mostneeded inrelation owatertreatment are:

1 gallon (g) =4.54609 itres

1 m3 (1000 1)=219.9736 (220)gallons

1 ft2=0.0929m2

1 m2= 10.76392 ft2

14.7p.s.i.= 1.0 1325bar= 1 atm

The term "parts per million" (ppm) wasformerly widely usedinrelationto thedosingofchemicals to water.This termimplies a elationship ofeithervolume with volumeorweight withweight, butsince thedensityofwater is almostexactly unity, this term is in effect synonymous with "milligrams per litre" (mgfl) and is

widely regardedasbeing so. However thepreferred usageismgll,which is anexpression ofmassorweightofchemical perunit volumeofwater, ratherthanppm.

Clarification tanks are rated in terms of flow per unit area per hour (gallons per ft2 per hour,in Imperialunits)with alternatively an upflowrate in feetperhour. In SI units,flow perunit areaper hour is m3 /m2 /hour or, cancelling m2, m/hour,whichisfrequently expressed as anupflow ate in mmIsec.

A standard rateofslow sand filtration is50gallons persquare footperday.

Converting toSI: 50/(220 x24x0.0929)=0.1019m3/m2 /hour=0.10m hr.

A standard rateof apidgravity iltration is 100gallons persquare footperhour.

Converting toSI: 100/ 220xO.0929)=4.8928 m3Im2 Ihour =4.9 mIhr.

5m/hr = 102.2 gallons per square footlhr.



64 FILTRATION



APPENDIX C 65

APPENDIX C: MUDBALL EVALUATION PROCEDURE1'4

FREQUENCYOFMUDBALL EVALUATION

Use this procedure onamonthlybasis,if mudballs inthe topof he filtermaterialareaproblem. An annualcheck is sufficient ifmudballsarenotnormally aproblem.

PROCEDUREFOR EVALUATION1. Backwash the filter to besampled and drain the filtertoat east 300 mm (12 inches) below the surface of

thetop media layer(or layerof nterest).

2. Push themudball sampler some 150 mm(6 inches) into thesand near one cornerof the filter. Tilt thehandle until itisnearly evel then lift the sampler ullofmedia andempty its contents intoabucket.

3. Takesamples neartheother cornersand at thecentre ofthe filter andplace

contents nbucket.

4. Hold a 10 mesh (9.5mm/ 8") sievein a bucket or tub ofwaterso that it is nearlysubmerged. Take ahandfulof thematerial removed fromthefilter andplaceitin thesieve. Gently raiseand lower he sieveabout 12 mm ('/2") at a timeuntil thefilter sand is separated from the mudballs. Shift themudballs toone sideby tipping thesubmerged sieve andshaking itgently.

5. Repeat the process until all the material taken from the filter has been washed in the sieve and themudballs separated. If he volume ofmudballs impedes thewashingprocess, move someof them o the

measuring cylinder used inthe nextstep.

6. Place 500 ml ofwater ina 1000ml graduated cylinder. Usea largeror smallercylinder depending on the

volume ofmudballs. When water hasceased to dripfromthemudballs on thesieve,transfer hem o thegraduated cylinder and record the new water level in it. The volume ofmudballs in ml is obtainedbysubtracting500 from the new water evel.

The total volume ofmaterial removed fromthe filter,if the 'mudballsampler' was full each time, would be3,540mJ. This can bechecked by drainingandmeasuring, by displacement, the volume ofsand washed fromthe mudballs. Thepercentage mudball volume iscalculated as

Mudball Volume ml)x 100

3,540

Mudball Volume (%) ConditionofFilteringMaterial

0.0- 0.1 Excellent0.1 -0.2 Vervood0.2-0.5 Good0.5-1.0 Fair1.0-2.5 FairlyBad2.5-5.0 Badflyer5.0 VeryBad

Thecondition of hefiltermaterial isevaluatedby reference to theprecedingTable.

Mudballs sinkmore readilyinanthracite thansand. Thereforemodify he aboveprocedure tocollectsamples fromthebottom 150 mm(6" )ofanthraciteby temporarily removing the upper ayers.



66 FILTRATION

FIGURE 14: MUDRALLSAMPLER

75 mm diametercopper tubing

150 flfll 1— 200



APPENDIXD 67

APPENDIX D: HYDRAULIC CALCULATIONS

HYDRAULIC CALCULATIONS5'6"1'12

FILTRATION RATE

In ordertocalculate hefiltration rate, the surface areaof the filter media(not he areawithin the filter walls)must firstbedetermined. The length of hefiltermedia surface maybemeasured between ilter wallsbut he

width, exclusive ofwashwater channels, must be measured when the filter isdrawndown forbackwashing,unless reliable drawings, from which measurements maybe taken, areavailable. Theoutputfromthe filter

for a specific period, say one hour, after it has fullymatured, maybe read from the flow meter, if one is

provided. Thefiltration rate is found by dividing heoutput nm3by thesurface areainm2.

Example:Length of ilter betweenwalls =3.5 mWidth offiltermedia =2.0m (fromwall towashwater channel)

Surface areaoffilter media = 7m2.

Outputoffilterin 1 hour(metered) =33.6m3

FiltrationRate =33.6m3perhr/7m2

=4.8m/hr.

The verage filtration ate forafilter runmaybe calculated similarly f he length of run (say 50 hours)andthe output say 1680m3)are known.Theaverage filtration ate=1680m3/50hours/7m2=4.8mIhour.

Theoutputof hefilter,ifameter s notprovided, maybeestimatedbyclosing the nlet valve andmeasuringthedrop in the water level in the filter in, say, 5 minutes. The estimated hourly filtration rate would betwelve imes the dropin level measured. In theexample abovethedrop in water level in 5 minuteswas 400

mm,so thedrop in an hourwould be 4.8 m, i.e., the filtration rate. The outputof he filter is the filterrate

bythe areaof hemedia:

4.8x7=33.6m3/hr.

The filtration rate for abankof filters all of the same size can be calculated by dividing themeteredoutputfromthebankby the otal filter area(the numberof iltersmultipliedby the ndividual filterarea).

UNITFILTERRUNVOLUME

The Unit Filter RunVolume (UFRV) canbeavaluable indicatorof ncipientor futureproblems inafilter,aswellasausefulmeasure ofcomparison between individual filters. The UFRV or thefilter runquoted in he

example above[output 1680m3 fromafilterofareaof 7m2I s 240m3 /m2.

In a case where the output from an individual filter is not metered, the UFRV can be calculated,byestimating the filter ratea number of timesduring the course of the run from themeasured drop in filter

level, andmultiplying theaverage filter ratebythe length of un. The filterrate, in theexample givenaboveis4.8 m/hr(or 4.8 m3/m2 /hr). If he length offilter run were 52 hoursand30minutes he UFRV for the

filterwouldbe4.8x52.5=252m3/m2.



68 FILTRATION

FILTER BACKWASHING

Backwashing is the reversal of the flow of water through a filter, possibly preceded or accompanied by air

scour, toremove accumulated matter emoved from themedia.

The hydraulics of filter operation have beenderived from a dimensional analysis of the factors influencingfriction losses, based on initial work by Darcy in France on friction losses in pipes. This work was

developed by Weisbach, Kozeny andCarman.

The Kozeny equation describes the pressure gradient. or laminar low, inabedofgranularmaterial,as

= 5(1—

h e3

where

Apis the pressure difference across a bed ofdepth h.e is the porosity of themedia,S is the specific surfaceperunit volume ofgrain (=6/d for spheres diameter d),

/1 is theabsoluteor dynamic viscosity ofwater,and

V is theapproach velocity of he water at the bed surface.

During backwashing the particles of filter media are suspended in the flow of water. Fluidization of the

filter bed occurs when the drag force or the head loss thrOughthe bed equals the submerged weightof the

media:

(pç-p)(1—e)pgwhere

g is the gravitational constant,

p. is thedensity of hemedia grains, and

p is the density ofwater

A p/h cannot exceed the above limit because the porosity 'e' automatically increases as a result of bed

expansion whichcorrects hehead oss. Theimportant parameters in theaboveequations andthe importantcharacteristicsof iltering materials) are thegrainsize (d) orsurface area(S),thedensity difference (Ps -

Pw)

and the porosity e'.

Backwash water is normally pumped, by fixed rate pumps, through the filters at the rate chosen by the

designer to suit themedia being used. The rate is chosento expand the bed by the requiredpercentage for

efficientcleaning without oss ofmedia over the weir walls. Losses ofmediamayoccur ifair scour pumpsare startedwhile high ratebackwash is inprogress, although interlocks to prevent this are normally provided.

While the density affects the fluidization threshold linearly, the grain size occurs as a square, so thata 10%

increase in grain size requiresa 20% increase in backwash rate. Sieves, as used in analysis of filtering

materials, are spaced at 20% intervals (4J2) and hence a material which is one sieve size coarser requires abackwash rate 44% higher. A specific grainsize must be used in makiug calculations with these equations,

although themediamaynotbe monosize. The parameter needed is the size which givesthe same hydraulicbehaviour asthemixture ofsizes in the media(same surface areaperunitvolume) and is called the hydraulic

size. Thismaybe calculated from the results ofa sieve analysis of the media. The method and an example

aregiveninthe standard published by the British Effluent &WaterAssociation6.



APPENDIX D 69

Calculation ofHydraulic Size:

1.Divide he %retainedon eachsuccessive sievebythesizeof the sieveaperture.2.Add

upthe

figureshus obtainedand divide

by100.

3.Obtain thereciprocal of heabovesum.4.Add10%tothereciprocal tocorrect he "retained" size tothe centre sizebetween adjacent sieves(seepreceding paragraph).

Example

Sieve Size RetainedonSieve Calculation

2.8 0.1 0.042.36 0.5 0.21

2.0 10.2 5.1

1.7 22.1 13.01.4 42.3 30.25

1.18 22.5 19.1

1.0 1.41.41

0.850 0.35 0.350.71 0.5 0.7

Divide totalcalculated by 100 =0.70Obtain reciprocal of0.70 = 1.43

Add10% =0.14HydraulicSize =1.57mm

Porosity depends on the shape of thegrains. A 1%change n porosity (e.g. from40% to41%)willproducea 9.5% increase inthebackwash fluidization threshold. The temperature of thewater affectsthe requiredbackwash rate as theviscosity ji) s reduced byhalf from 1.781 at0°C to 0.798 at30°C. Thisrangeofwater

temperatures wouldbe extremely unlikely to be encountered in Ireland, but the temperature effect mightcombine witha oss offines fromthemedia, over time, to reduce the effectiveness of thebackwashing incleaning themedia horoughly.

Where flow velocities are abovethe laminarregion, the Carman modification of theKozeny equation isappropriate:

R= 5Re1+O.4Re°'

iV2

where

('11 V't1_e)PV2 and Re

pV2 e3 S(1—e)ji

where R1 is theshearstress at he surface of hegrains

Re is theparticle Reynolds Number

These equations, although presenting problems for manual computation, are readily managed in simplecomputer programs.



70 FILTRATION



REFERENCESANDREADING LIST 71

REFERENCES AND READING LIST

1. US EPA(1994) WaterTreatmentPlantOperationVolume1:AFieldStudy Training Program.ThirdEdition.

2. Huisman, L. & Wood,W.E.(1974)SlowSandFiltration, Geneva, WorldHealthOrganisation.

3. Smethurst, George. (1990)Basic Water Treatment. Second Edition,ThomasTelford Ltd.,London.

4. American Waterworks Association. Water QualityandTreatment AHandbookofCommunityWater

Supplies. FourthEdition. A.W.W.A., 6666 W.QuincyAve.,Denver,CO 80235.

5. Stevenson, D.G.,(1994)TheSpecification ofFilteringMaterials orRapidGravityFiltration inJournal

of nstitution ofWater andEnvironmentalManagement, Volume8,October 1994.

6. British Effluent and WaterAssociation (1993)Standard or theSpecflcation,ApprovalandTesting ofGranularFilteringMaterials.BEWA:P.18.93

7. Twort, A.C., Law, F. M., Crowley,F.W.,andRatanayaka, D.D.,(1994) Water Supply. 4thEdition.

EdwardArnold, London.

8. EuropeanCommunities(QualityofWater Intended orHumanConsumption)Regulations 1988. (S.!.

No 84of1988) Stationary Office, Dublin.

9. Safety,HealthandWelfareatWork Act,1989 (No.7of1989). Stationery Office,Dublin.

10. Safety,Health and WelfareatWork(GeneralApplication )Regulations,1993 (S.I. No. 44of1993)

Stationery Office, Dublin.

11.CodeofPractice ortheSafety,Health and WelfareatWork( ChemicalAgents )Regulations,1994 (S.!.

No. 445of1994). Stationery Office, Dublin.

12.Peavey,Howard, Rowe,Donald andTchobanoglous (1986) Environmental Engineering McGraw -

Hill,NewYork.

13. Engineering School, UCD(1972). Course Notes- Water QualityManagement.Department ofCivil

Engineering, University College,Dublin.

14.Flanagan,P.J., (1992)ParametersofWater Quality - Interpretationand Standards.SecondEdition.

Environmental Research Unit, Dublin.

15.Mc Evatt(1973) MetricGuide 1. ThirdEdition. An ForasForbartha, Dublin.

16.Department of heEnvironment (1990)Cryptosporidium inWaterSupplies: ReportofheGroupof

Experts. Chairman: SirJohnBadenoch. HMSO,London.

17.Department oftheEnvironment (1994)ProceedingsofWorkshopon Cryptosporidium inWaterSupplies.

Editedby A. Dawson andALloyd. HMSO, London.



72 FILTRATION



USER COMMENT FORM 73

USER COMMENT FORM

NOTE: Completed omments tobeforwarded to: The EnvironmentalManagement and Planning Division,Environmental Protection Agency, Ardcavan, Wexford

DocumentTitle: FiltrationManual

CONTENTS:

STYLE:

INFORMATION:

SUGGESTIONS FORFUTURE EDITIONS:

NAME ORGANISATION

ADDRESS

DATE PHONE FAX



74 FILTRATION



EPA PUBLICATIONS 75

Recent Environmental Protection AgencyPublications

LandfillManual Landfill Monitoring. (1995). £15

Landfill Manual InvestigationsforLandfills. (1995). £15

Waste Water TreatmentManual PreliminaryTreatment. (1995). £15

The EPAproducesawidevariety ofreports, guidancenotes,manuals andcodes of

practicerelevant o protection of he natural environment. A full list ofpublications is

available on request. (See below).

Order Form

Ifyouwish toorder further copiesof hispublicationorany of hose listedabove,simply

complete this order form and returnit toEPA, Publications'Sales, St. Martin'sHouse,WaterlooRd., Dublin4,enclosing acheque orpostal order for the pricequoted.

Publication Title No. ofCopies

_______________ Li

_______ U

Tickhere to receive a ull listofEPApublications(no charge) Li

Name Address



76 FILTRATION



Environmental Protection AgencyESTABLISHED

The EnvironmentalProtectionAgencyAct,1992,was enactedon 23

April, 1992 and under

this legislation the

Agency wasformallyestablished on 26 July,1993.

RESPONSIBILITIES

TheAgencyhasa wide

range ofstatutory duties

andpowersundertheAct.

The mainresponsibilitiesof theAgency include the

following:

• the licensing and

regulation oflarge/complexindustrial and other

processes with

significant pollutingpotential, on thebasis

of integrated pollutioncontrol (IPC) and the

application ofbestavailable technologiesforthis

purpose;• the monitoringofenviromnental quality,

including theestablishment ofdatabases towhich he

public willhaveand the

through, for example,theencouragement ofthe use ofenvironmental audits,theestablishmentofaneco-labellingscheme, thesetting ofenvironmental quality

objectives and the

issuing ofcodes ofpractice on matters

affecting the

environment;• thepromotion and co-

ordination ofenvironmental

research; and

generally overseeingtheperformance by

local authoritiesof

theirstatutoryenvironmental

protection functions.

STATUS

TheAgency s an

independent public body.Its sponsor inGovernment is the

Department of heEnvironment.

Independence is assured

through theselection

procedures forthe

DirectorGeneral and

Directors and the

freedom, asprovided in

ORGANISATIONThe Agency'sheadquarters are

located in Wexford and it

operates fiveregional

inspectorates, located n

Dublin, Cork,Kilkenny,Castlebar andMonaghan.

MANAGEMENTThe Agency is managed

by afull-time Executive

Boardconsisting ofa

DirectorGeneral and fourDirectors. TheExecutive

Board is appointed by the

Government followingdetailed procedures laid

downintheAct.

ADVISORY

COMMITtEE

The Agency is assistedbyanAdvisory Committee

of welve

members.Themembers

areappointed by the

Minister for the

Environment and are

selected mainly fromthose nominated byorganisationswith an

interest inenvironmental

anddevelopmentalmatters. TheCommittee

has been a wide

epa water treatment manual filtration 1

Documents