3 4 1 . 2

7 0 H A

Handbook of

PUBLISHED BY PUBLIC WORKS PUBLICATIONS

UNDERDRAWS OF ENDURING

VITRIFIED CLAY

Foreword

LIBRARYInternational Reference Certtltfor Comrr-'nltv V-'^r Suppfy

TRICKLING FILTERS were in use in 3,506 municipal' wastewater treatment plants in the United States in

1962, as reported in a statistical summary issued by theU. S. Public Health Service in 1964, the latest compila-tion available. Of these plants, 2,135 employed standardand 1,371, high rate filters. The total population servedby these facilities at that time was 23,006,271. While asingle stage filter was used in many cities as the solesecondary treatment process, in over 2,600 plants theywere employed in multiple stages or in conjunction withother biological processes as intermediate or final steps.With the increasing need for advanced waste treatment,research is being directed toward taking advantage ofthe potential versatility of niters in developing proc-esses that will exceed ordinary secondary treatment inperformance.

The growth in popularity of trickling filters is illus-trated by periodic census studies. In 1940, there were1,486 plants in use, only a few of which were high rate.In 1945, 1,581 plants were reported, of which 122 werehigh rate; most of the construction in this war periodwas for military purposes, municipal construction ofwastewater treatment plants being severely curtailed.The 1957 report showed the existence of 1,870 standardrate and 812 high rate niters. From 1957 to 1962, therewas an increase of 265 standard rate and 559 high ratefilters. None of these figures includes institutional or in-dustrial data. For example, most secondary treatmentdevices for service areas on major turnpikes employthis method.

The first use of trickling filters of record was in 1889,when the Lawrence Experiment Station of the Massa-chusetts State Board of Health put two filters into ser-vice using gravel media. Their use led to an understand-ing of one of the essential features of this method oftreatment—the slow movement of the liquid in filmsover the surface of the stones while in contact with air.

Even though these first filters were developed in theUnited States, English engineers developed the methodand the means of applying the wastewater as a sprayto the surface of the filter; and, highly important, the

use of "false bottoms" to support the media and, at thesame time, to eliminate clogging at the filter floor level.

The first trickling filter installation for a city in thiscountry was recommended for Atlanta, Ga., in 1903, butit was not constructed until 1910. However, a small ex-perimental unit had been installed in 1901 in Madison,Wise., and from this and from test and pilot plants val-uable design information was obtained. The plant forColumbus, Ohio, put into operation in 1908, provided10 acres of filters. Also, in 1908, a one-acre plant wasbuilt for Reading, Pa., and another small plant forWashington, Pa. Trickling filters at Mt. Vernon, N. Y.,were put into operation in 1910, and the large plant atBaltimore, Md., with 14 acres of filters, was completedin 1911. Other installations were made for the munici-palities of York, Allentown and Meadville, Pa., Rome,N. Y., and North Plainfield, N. J.

This growth of use of trickling filters produced oper-ating data on which many improvements have beenbased. One of the early observations related to thenecessity of having a "false floor" to prevent the sus-pended matter in the wastewater from accumulatingon the floor of the filter and interfering with drainageand ventilation. The use of an inverted half-round pipewas probably most common. However, at Baltimore, aseries of channels were designed with perforated tilecovers.

During the past twenty-five years, the use of clayfilter bottom blocks or underdrains has grown and itis a very rare installation where such blocks are notused. The manufacturers making filter underdrains,united in the Trickling Filter Floor Institute, engage inresearch, improvement of manufacture, developmentof specifications, and advertising to contribute to themore effective use of trickling filters.

This booklet, the fifth edition of the "Handbook ofTrickling Filter Design," is a product of the TricklingFilter Floor Institute in cooperation with PUBLIC WORKSMagazine. Its purpose is to furnish technical data andprovide information whereby trickling filters will do astill better and more economical job in waste treatment.

Copyright 1970, Public Works Journal Corporation, 200 South Broad Street, Ridgewood, New Jersey

Wastewater Treatment with Trickling Filters

CONVENTIONAL was t ewa te rtreatment procedures involve

solids removal and stabilization ofthe organic compounds dissolved inthe liquid fraction. The latter stepis usually accomplished by biologi-cal oxidation, in which a culture ofmicroorganisms "feeds" upon theorganics. The devices used for bi-ological oxidation establish aerobicconditions to develop the cultureand keep it active for maximum ef-fectiveness. This means providingan environment of oxygen or air.Consequently, as a biological treat-ment medium, a trickling filter isessentially an aeration device andprovides very little mechanical re-moval of suspended solids as theterm "filter" might imply.

The adaptability of trickling fil-ters to varying organic and hy-draulic loadings has resulted intheir wide usage. Many variationsin design have been developed ca-pable of producing almost any de-sired end results. All of these varia-tions, however, appear to retain theability to operate in cold weatheror in hot, or with overloads, withonly reasonably skilled supervision.

Certain conditions are necessaryfor most effective operation. Theseinclude: i ) Proper pretreatment ofthe waste to remove solids, usuallyattained by passage through a sedi-mentation tank; 2) reasonably uni-form application of the settled wasteto the surface of the filter, as by arotary distributor or spray nozzles;3) use of properly sized filter me-dia, free from fines and substantiallyuniform in grading; 4) provision ofunderdrainage facilities for theprompt and complete removal of thematerial passed through the filterbed; and 5) assurance of adequateventilation so that each unit of thefilter receives the air needed for theoxidation process.

Filters are classified according tothe organic loading for which theyare designed, as high rate or stand-ard (or low rate); and by depth, asdeep or shallow.

The amount of organic loading, inpounds of BOD applied per cubicyard, 1000 cu. ft. or acre-foot ofmedia determines whether the filteris to be termed high rate or stand-ard. Though there is no definite

Table 1—Abbreviations

5-day Biochemical Oxygen De-

mand BODSuspended Solids SSGallons per minute gpmGallons per day gpdMillion gallons per day mgdGallons per person per day . • • gpcdAcre-foot = 43,560 cu. ft atMillion gallons per acre per day mgadMilligrams per liter mg/L

standard, a generally used loadingfor a standard filter approximates,as a maximum, 600 lbs. of BOD peracre-foot, 13.7 lbs. per 1000 cu. ft.or 0.375 lb. per cu. yd. Though manystandard filters are carrying heavierloadings, they were not generallydesigned to do so initially. TheManual of Practice, Water Pollu-tion Control Federation lists thestandard rate filter range as 5 to 25lbs. per 1,000 cu. ft. High rate filtersare usually designed for an appliedloading, based on settled sewagestrength, of 2,000 to 3,000 lbs. peracre-foot or 1% to 1% pounds ofBOD per cu. yd. However, theWPCF Manual of Practice gives therange as 25 to 300 lbs. per 1,000 cu.ft., which is 1,090 to 13,100 lbs. peracre-foot.

Just as in the case of organic

loading, there is no clear line todistinguish deep from shallow fil-ters. A generally used standard inrespect to depth considers a filterin which the media is less than 4.5ft. in depth as a shallow filter; andone with media more than 4.5 feetin depth as a deep filter. The depthis the average distance from the topof the underdrain floor to the topsurface of the filter media.

Definitions and AbbreviationsFor the purpose of this text, the

abbreviations shown in Table 1 willbe used. Definitions of other termsused are as follows:

The organic load on a filter is theBOD content in pounds that is ap-plied to the filter. It is usually theBOD content of the raw sewage lessthe BOD that is removed by the pre-treatment devices such as screensand clarifiers. For instance, if thesewage flow is 1 million gallons perday and the BOD of the raw sewageis 240 mg/L, the total BOD will be2,000 pounds. If 30 percent is re-moved by primary treatment de-vices, the organic load applied tothe filter will be 2,000 x 0.70 = 1,400pounds BOD.

The volume or hydraulic loadingon a filter is the amount of sewageor waste applied to it, computed on

Courtesy W. S. Dickey Clay Mfg. Co.

• ALL elements of a trickling filter are shown in this cutaway view. A is the influentline; B, the drainage floor; C, a vent; D, the media; E, distributor feed column; andF, the rotary distributor. The various functions are described in detail in the text.

the basis of millions of gallons peracre of filter surface per day. If thesewage flow is 1 mgd and this is ap-plied to a filter which has a surfacearea of 0.10 acre, the volume loadingwill be 10 million gallons per acreper day.

A single-stage filter is one inwhich treatment is accomplished byone or more passes of the sewagethrough the filter. This term is usu-ally restricted to high rate filtersusing recirculation. If there are twoor more filters in a single-stageplant, these filters are operated inparallel. In a two-stage filter sys-tem there are two filters in series,with the effluent from the first orprimary filter being passed throughthe secondary filter. Recirculationis usually employed on both, niters.Large two-stage treatment plantsmay have four, six or more niters,each pair of them working as a team.

Recirculation is the return of aportion of the sewage or wastewhich has already passed throughthe filter, and sometimes also hasbeen settled, for another passagethrough the filter. This return maybe directly to the filter or it may beto the inlet of the primary settlingtank.

The primary clarifier, primarytank or primary settling tank is thesedimentation device ahead of thefilter which pretreats and preparesthe waste for application to the fil-ters. An intermediate clarifier ortank is the sedimentation device be-tween any two filters. A secondaryclarifier or secondary or final set-tling tank is the sedimentation de-vice following the final filter. Thisusually provides the final treatmentfor the waste except for chlorina-tion.

In considering trickling filters, thefilter or niters and the followingclarifier should always be consid-ered as a unit in both design andoperation. There is no place in sew-age treatment for a trickling filterwithout a final clarifier in which tosettle the filter effluent. (This doesnot apply to the first stage filter ofa two-stage filter system.)

Other Units of Express/onHydraulic load, including recircu-

lated flow if used, is in gallons offlow per square foot of surface areaof the filter. The standard rate filterrange will be 25 to 100; the highrate range will be 200 to 1,000. Therate in million gallons per acre perday multiplied by 23 will give gal-lons per square foot per day. Thusa rate of 20 mgad equals 460 gsfd.

The rate in pounds of BOD peracre foot per day multiplied by 0.023

Courtesy Koppers Co.TYPICAL distributor in operation at Salisbury, Md. wastewater treatment plant.

equals the pounds of BOD per 1,000cu. ft. per day. Conversion frompounds of BOD per cubic yard perday to pounds of BOD per 1,000 cu.ft. per day is accomplished by multi-plying by 37. Thus 2,420 pounds ofBOD per acre foot, multiplied by0.023 represents 55.6 pounds per1,000 cu. ft., as does the equivalent1% lbs. per cu. yd. multiplied by 37.For high rate niters, the recirculatedflow is not considered in contribut-ing an additional BOD load.

Percentage of recirculation equalsrecirculated flow divided by wasteflow multiplied by 100. The standardrate filter range is 0 to 50 and thehigh rate range is 25 to 300.

Comparison of Filter TypesThere is no essential difference in

construction between the high rateand the standard or low rate filter.The same media, the same under-drains and the same ventilation fa-cilities are provided in either case.The distributors are essentially thesame, except that those used forhigh rate filters must be designedfor larger volume rates of applica-tion. The potential BOD removalappears to exist in every filter; therate and method of application ofthe sewage to the filter determineshow and to what extent this poten-tial is utilized. The high rate filter,because it is loaded much moreheavily—the BOD loading may bethree to six times as great as thaton the standard rate filter-—is cor-respondingly smaller and usuallycosts less to build.

There are differences in the qual-ity of the effluents from high rateand standard niters. The latter pro-duce a higher degree of nitrification,so that strengths of effluents fromthe two types of niters, stated inmilligrams per liter of BOD, cannotbe directly compared. This is espe-

cially true of single-stage high ratefilters; two-stage high rate nitersmay produce some degree of nitri-fication. However, local conditionsusually govern in determining thedegree of treatment needed and theactual reduction in the organic loaddischarged to the stream is a mostimportant factor.

If local conditions dictate the se-lection of a plant to produce an ef-fluent low in BOD, well nitrifiedand highly stable, full considerationshould be given to a deep filter anda low rate of loading. If a less com-pletely stabilized effluent is permis-sible, with a slightly higher BODcontent, a single-stage high rate fil-ter may be selected because it isusually lower in initial cost. If thewaste is very strong and a good ef-fluent is necessary, a two-stage filtermay be preferable, since this com-bines many of the cost-saving fea-tures of the single-stage high ratefilter with an effluent that may ap-proach that of the standard filter inBOD content.

Pre-TreatmenfStandard practice is to provide

pre-treatment ahead of filters bymeans of settling tanks, either roundor rectangular, equipped with me-chanical apparatus for removing thesludge and preceded by a comminu-tor and properly cleaned screens.

Standard practice usually pro-vides for a detention period of abouttwo hours in the primary settlingtank for design flow plus recircula-tion, in the case of a high rate filter.However, there is considerable vari-ation in requirements among theseveral states and the applicableregulations should be determinedand followed.

Surface overflow rates, tankdepths and other factors are alsosubject to state regulations and

'-•s"^V **i %

Courtesy P.F.T. Div. Rex Cliainbelt, Inc.AERIAL view of Winston-Salem's treatment plant and four 200 ft. diam. filters.

these may vary materially amongthe states.

In the treatment of cannery wastesand creamery products, some engi-neers use a fine screen and a smallequalizing tank for pre-treatmentinstead of the usual primary tank.This is based on studies and obser-vations which indicate the "micro-bial forest" on the filter media cantreat any material not large enoughto interfere with the application ofthe sewage to the filter.

Intermediate andPost-Treatment

Intermediate clarifiers have beenused in a relatively small numberof plants. They are of most valuewhere the sewage or industrialwaste is very strong. Detentionperiods are usually somewhat lessthan for primary clarifiers.

The secondary clarifier is the keyfactor in the design of filters for bestoperation. It is not the function ofthe filter to remove suspended mat-ter, but to change this to a relativelystable form. This material, and thespent gelatinous film which sloughsoff the filter media, continuously inthe high rate filter and from time totime in the standard filter, must beremoved by the clarifier.

As with primary tanks, the deten-tion period, overflow rate, weiroverflow rate and tank depth of thesecondary clarifier are subject tostate health department regulations.In general, detention periods of 2 to3 hours should be provided for de-sign flow plus recirculation, if thesecondary clarifier is involved in therecirculation process. Where recir-culation is used, the necessary vol-ume may be taken from a launderinstalled in the secondary tank.

Effects of TemperatureTrickling filters operate better

and produce better effluents inwarm, weather periods. However,though they are affected in some

respects by low temperatures, ex-treme cold is necessary to createany serious adverse effect. In theNorth Central States, it is fairlystandard practice to provide covers,either of prestressed concrete or ofwood, for filter units, for experiencehas shown that operating results areimproved and troubles from icingare eliminated.

Cold weather effects may be oftwo kinds: 1) the surface of thefilter may become iced; and 2) theefficiency of the filter in respect toBOD removal may be impaired.

In ice formation, to which snow-fall may contribute, wind appearsto be an important factor as it drivesspray to the downwind side wheredeposits may build up. However,difficulties are restricted to the reallycold areas. No such troubles werereported in the operation of un-covered filters at more than 300Army posts during World War II,though some of these were locatedin low temperature areas.

It is good practice to providebleed valves in the distributor sup-port posts of rotary distributors andin the distribution system, of spraynozzle installations so as to preventfreezing and consequent damage ifthe flow of sewage stops or becomesso small that the dosing siphonswork only at long intervals. Insome plants, a small pipe is tappedinto the feed pipe at the distributorbase and is left open all winter.

To determine the effect of tem-perature on the efficiency of BODremoval by low rate trickling filters,the records of nine military sewagetreatment plants were studied by aSub-Committee on Waste Disposalof the National Research Council.These were all standard rate plants.Results obtained in July, Augustand September of 1943 were com-pared with results in December,1943, and January and February,1944. With adjustment for differ-ences in loading during the two

periods, there was a small but sig-nificant drop in efficiency duringthe colder months during which theair temperature averaged 38.1°Flower than the summer months.The subcommittee reports that:"With military sewage of the usualconcentration, this difference inefficiency results in effluents with aBOD about 20 percent larger in win-ter than in summer."

On this basis, it was estimatedthat a filter should be about 43 percent larger if it is to operate as wellat an average temperature of around40°F as the normal filter will oper-ate during the warm summermonths. The army designs of WorldWar II specified that filters in thenorthern states should be 25 per-cent larger than those in southernareas.

Under the best of conditions, sometime is required to build up thebiologically active film necessary onthe filter media. The time requireddepends on the temperature andprobably on a variety of local con-ditions. Usually 3 to 4 weeks arerequired. If the plant is started latein the fall, when air and sewagetemperatures are low, an evenlonger period will be required be-fore adequate treatment results canbe expected. A change from highrate to low rate operation, or viceversa, will probably require 4 to 6weeks to develop the proper bio-logical growth. Changes in rate ofloading are also likely to requireconsiderable periods before resultstruly reflect the change.

ApplicabilityTrickling filters are applicable to

weak or strong domestic waste-waters and to many types of indus-trial wastes. By utilizing recircula-tion, whereby the incoming sewageis reduced in strength, even exceed-ingly strong wastes can be treatedsatisfactorily and economically. Re-circulation may be applied to eitherhigh rate or standard filters. Highrate and standard filters may becombined in a single treatment plant,a procedure especially adaptable tothe production of an excellent efflu-ent. In such cases, the high ratefilter is used as the primary filter toremove as much as possible of theorganic load; and the effluent fromthe high rate filter is applied to thestandard filter. Final treatment, fol-lowing the high rate filter, may alsobe provided by an activated sludgeplant. In either case, the high ratefilter markedly reduces the organicload on the secondary unit anddelivers to it a waste of uniformstrength.

Media, Drainage and Ventilation

THE five essential elements in atrickling filter are the base sup-

port; the underdrain blocks whichcarry off effluent and provide venti-lation; the containing wall; themedia to support the growth ofmicroorganisms; and the distributorfor applying settled wastewater tothe filter.

Filter Floor and Center ChannelsThe floor should provide a firm

and uniform base on which to buildthe remainder of the filter. It shouldbe of concrete, poured on a well-compacted earth base. Since the loadcarried by the floor is not great,a concrete thickness of 4 to 6 inchesis ample. Light reinforcement is de-sirable for not only must the floorbe uniform initially to facilitategood drainages, but it must retainthat uniformity for the life of theplant The floor may be poured inone run, or there may be joints. Ifused, joints should run in the di-rection of the floor slope, and rein-forcing dowels should be carriedthrough the joints. Floor slope to-ward the drain channels should benot less than 2 in., for any semi-diameter, and is usually about 0.3ft. per 100 ft. There are various de-signs, but the most common practiceinvolves bisecting the filter floorwith a center channel toward whichthe two halves of the filter slope.

Precise analysis of the flow in thecenter channel is difficult becauseof the interference with the flow dueto the discharges from the manyunderdrains entering at right anglesto the main flow. Generally, satisfac-tory capacity will be provided by across-section and a slope that will,as determined by the usual flowformulas or charts, carry the maxi-mum possible flow that may occurwith the liquid surface in the chan-nel slightly below the bottom of theunderdrain channels.

Design of center channels differssomewhat for shallow high ratefilters and for standard rate deepfilters. In a shallow filter, the feedpipe for the distributor must gener-ally pass under the filter floor andup through the distributor support.In the standard rate deep filter, thefeed pipe is usually passed throughthe body of the media. This is largely

because of the relative sizes requiredfor the feed. In a standard ratefilter 100 ft. in diameter with a rateof surface application of 2 mgad, theflow to the filter will average 0.36mgd, which can be carried by an8-inch pipe (or even a 6-in.). In ahigh rate filter, of the same size,using a surface rate of application of20 mgad, the feed pipe will have tocarry an average of 3.6 mgd and a16-in. or 18-in. pipe is required. Ifit is necessary to conserve head, aneven larger pipe may be used. Thelarger pipe would block off an un-desirably large portion of the filterif the pipe were carried through themedia instead of beneath the floor.

The center channel in the standardrate filter can be smaller because itcarries a lesser volume of waste-water. As a result, it is often passedthrough the distributor support,though sometimes split or carriedaround it. In the high rate filter,the center channel must be largerand consequently cannot be carriedthrough the distributor support. Itis common practice to offset thechannel from the center enough toclear the support column. A fewlarge filters have been built with

two channels and fewer have usedperipheral channels, the floor slop-ing to the circumference.

Center channels may be rectangu-lar in cross section or the bottommay be semi-circular. In either case,channel covers are necessary to sup-port the superimposed media. Thesecovers will be described under alater heading.

To facilitate cleaning and inspec-tion, the center channel may be ex-tended through the filter wall atboth ends, terminating at the upperend in a small open box coveredwith a grating. The box should be ofsufficient size to permit entrance.

The value of ample ventilation isproved by the excellent operatingrecords of many filters where vitri-fied clay underdrain blocks havebeen used. Some engineers add ver-tical vent pipes around the innercircumference of the walls at theperiphery of the filter and othersalso extend the drainage ductsthrough the filter walls. This is ad-vantageous if provision for flushingof the channels is desired, as whenindustrial wastes are treated.

The absolute necessity for ade-quate drainage and ventilation in a

DICKEY high-rate blocks were used in the trickling filter at Springdale, Ark.

5

• STANDARD rate filter underdrainblocks made by the member firms ofthe Trickling Filter Floor Institute. Allare made to conform to ASTM specifi-cations given in this Handbook and areof corrosion resistant vitrified clay.

trickling filter should be recognized.Complete systems of vitrified clayblocks are laid directly on the filterfloor. These units are not affectedby the action of wastewater and thusare as long lived as the filter itself.The units of these systems arequadrangular prisms having open-ings in the upper face which collectthe wastewater passing through thefilter. Ample conduit capacity isprovided for rapidly discharging theeffluent, and for admitting air intothe filter for ventilation and aera-tion. The blocks are laid on a thinlayer of grout or of dry mortar,rapidly and precisely, so that thedrainage channels are automaticallyaligned. The strength of the blocksis greatly in excess of that requiredunder any condition of service.ASTM specifications for these blocksare given on pages 9 and 10.

Underdrain BlocksUnderdrain blocks differ in de-

sign primarily according to use instandard or high rate filters. Thelatter are deeper and have channelsof greater capacity for carrying offthe filter effluent. Actually, thestandard blocks have the ability tohandle any application rate up to 50mgad on a 200-ft. diameter filter,but the high rate blocks provideadditional space for air penetrationinto the filter and are preferred bysome engineers, especially for indus-trial waste applications.

The blocks described below com-ply with all requirements of ASTMSpecification C159. All blocks are ofvitrified clay which experience hasshown to be resistant to all elementsin either domestic or industrialwastewater; to have excellent flowcharacteristics because of the smoothsurface; and to resist attachment ofgrowths which might tend to clogthe ventilating openings or flowchannels.

Standard BlocksDICKEY, made by the W. S.

Dickey Clay Mfg. Co., are 18 in.long, 11% in. wide and 4Vi in. deep,with 18 top openings.

POMONA, made by Pomona PipeProducts, are 13% in. long, 12 in.wide and 5% in. deep.

TRANSLOT, made by Can-TexIndustries, Inc., a division of HarscoCorp., are 11% in. wide, 18 in. longand 4LVA in. deep.

High Rate BlocksTRANSLOT "Hi-Rate" blocks

are 18 in. long, 11% in. wide and7]/i in. deep. They have three flowchannels and five transverse topslots. These are made by Can-TexIndustries, Inc., a division of HarscoCorp.

POMONA "Rapid-Flo" high rateblocks are 13 V2 in. long, 11% in. wideand 724 in. deep, with three slotopenings 1% in. wide—made byPomona Pipe Products.

DICKEY, made by W. S. DickeyClay Mfg. Co. are 18 in. long, 11% in.wide and 7M> in. deep with two flowchannels and six transverse top slots.

It is important to insure in thespecifications the use of good ma-terials and installation methods. Aspecification which will assure goodunderdrainage construction follows(the latest ASTM standard revisionnumber should be used).

Underdrains—The contractor willfurnish and install vitrified clay un-derdrains as shown on the plans.These shall be laid in a dry mortarbed, on the floor of the filter before

H HIGH-RATE filter blocks made bymembers of the TFFI. These high-rateblocks feature deeper channels by twoinches or more than those used in thestandard-rate filters. In all other re-spects, the blocks are quite similar.

the stone is placed. Underdrainsmust comply with standard specifi-cations ASTM C159, and shall beequal and similar to those manu-factured by members of the TFFI.The mortar shall consist of sand andcement, 1 cement to 6 sand. Afterthe underdrains are laid and beforethe stone is placed, the dry mortarbed shall be wetted by sprinkling.Blocks must be laid in true align-ment, with cross joints staggeredin longitudinal rows at right anglesto the center drains.

Angles, Cover Blocks,Ventilation and Laying

For use in circular filters, blockscut to 22y2°, 30°, 45° and 67M>° areavailable to match approximatelythe curvature of the filter walls, andcross-sections at the pier, effluentchannel and perimeter wall.

Also short sections of standardblocks are provided to fill out runsand eliminate the necessity for cut-ting blocks on the job. Vent blocksto support pipe riser vents to thesurface of the filters are furnished;and also wall reducers for carryingventilation ducts through the filterwalls for use where flushing oradded ventilation is deemed desir-able.

In ordering blocks for a filter, itis necessary only to provide themanufacturer with a plan of thefilter, showing essential dimensions.Shipments of blocks are then madewith the correct number of straightblocks, angle blocks, short sections,specials and cover blocks.

Guidance in laying is also pro-vided by the manufacturers to in-sure good practice with the mini-mum labor requirement.

Standard cover blocks for the cen-ter drainage channel are normallylimited to 30 in. in length. At least3 in. of bearing should be allowedat either end. If it is necessary toprovide a center channel wider than24 in., either a central supportshould be used or a special channelcover designed.

Filter Design DefaltsFilter depths range all of the way

from 3 ft to 10 ft. Recently experi-mental work has been carried onwith filters up to 24 ft. deep; andsome deeper test units have beenbuilt. Low rate filters, however, areusually designed with minimumdepths of 5 to 6 ft., and with a maxi-mum of 9 to 10 ft.; most are about6 ft deep, which depth is more orless standard. High rate filters varyquite widely. The Biofilter is de-signed with depths all the way from3 to 6 ft., but 4 to 5 ft. is probablymost used. The Aerofilter is nor-mally not less than 6 ft. deep andmay be deeper. The Accelo filteris normally 6 ft. The depth functiondepends much on local topographyand on cost factors. No matter whatthe depth or what the volume load-ing, the underdrains previously de-scribed have ample load, drainageand ventilation capacity. That is,they are strong enough to carry theweight of the deepest filter, with avery wide margin of safety. Thedrainage channels are sufficientlylarge to handle properly the flowsfrom, a filter 200 ft in diameteroperating at a volumetric rate of50 million gallons per acre per day.The top openings are ample in sizeto provide the necessary air to in-sure good operating results.

Much experimental work has beendone in attempts to improve filter

operation by blowing air throughthe media; even, oxygen has beenused. No results have been obtainedwhich indicate that added ventila-tion, over and above that providedby the standard vitrified clay under-drains, is helpful in better opera-tion.

Wall ConstructionMost walls for filters have been

of reinforced concrete. When used,such walls are reinforced circum-ferentially, using the hoop formulafor design, with some vertical re-inforcing extending into the foot-ing. Walls of this type are usually8 to 12 in. thick.

It has been proposed to constructfilters without walls, allowing thefilter media to seek its angle ofrepose. It will generally be foundthat such construction is not eco-nomical. The additional stone re-quired for the peripheral unusedfilter media, plus the cost of ex-tending the filter floor out to the toeof the slope of the stone, will usu-ally exceed the cost of constructingwalls to retain the stone. Moreover,experience has shown that this mar-ginal area of stone, which is partlyor occasionally wetted is liable tobe a producer of filter flies.

Filter MediaCrushed stone is the most widely

used media for trickling filters.Crushed slag is also used, especiallyin areas where it has a price ad-vantage over stone. Rounded gravelis sometimes used but some reportsindicate it does not give as goodresults as do the angular aggregatesresulting from crushing. However,in other plants well-graded gravelhas produced excellent effluents,although an approximately cubicalshape appears preferable. Federalspecifications SS-S-448 r e q u i r e

media "as nearly equidimensionalas possible. Pieces that are elon-gated or flat, or both, shall not ex-ceed 10 percent by weight. Theseshapes are defined as having lengthexceeding width, and width exceed-ing thickness in ratio of more than1.5 to 1.0, dimensions being maxi-mum for the size of pieces underconsideration."

The media should be free of fineparticles, grease and oil and shouldbe properly screened and washed toremove dust and dirt before place-ment in the niters. Even with wash-ing, enough sand and fine stoneparticles may remain to cause trou-ble when filter operation is begun.Unless precautions are taken toprevent it, fine inorganic materialwill be carried to the secondaryclarifier and thence to the digester,where it may cause further oper-ating difficulties. In extreme cases,the mixture of sludge and dust mayclog the sludge lines. It is goodpractice for the designer to makeprovision, either temporary or per-manent, to permit the filter effluentto be bypassed around the secondaryclarifier until after a sufficientlylong operating period to assure thatall fine particles have been removed.Slag media should be free from allfree iron.

Media SpecificationsThere are differences of opinion

in regard to stone and slag mediasize. The modem trend is towarduse of larger aggregates—2V2-inchminimum. Uniformity is also impor-tant, and particles smaller than thelowest limit should not exceed 10percent by weight of the total.

Federal Specifications SS-S-448state that the filter media shall beas uniform in size as possible and,when tested with standard labora-tory sieves having square openings,

POMONA Rapid-Flo floor block in a 93-ft. dia. high rate filter at Lynehburg, Va.

Table 1—Federal Specifications for Filter Media Grading Sizes(Specification SS-S-448, Int. Amendment-1)

Nominal Size(square

ODeninps^ ' f ^ * • • P • i j^«u

in inches)

3y2 to 2V23V2 to 2

3 to 23 to IV2

21/2 to IV2

4 in.

100100100100100

Amount Passing Sieve (square openings) percent

3'/2 in.

90-10090-100

100100100

3 in.

90-10090-100

100

2V2 in.

0-10

90-100

2 in.

0-50-100-10

by weight

V/2 in. 1 in,

0-50-5

0-10 0-50-10 0-5

shall conform to the gradings shownin Table 1 of this text. Any of thefirst three sizes given are preferableto the smaller sizes, for domesticwastewater.

Commercial aggregate specifica-tions are now generally based onthe use of screens having squareopenings. Federal SS-S-448 hasbeen cited. ASTM specificationsC88, C136 and D75 apply to aggre-gates and include grading, sound-ness, wear and other factors.

In recent years, other materialshave been developed by certainmanufacturers. Among these areplastic sheets (polystyrene or poly-vinyl chloride) and redwood slats.The plastic sheets are assembled in"honeycomb" modules small andlight enough to be placed by hand.Corrugated or alternate corrugatedand flat sheets are used. The mod-ules have an effective surface areaof 27 sq. ft. per cu. ft and have

greater than 90 percent void space.The weight is 2 to 3 lbs per cu. ft.Advantages are light weight, ease ofplacement and superior ventilationqualities. These permit the con-struction of deep filters—up to 40 ftPVC modules are made by DowChemical Co., B. F. Goodrich Indus-trial Products Co. and Ethyl Corp.

The redwood slats are also avail-able in module form. These are roughsawed to provide purchase for bio-logical growth. They are % in. tihick,V/z in. wide and either 3 5 ^ or 47V2in. long. Spacing is 0.69 in. betweenedges. They are furnished by Del-Pak Corp.

Placing the MediaNo generally acceptable method

of placing the media in the filterhas been developed which will in-sure that pockets of fines do notoccasionally occur in filters. If theseare deep in the filter, they may re-

• CHANNEL block made by Can-Tex Industries being placed in one of the filterswhich serves the Central Sewage System of the Trinity Valley Authority near Dallas.

duce filter capacity and effluentquality. However, shallow surfaceponding is not always due to suchfines and often does not affect theoperation of the filter. In view of theexcellent operating results obtainedfrom so many filters, it is probablethat usual methods of controllingfines are adequate. However, itshould not be forgotten.

The methods used in placing mediain the Fort Worth filters representvery good practice. The specifica-tions required that, in the placing ofthe filter media, breakage and seg-regation and the entrance of fineparticles into the filter "shall beprevented. Media shall be placed inthe filter using a belt conveyor, byboxes, or by other approved meth-ods, which method does not requiretraffic of any type over the top ofthe stone placed in the bed, whetheron the stone or on mats placed onthe stone." It was also requiredthat "immediately prior to finalplacing, the media shall be passedover a suitable screen with not lessthan 1-inch square openings."

DistributorsRotary distributors, which are

made for all sizes of filters from thesmallest to the largest are generallyused for distributing the sewageover the stone. Spray nozzles arealso available for rectangular instal-lations and special disc distributorsfor some types of small filters. Therotary distributor has the advantage—important in most plants—of re-quiring very little head for its opera-tion.

Rotary distributors are made bythe following firms: Ralph B. CarterCo.; Dorr-Oliver Inc.; Eimco Corp.;General Filter Co.; Infilco Div. ofFuller Co.; Water Pollution ControlDiv. of Keene Corp.; Lakeside Engi-neering Co.; Link-Belt Div. of FMCCorp.; Pacific Flush Tank Div. ofRex Chainbelt Inc.; Smith andLoveless; Walker Process Equip-ment; and Yeomans Bros. Nozzlesfor rectangular filters are furnishedby Pacific Flush Tank Div. Specialdisc distributors are made by Lake-side and Yeomans.

Any of these manufacturers willsupply the detailed drawings cover-ing the installation of equipment

SCOPE

1. These specifications cover vitrified clay filter block made from clay or shale or mixtures thereof.

TYPES

2. (a) Two types of filter block are covered:L—A oneMeGe filter, block suMa|ite. for-use' ki- constructing a -single-course trickling,.

^^^^^^^isi^^^Afiriri^^T ductjs iftspugh; ;$& ls»^B^sortion oi jhfe;

i | | | | • in the

fi

P. ft—A tw^'iSce filter Hock sysifoor, that prOvMes d^nctge and

p i l ' s or i£fi$r§ fbnss eosf^^©^. dr<filter bed sna for passage: ti{(Sir intoafta aefa#on grilles for passage of li<

..jsuitabje. fpr use in construotig^ a two-course trie&tion. VS%en placed or agsemiMd, a course of the

for conVe.y^noe of liquidstjpr bed. Ah assembly of the y^ 'er unijs formssfrom, and air into, the filteiiMg': ihedia.

(b) When in position of use both types of block shall completely cover the floor of the tricklingfilter bed.(c) The purchaser should specify whether block shall be type I-S (standard-rate), type I-H (high-rate), or type II.

COMPRESSIVE STRENGTH

3. (a) The average compressive strength of five full-size filter blocks shall be not less than 600 psi,based on gross area of the block, with the load applied in the same direction as in service. Noindividual block shall have a compressive strength less than 500 psi.(b) Both component units of type II systems, tested individually, shall conform to the compressivestrength requirement of Paragraph (a).

ABSORPTION

4. The average water absorption of five filter block, by 1-hr submersion in boiling water, shall notexceed 6 per cent of the dry weight of the block.

ACID RESISTANCE

5. Filter block shall be acceptable only when the percentage of acid-soluble matter for specimensrepresenting such block does not exceed 0.25 per cent by weight.

SHAPE

6. (a) Type I filter block shall be square or rectangular in plan, as laid. When placed in parallelrows on the subfloor of the filter bed, cells in a single course of type I filter block shall form con-tinuous drainage ducts, and the apertures or slots shall provide for drainage and aeration offilter media.(b) Units for the lower or drainage course and units for the upper or aeration course of type II filterblock shall be square or rectangular in plan, as laid. When placed in parallel rows on the sub-floor of the filter bed, the lower units shall form continuous drainage channels. Upper units, whenlaid in parallel rows on lower units, shall provide for drainage and aeration of filter media.

PERMISSIBLE VARIATIONS IN DIMENSIONS

7. (a) Length, width, or height of a type I filter block shall not vary by more than l/4 in. per ft. overor under the nominal dimension stated by the manufacturer.(b) Lower drainage channels and upper aeration grilles of type II block shall not vary by morethan % in. per ft. over or under the nominal dimension stated by the manufacturer.

APERTURES

8. (a) Apertures in the top of type I filter block shallbe not more than 1 '/2 in. in width, but may varyin length. Aperture area shall be not less than2.0 per cent of the top area of the filter blockwhen in position of use.(b) Apertures in the top, or aeration course oftype II filter block shall be not more Aon 1'/»in. in width, but may vary in length, .stfpar un-obstructed aperture area shall be not; MSs than30 per cent of the top area of the bl6;<*K whenin position of use.

Table I.—Cross-SectionalDrainage Areas

Minimum Effec/iV&A&e.a,Sq. in. per fool oPffldth

20

block, and

f in. in

SHELL AND WEB THICKNESSES

9. . (a -.fexferior w@Ss- (sheffe) of type Ishifiri be not less than S/16 in. in thiinffefet webs shall be riot less thanthickness.(b) Exterior walls (shells) and interior webs andstruts of type II filter block comprising the loweror drainage course shall not be less than 9/16in. in thickness. Exterior walls (shells) and longi-tudinal webs of type II filter block comprisingthe top or aeration course shall be not less thanl'/g m- in thickness. Stiffening webs for stiffen-ing the longitudinal webs, shall be not less than9/16 in. in thickness.

DRAINAGE CHANNELS

10. (a) Effective cross-sectional drainage area shallbe defined and measured as the sectional arealying below the lowest level of the apertureslots. Aggregate effective cross-sectional areaof drainage channels in a filter block shall benot less than shown in Table I.(b) The bottom of the drainage channels of anyfilter block shall be curved or narrowed in widthto form a trough in the bottom of each channelto accelerate flow of effluent and to minimizestranding of solids in such channels.(c) The height of drainage channels in any filterblock shall be not less than 2!/2 in., measuredfrom the lowest level of the aperture slots.(d) There shall be not less than two paralleldrainage channels per foot of width of a type Ifilter block.

WORKMANSHIP AND FINISH

11. (a) Filter block may be either glazed or un-glazed. If glazed filter block are specified, atleast 90 per cent of the inner and outer surfacesof the block shall be covered with glaze, except-ing the wire-cut ends where it may be entirelyabsent.

FOLD FLAP OUT

m(b) The lower bearing surface of filter blockshall not vary from a plane by more than 3/16in. per foot of the greatest length or width of theblock.(c) All filter block shall be well burned and sub-stantially free of laminations and fractures. Allblock shall be free of open cracks exceedingV/2 in. in length (such cracks shall not extendcompletely through a connecting portion of theweb or shell), and shall be free of other cracksexceeding 3 in. in length that do not extendthrough the webs or shells. All block shall befree of chips larger than 2 by 1 by 3/16 in. andof blisters whose diameters exceed 2 in. andwhich project more than !/8 in. above the nor-mal surface.

MARKING12. All filter block shall bear the name, initials, or

trade-mark of the manufacturer and shall bemarked to show compliance with ASTM Speci-fications C 159. These marks shall be indentedon the exterior of the block and shall be legible.

REJECTION

13. If test specimens fail to conform to requirementsfor compressive strength, absorption, or acidresistance, the manufacturer may sort the ship-ment, and new samples shall be selected bythe purchaser from the retained lot and testedat the expense of the manufacturer. In case thesecond set of test specimens fails to meet re-quirements, the entire lot shall be rejected.

EXPENSE OF TESTS

14. Except as specified in Section 13, expense ofinspection and testing shall be borne by thepurchaser.

SAMPLING AND TESTING

15. (a) The purchaser or his authorized representa-tive shall be accorded proper facilities for sam-pling and inspection of filter block, both at theplace of manufacture and at the site of the work.At least ten days from the time of samplingshall be allowed for completion of tests.(b) Filter block shall be sampled and tested inaccordance with the Methods of Sampling andTesting Structural Clay Tile (ASTM Designa-tion: C 112), except that resistance of block tothe action of acids shall be determined in ac-cordance with the applicable portions of Sec-tions 14 to 18 of the Methods of Testing ClayPipe (ASTM Designation: C 301).

GUARANTEEOF VITRIFIED CLAY UNDERDRAW FLOOR BLOCKS

Warrant and Guarantee to

that all vitrified clay underdrain blocks manufactured by the above said company

and furnished for installation in

under contract dated

between. .and.are proof against impairment of service or failure due to rust, corrosion, decay,chemical decomposition and/or disintegration caused by any action of acids,alkalis, sewage, sewer gases, ground waters or industrial wastes, excepting Hydro-fluoric Acid only. Also guaranteed against destruction by rats, rodents or termites.

The undersigned bind themselves for a period of

FIFTY YEARSafter date hereof to furnish to the above namedfree of cost, vitrified clay underdrain blocks in the same quantity and sizes equalto any of the vitrified clay underdrain blocks manufactured and furnished by abovenamed manufacturer under said contract which may have been damaged, destroyedor impaired in service by reason of any of the above named causes.

It is expressly understood that the conditions of This Guarantee do not providefor Furnishing of Blocks which have been broken or impaired by reason ofImproper Handling, Placement or Construction.

The. . Company

By -Title-

Signed this- _day of- -19-

RICKLING FILTERLOOR INSTITUTE

Standard Rate Filters

THE first trickling filters were ofthe standard rate type, designed

to use the concept of low rates ofapplication, up to 600 lbs. of BODper acre-foot, on relatively deepbeds of rock or crushed stone. Thepopular method of distribution ofwastewater over the media was bymeans of fixed nozzles, with dis-tributing manifolds buried in themedia.

In the 1930's, the trend switchedto circular beds with rotating dis-tributors which consisted of armsextending over the diameter of thefilter suspended from a central in-fluent column. In this same decade,the concept of continuous applica-tion was developed, which resultedin recycling the effluent in a pre-determined ratio to influent on anaverage flow basis. The recirculationprinciple also helped maintain ap-plication to the beds, regardless ofinfluent flow conditions. This devicebecame known as a high rate filterbecause the overall dosage is muchhigher than that of standard niters.

Both high rate and standard ratefilters are in use today, though thetrend is toward higher loadings andhigh rate filter construction.

Many studies have been made ontrickling filter performance, the mostnotable of which was that of theCommittee on Sewage Treatment atMilitary Installations of the NationalReset rch Council. This was done in1944-45 and was reported in theSewage Works Journal, September,1946.

Trickling filters are constructedin varying depths to carry varyingorganic loadings with equally vary-ing hydraulic loadings. Thus, thereare problems in determining whatmay be expected of a proposed filterunder varying loading conditions.The National Research CouncilStudy (hereafter referred to asNRC) established empirical for-mulas which can be used to deter-mine quite closely what may beexpected under given conditions ofloading. Other formulas have alsobeen derived, but these relate gen-erally to high rate filters.

What might be termed an averagedepth for standard filters is 6 ft. butactual depths of installations in usevary from 5 to 10 ft. It has been

contended that the greater depthsresults in increased nitrification ofthe effluent. The NRC study referredto above stated that "no significantdegree of nitrification occurred inplants with a daily BOD load in ex-cess of 500 pounds per equivalentacre-foot." Some plants "did notproduce a significant amount of ni-trate despite low loading."

The BOD loading standards forstandard filters are by no meansfixed. The upper limit is generallyabout 600 lbs. of BOD per acre-foot per day, but a loading of theorder of 400 lbs. is more common.On 15 New Jersey plants studiedby Forman and Shaw, applied BODloadings averaged 367 pounds ofBOD per acre-foot. The averageloading on 10 Army plants as re-ported by the NRC Committee was377 lbs. In general, primary con-sideration in designing and in de-termining the load on standardfilters should be based on local con-ditions and on State Health De-partment or other regulatory agencystandards. Heavier loadings are pos-sible in warm climates than in coldclimates.

Volumetric LoadingThe volumetric loading on stand-

ard filters averages about 300,000gallons per day per acre-foot; or,for a 6-ft. deep filter, a surface load-ing rate of about 1.8 mgad. There is,however, no fixed standard. Basedon a BOD strength of raw sewageof 240 mg/L, a flow of 1 mgd, a re-

moval of 30 percent in the primarysettling tank, and an organic loadingof 400 pounds of BOD per acre-foot,there would be required 3.5 acre-feet of filter. In this case the volu-metric loading rate would be about285,000 gallons per acre-foot perday, or about 1.71 mgd per acre ofsurface. As in this example, the or-ganic loading will often govern thevolumetric loading. The medianvolumetric loading on the 10 Armyplants referred to was 1.13 mgad,a rather low value.

The media used in standardtrickling filters should comply withthe requirements given elsewherein this text Specifications shouldinclude resistance to freezing (notneeded in far southern states); sizerange; percentage of fines; screen-ing before placement; method ofplacement in the filter; and otheritems contained in the sample speci-fications given.

For underdrainage and ventila-tion, specially designed blocks, alsodescribed in another chapter, shouldbe used. Specifications for suchblocks are included elsewhere. Cen-ter channels in standard filters aresmall in comparison with those re-quired for high rate filters, andstandard channel cover blocksshould be used. Vent blocks foropenings through the wall, andcover blocks for vent risers, areavailable from the same sources asare underdrain blocks and channelcovers.

Rotary distributors are normally

• LARGE areas occupied by standard rate filters compared with those used by highrate units have favored construction of the latter type in modern design practice.

11

used for applying the sewage to thefilter. Most of these are revolvedby the reaction to the dischargethrough the orifices which areplaced on opposite sides of the op-posing arms. Most distributors willoperate satisfactorily with a mini-mum discharge that is 40 percent ofthe maximum, that is, they willhandle flows up to 2% times theminimum, but 4 ft. of head shouldnormally be provided with this ratio.

The flow of wastewater does notalways fall within the operatinglimits of the distributor, and a dos-ing tank is usually provided. Thehigh water level of the dosing tankmay be 2 to AVz ft. above the levelof the orifices in the distributorarms. However, where local condi-tions require that every possibleinch of fall be conserved, specialdesigns are available and the manu-facturer's engineers should be con-sulted. Distributors having 4 arms,arranged so that 2 arms serve forlow flows and 4 arms for high flowspermit operation with flows up toabout 5 times the minimum. Thesame result can be obtained by hav-ing two channels in each arm, anupper and a lower one. In all caseswhere high flows may occur, thehydraulics of the siphon, the feedpipe and the distributor should becalculated carefully to insure thatthe friction head is not so great asto cause the siphon chamber tooverflow.

Head Losses

The head losses between highwater level in the dosing tank andthe top of the filter media include:a) The entrance loss in feet due todischarge into the siphon pipingwhich may be assumed at 4v2/2g,where v is the velocity in the filterfeed pipe; b) the dosing tank lossor the drop in level in the dosingtank required to fill the distributorarms which equals the volume ofthe arms in cubic feet divided bythe area of the dosing tank in sq.ft.; c) the friction loss in the feedpipe which is computed from pipeflow tables with allowances forbends and for increase in velocityhead if the feed pipe is smaller thanthe siphon; d) the distributor losswhich should be obtained from themanufacturer; 12 in. is the usualpermissible minimum; and e) thedistance from the center line of thedistributor arm to the top of themedia which is usually 6 in.

The dosing tank should provideabout 2 minutes detention at twicethe average rate of flow. This allowsfrequent dosing and keeps the sizeof the dosing tank within practical

limits. Since every foot of draw-down in the dosing tank representsso much head loss, and since thetop layer in the dosing tank is dis-tributed most effectively over thefilter surface, it is customary tomaintain a drawdown of 10 in. butnot over 12 in. The total loss ofhead from the high water level ofthe dosing siphon can be determinedfrom these data.

Where dosing tanks are not em-ployed and the wastewater ispumped, the distributor must bedesigned to take the flow as it is re-ceived.

PretreatmentPretreatment is normally by set-

tling; a period of 2 to 2.5 hoursshould be provided, with properallowance for future-growth. Sincerecirculation is rarely used withstandard rate filters, though it canbe, settling tank capacity is based onthe expected flow of wastewater atthe date selected for design. Reten-tion period, overflow rate, depth oftank, weir design and some otherfactors are subject to State Boardof Health or other regulatory agencyrequirements and may differ mate-rially from state to state.

In design, a BOD removal of 30percent to 35 percent in the pri-mary settling tank is used, depend-ing on overflow rate and retentionperiod. The specific figure to beused will generally depend on StateBoard of Health or other regulatoryagency requirements.

In general, the above applies alsoto secondary settling tanks follow-ing the filter.

Filter PerformanceThough standard filters are quite

uniformly stable in. day-to-day per-formance, there are variations ineffluent quality from filter to filterand even occasionally from monthto month, for which adequate ex-planations are not available. Theformula which follows is based onthe curve of best fit for 10 Armyplants studied by the NRC Commit-tee previously mentioned. It showswhat a standard filter may be ex-pected to do on the average. Therewill be variations from such aver-age, especially seasonal variations.The 10 Army plants produced efflu-ents ranging from 14 to 89 mg/L,BOD with an average of 38.0 mg/L.

Design ComputationsThe formula shown below was

derived by the NRC subcommitteepreviously referred to. It measuresthe probable efficiency of the filterand the following clarifier. It will

not always indicate what a particu-lar filter is doing, but it will closelyapproach what may reasonably beexpected from a well designed, wellconstructed and properly operatedfilter. The efficiency of such a filtershould be:

100E =

1 + 0.0085

where E is the efficiency of the filterand following clarifier in terms ofthe percent removal of applied BOD;and L is the organic loading inpounds of BOD per acre-foot thatis applied to the filter. Since, in afilter of the type discussed in thischapter, there is no recirculation,L is the applied BOD in poundsdivided by the filter volume in acrefeet

The procedure in using the for-mula is shown in the following il-lustrative examples.

With a flow of 1 mgd, a raw sew-age BOD of 240 mg/L and 30 per-cent removal in the primary clari-fier, the BOD applied to the filteris 240 x 8.33 x 0.70, or 1400 lbs.Assume an organic loading of 400lbs. of BOD per acre-foot, whichwill be the value of L in the formula.There will be required 1400 H- 400,or 3.5 acre-feet of media. With afilter depth of 6 ft., the area of thefilter will be 0.583 acre. The effi-ciency of removal by the filter andfollowing clarifier by the formulawill be 85.5 percent. The effluentfrom the secondary clarifier willcontain 1400 (1.00-0.855) or 203pounds of BOD which is the equiva-lent of slightly more than 24 mg/L.The overall reduction through theplant will be 90 percent.

The filter in the above examplehas been loaded at what is prob-ably the best standard loading. Theeffect of using a somewhat higherloading will now be shown. Assumea loading of 600 pounds of BOD peracre-foot which will be L in theformula. With a depth of 6 ft, thefilter area will be 0.39 acre and thefilter volume will be 2.33 acre-ft. Ethen is 82.8 percent and the effluentwill contain 29 mg/L of BOD. Theoverall reduction through the plantwill be about 88 percent.

The standard filter, properlyloaded, gives very fine results inBOD removal, with the productionof a stable effluent. The volume ofmedia and the area of filter re-auired is large. If these factors areof importance from the viewpointsof space required or of cost, it isbetter to use a high rate filter thanto overload a standard filter.

12

A Nomographic Solution for Design

THE PURPOSE of the nomographI presented here is primarily to

reduce the computation time ior apreliminary investigation of thesize of units for a trickling filtercomplete treatment sewage plant.While nomographic solutions cannotoffer the accuracy of mathematicalcomputation, they can be used togive quickly an estimate of gen-eral unit sizes for a proposed plant.The nomograph for the design oftrickling filters is essentially cor-rect in terms of its representationof the NRC formula, but great pre-cision is lost in scale distributions.Consequently, it is suggested thatthis nomograph be used primarilyfor initial estimates or gross esti-mates of trickling filter sizes.

Another use is for the computa-tion of the theoretical percentageof BOD removal in any filter, giventhe filter diameter, the depth, re-circulation factor, and the poundsof BOD applied per day. Computa-tion of filter dimensions can bemade by selecting 'the desiredpercent of BOD removal and thepounds of BOD to be applied tothe filter per day. The filter beddiameter, the depth, and the re-circulation ratio can then be de-termined.

The design limits are not indicat-ed on the nomographs. It is sug-gested that those utilizing thisgraph familiarize themselves withthe appropriate state limitations.

For using the nomograph for fil-ter design only three values mustbe known: 1) The percent BODremoval desired through the filter;2) the pounds of BOD which willbe applied to the filter per day; and3) the total gallons of waste (inmgd) applied to the filter. (A validassumption is that 65 percent of thedaily raw sewage BOD will be ap-plied to the filter.) Knowing thesevalues, it is then possible to ex-amine and select comparative sizes,depths, rates of application and re-circulation, if required.

Step 1. Through the scale "%

BOD Removal" locate the % ofBOD to be removed from the filter(within the latitude ranges as in-dicated on these scales).

Step Z. Extend a line throughvalue on "% BOD Removal" scalethrough intersection of R-R andT-T until it strikes a point on"Pounds of BOD Applied per Acre-Foot" scale.

Step 3. Join known values of"Pounds of BOD Applied to Fil-ter Per Day with the point locatedin Step 2 and note intersection ofScale R-R.

Step 4. Project this point on ScaleR-R along the parallel guide linesto a point of intersection on scaleT-T.

Step 5. The point located in Step4 (above) becomes the pivot pointfor relating depth and area figuresto meet design requirements asfound in Step 2. With a straightedge pass a line through pivot pointon scale T-T and read values fromscales "Effective Depth (Ft)" and"Filter Bed Diameter (Ft)". Anycombination of Effective Depth andBed Diameter will meet designcriteria in terms of organic loading.Note: The concept of effectivedepth is related to recirculation inhigh rate filtration. In a standardrate filter, the effective depth isequal to the actual depth and re-circulation is zero. The effect ofrecirculation, or the recirculationfactor F where:

1 -t R/IF =

(1 + 0.1 R/I) -allows, in the NRC formula fortrickling filters, for a mathematicalcombination of depth and F, whichis labeled "effective depth." Con-sequently the recirculation factorF has the effect of increasing thedepth. The actual depth can thenbe found through a nomographicsolution shown on the chart whichrelates the effective depth througha recirculation ratio R/I.

Step 6. Practical selection of depthand area-diameter values will

be dependent upon design criteriaand economic considerations to beevaluated by the designer.

Step 7. The "Hydraulic LoadingRate" may be checked by passinga line through "Filter Bed Diame-ter" as selected in Step 6 and the"Daily Flow to the Plant (mgd)"to the "Hydraulic Loading Rate(mgad)" scale. Area and flowcombinations should fall into thecategory of high or standard ratedepending upon criteria chosen. Ifthis does not occur, adjustment offilter diameter and depth should bemade to accommodate the hydraulicloading rate.

When an existing filter is in op-eration the theoretical efficiency canbe easily computed with this nomo-graph. All the following factorsshould be known from the installa-tion: 1) Filter diameter; 2) actualdepth; 3) recirculation ratio; and4) pounds of BOD applied to thefilter.

Step 1. If recirculation is used,join values of "Actual Depth" and"Recirculation Ratio" to determinethe "Effective Depth."

Step 2. Join value of "EffectiveDepth", as found in Step 1, withvalues of "Filter Bed Diameter"and note point of intersection withscale T-T.

Step 3. Project point at T-Talong parallel lines to intersectionwith scale R-R.

Step 4. Through R-R and valueof "Pounds of BOD Applied to Fil-ter" locate value of "Pounds of BODApplied per Acre-Foot."

Step 5. Through value found inStep 4 and point of intersectionof scales R-R and T-T join a lineto intersection with "% BOD Re-moval" scales. This value yieldsthe theoretical "% BOD Removal"for the filter of given size.

The red lines indicate an ex-ample worked out for high ratefilters; the green, for a standardfilter.

From an article by Walter R.Lynn, PUBLIC WORKS, August1957

13

POUNDS OF B.O.D. APPLIED PEF

l\) Ol -f> O) <T> -W0D(0OO O O O O O O O O

J I I . 1 . 1 . 1 i l i | i l i l

no

ro Ol

POUNDS OF B.O.D APPLIED PE

Ol

o

CUB

70

II

70Q

CO

OQ

no

oCn

15—1

T lmo

m

om

HX

—_

1 0 -9 -8 -7 -6 -

5 -

4 -

3 J

I • IOloooo

roOl

ooo

1 •• 1 • 1— <£> 00

Pooooooooo

11',-Jooo

1 10)"boo

1 1Ol

ooo

1 1ooo

1 1Ol

'ooo

1 1ro"boo

I' l ' l— ID CO

ooooooo

11->loo

11oo

1 1 'Ol

oo

1

cc

POUNDS OF B.O.D. APPLIE

14

- FOOT

fr3O

_, 0E> U> O'o o*o"o"oo o oooo o ooo1. 1.1.1.1

DESIGN OF TRICKLING FILTERS(U.M.R.B. CRITERIA)

Based upon N.R.C. formula:

100P — 1 + 0.0085

ADF

Where: P = % B.O.D. removalW = lbs. of B.O.D. appliedA = area of filter in acresD = depth of bed in feetF = recirculation factor

InoOo

I'TM'I-ICONO o O Oo

1mo

1 1o

1 1 'o

1 'wo

1 •noO

1—o

TO FILTER PER DAY

15

The Design of High Rate Filters

• HIGH RATE filter plant servingSalina, Kansas, with 115-ft. distribu-tors is shown above. At right is a 70-ff. diameter distributor dosing atthe rate of 25 mgad at Suffern, N. Y.

WALLACE J. BENZIE. P. E.Director,

Sanitary Sewers and Engineering,Genesee County Drain Commission,

Flint, Michigan

H IGH RATE trickling filters arethose in which some portion of

the filter discharge or final effluentis returned to the influent of theprimary sedimentation tank whereit joins the raw sewage; or to thefilter influent where it joins theprimary effluent. Plants can be de-signed to give any degree of treat-ment between sedimentation andcomplete treatment. Filters areusually dosed at 5 to 10 times therates employed on standard filters,and with careful design, they canproduce an effluent as low in sus-pended solids and BOD as a stand-ard filter but with little, if any,evidence of nitrification.

The much heavier organic load-ings used with high rate filters haveresulted in the use of a variety ofterms to designate such loading. Therather commonly used basis, withstandard filters, of pounds of BOD

per acre-foot, have been applied tohigh rate filters, and loadings ofBOD of 2,540 to more than 3,000pounds per acre-foot have beencommon. Some engineers have re-duced this loading to a cubic yardbasis, dividing the acre-foot load-ings by 1,613. There has also beensome use of 1,000 cu. ft. as a loadingbasis, and it has been recommendedby a committee of the Water Pollu-tion Control Federation that suchusage be standardized in sewagetreatment practice. Since 1,000 cu.ft. equals 37 cu. yds., the cubic yardloading is multiplied by 37 to obtain

the 1,000-cu. ft. loading. Thus, 3,000lbs. per acre-foot is the same as1.86 lbs. per cu. yd. or 68.8 lbs. per1,000 cu. ft. There may be a slightsaving in time in making designcomputations by either the cu. yd.or the 1,000-cu. ft. basis, but mostformulas now in general use, fordetermining filter efficiency arebased on the acre-foot.

There are three principal typesof high rate filter in common usetoday, based on rate of feeding,recirculation or loading.

In one type of filter operation, aportion of the trickling filter effluent

16

I RECIRCULATION design arrangements available are varied.In the diagram above, larger circles represent filters; small

circles are clarifiers; and heavy lines, recirculation paths.BOD reductions are somewhat dependent on layout employed.

is recirculated to the primarysettling tank influent. It is some-times desirable to add a secondstage filter to this system where itis necessary to provide for addition-al treatment. Loading rates and re-circulation ratios can be modifiedto provide the desirable effluentquality. Organic loadings are in theorder of 2,500 to 3,000 lbs. of BODper acre foot, or 1.5 to 1.9 lbs. percu. yd. based on the strength ofthe primary tank effluent. The vol-ume loadings range from 10 to 30million gallons per acre per day(mgad).

The unsettled filter effluent mayalso be recirculated directly back tothe filter influent with no inter-mediate settling provided. This re-circulation principle can be usedon either standard or high ratefilters.

In addition, primary effluent maybe distributed over a maximumarea of the filter at one time withrecirculation only at times necessaryto maintain the flow requirementsto provide for satisfactory distribu-tion of the waste onto the filtermedia. Recommended organic load-ings range from 3,000 to 3,200 lbs.of BOD per acre foot or 1.9 to 2.0lbs. per cu. yd. based upon the ap-plied sewage.

Additional treatment may be pro-vided by utilizing a second stagefilter. Because of the heavier or-ganic load applied to a high ratefilter per unit of filter volume, theeffluent from a single stage highrate filter will be higher in BODthan that from a lightly loadedfilter. To overcome this factor, theeffluent from the filter may be mixedwith the incoming raw sewage andpassed again through the filter; ortwo filters, in series. These pro-cedures may result in treatmentcomparable to that from a filter that

is loaded less heavily, while thesmaller units required in the highrate process compared with stand-ard filters may reduce constructioncosts materially.

FlowsheetsFlowsheets most frequently used

are shown in the accompanyingdiagram. The choice of flowsheet isdependent upon several factors, in-cluding the strength of the rawsewage, the quality of effluent de-sired and, to some extent, on thesize of the plant. In a plant wheremultiple filter unite are requiredbecause of size, local requirementsor other reasons, it is frequentlydesirable to arrange for two-stageoperation with small additional cost.A single-stage plant is differentiatedfrom a two-stage plant by the factthat in the latter the sewage ispassed successively through a pri-mary filter and through a secondaryfilter. Recirculation may be appliedto both filters.

Thus, a single-stage filter is onein which the filters, if there aremore than one, are operated inparallel. Sewage may be, and usual-ly is, recirculated through single-stage filters, the effluent from eachfilter being passed to a secondarysettling tank after or before with-drawing the portion that is to berecirculated.

In a two-stage filter, the effluentfrom the primary filter or filters,after withdrawing that portion tobe recirculated and returning it toand through the primary filter,passes to a second filter; and theeffluent from the second filter, lessrecirculation, is discharged to thefinal settling tank.

RecirculationIn all high rate trickling filter

plants, some form or pattern of re-

circulation is always used. The ratioof the volume of flow recirculatedto the volume of average rawsewage flow is defined as the re-circulation ratio R. Where morethan one recirculation means isused in a plant, the ratios are dis-tinguished as Rj, R2, etc. The bene-ficial effects of recirculation havebeen demonstrated.

Return of recirculated filter ef-fluent to the raw sewage has provedbeneficial in reducing odors par-ticularly in initial stages of plantoperation when the sewage flow isfrequently much below design con-ditions and often septic. Recircu-lated flow is also effective in re-ducing scum, particularly whentreating stronger sewage.

There are many variations in thearrangements for recirculation. Ina single-stage filter, the most com-mon arrangement is to return a por-tion of the filter effluent to the in-let of the primary settling tank.In the two-stage plants, the mostused plan is to return a portion ofthe primary filter effluent for an-other passage through the primaryfilter; and a portion of the secondaryfilter effluent for another passagethrough the secondary filter.

However, excessive recirculationis not necessary. Very good effluentscan be obtained with recirculationratios of 0.5 to 1.

Loadings on high rate filters mustbe considered from two bases: Theorganic loading in pounds of BODper acre-foot, cubic yard, 1,000cubic feet, or square foot of surfacearea; and the rate at which sewageis applied to the surface, usually inmgad.

The organic loading is based onthe strength of the sewage appliedto the filter. In design, it is com-mon practice to allow 30 to 35 per-cent removal of BOD in the pri-

17

mary settling tank and to assumethat 65 to 70 percent of the rawsewage BOD will be applied to thefilter.

Much heavier loadings than thosestated above are applied to theprimary filters of the two-stageplants. In a two-stage plant, thesame volume of filter media is usedas for a single-stage plant, with themedia divided, usually equally, be-tween two filters. The entire or-ganic load, after primary settling,is applied to the primary filter. Thismay result in loadings of 3 to 4 lbs.per cu. yd. or 4,800 to 6,400 lbs. peracre foot. In effect, the primaryfilter becomes what has been calledin the past, a roughing filter, withthe principal function of removingBOD and preparing the sewage forthe final passage through the secondfilter.

Volume loadings range from 10to around 30 mgad, based on thesurface application rate. Assuminga sewage flow of one mgad, a re-circulation ratio of 0.5 to 1, in atwo-stage filter, the BOD of theraw sewage is 240 mg/L and theremoval in the primary settlingtank is 30 percent, the organic loadon the filter will be 8.33 x 240 x0.70 = 1,400 lbs. If the unit organicloading in the filter is 1.75 lbs. percu. yd., 800 cu. yds. of media willbe required, divided into two filters,each containing 400 cu. yds. If thefilters are 5 ft. deep, the surfacearea of each will be 2,160 sq. ft. orabout 0.05 acre. Since the daily flowof sewage plus recirculation effluentwill amount to 1.5 mgad, the surfaceload on the filter will be about 30mgad.

If a single-stage filter were used,all of the 800 cu. yds. of filter mediawould be in a single filter, and avolume loading rate would be halfas great, or 15 mgad. Few plantsare designed for rates much above30 mgad and 50 mgad is about themaximum that has been used.

Predicting PerformanceWhen laying out new high rate

filter plants, designers need to esti-mate or predict their performance.This can be done with reasonableconfidence through the applicationof certain formulas developed fromanalysis of operating data fromexisting high rate plants. It shouldnot be concluded that sewage treat-ment has become an exact sciencewith precise performance predicta-ble by formulas. Such formulas do,however, serve as a tool or guideto the experienced designer. Onesuch basis is the accompanying TenStates Standard Curve. This showsthat with BOD applied to the filter

at the rate of 40 lbs. per 1,000 cu. tt.there will be 28 lbs. per 1,000 cu. ft.removed. Other expected removalsare readily obtainable from thiscurve.

Another method of predictingperformance preferred by some de-signers is on the basis of the NRCformulas developed specifically fromtrickling filter installations servingmilitary bases in World War II. Inthese formulas, weight is given tothe organic loading placed on thefilter, and also to the ratio of re-circulation, since it was found thatthese factors affect the quality ofthe effluent. The volume of filtermedia is stated in acre-feet. Theorganic loading is a function of theBOD loading, the acre-foot volumeof the filter, and the recirculationfactor.

Since, in recirculation, part of theorganic material is removed in eachpassage through the filter, a re-circulation factor F is derived fromthe recirculation ratio by means ofthe following formula, where R isthe recirculation ratio:

ume in acre-feet, F the recirculationfactor and W the pounds of appliedBOD:

WL = . . . (2)

V X F

Thus, if applied BOD is 1,400pounds, V is 0.5 and F is 1.65, L =1700.

The filter efficiency formula is:

100Tji . . . . (3)

1 + RF =

(1 + 0.1R)2. . . ( 1 )

For a 1 to 1 recirculation ratio,F is 1.65 and for a ratio of 0.5 to 1,F is 1.36.

The percent efficiency of removalof BOD in a trickling filter and itsfollowing clarifier is expressed byformula (3) below, where E is theefficiency in percent of removal oforganic matter and L is the organicloading. L is derived by formula (2)below and, for filters with recir-culation, is based on pound of BODapplied, filter volume and recircula-tion. With V representing filter vol-

1 + 0.0085 V L

and where L = 1,700, the efficiencyof the filter and accompanyingclarifier will be 73.7; that is 73.7percent of the applied BOD will beremoved in the filter and final clari-fier (not considering the removalin the primary clarifier.)

Thus, for the example cited, re-moval in the filter and followingclarifier will be 1,400 X 0.737 or1,032 pounds. BOD remaining in theeffluent will be 1,400 - 1032 = 368pounds. Overall removal in theplant, including that in the primaryclarifier will be 81.6 percent.

The above formula applies to anysingle stage high rate filter no mat-ter what the rate of loading. It alsoapplies to the primary filter of atwo stage plant; but it does not ap-ply to the second stage filter of suchan installation.

In predicting the performance ofthe second stage of a two-stageplant, the reduced treatability of theorganic waste must be taken intoaccount. A formulation is given inan article, "Design Considerationsfor Biological Filters," PUBLICWORKS, February, 1967. Compila-tions of data on the operation of two-

BOD REMOVAL curve used in Ten States Standards for filters and final settling.

ooo

oom

30

20

10

r

B Poin

tci

ent

d Th

isIn

suff

Be

yon

10 20 30 40

BOD Applied - Pounds per 1000 Cu. Ft.

50

18

stage filters are available. Two no-table references are "Sewage Treat-ment at Military Installations,"Sewage Works Journal, Vol. 18, No.5 (1946) and R. S. Rankin, Trans-actions ASCE, Vol. 120, p. 823(1955).

ClarifiersPrimary and secondary clarifiers

will normally be designed in ac-cordance with the State Board ofHeal th requirements; commonlyrecommended detention is about 2hours in primary settling tanks,with overflow rates of 650 to 800gallons per sq. ft per day; and thesame for secondary clarifiers. De-pending on recirculation practices,clariiier capacity may be based onraw sewage flow; as when recircula-tion is merely to maintain distribu-tor operation during low flow peri-ods; or on raw sewage flow plus re-circulation, when strong sewage orother factors dictate recirculation toprovide dilution. The size of theprimary clarifier is usually selectedon the basis of average 24-hour rawsewage flow plus recirculated flowreturned to the clarifier influent.

Treatment ResultsSingle-stage high rate filters will

normally produce effluents of theorder of 40 to 45 mg/L BOD whenloaded in the range of 2,400 to 2,800pounds of BOD per acre-foot. Two-stage filters, with the same loadingwill normally produce effluents inthe range of 24 to 28 mg/L BOD.The Sandusky, Michigan, treatmentplant has a filter volume of 9,850cu. ft. and was designed with a re-circulation ratio of 1:1. This planthas shown reductions in BODthrough the filter of more than 80percent when loaded at 54 poundsof BOD per 1,000 cu. ft. This plantis quite typical in its operating re-sults, of a number of plants of thistype existing in Michigan today.

The performance of trickling fil-ters is affected by several factorsincluding media size, depth ofmedia, ventilation, hydraulic rate ofapplication, organic loading andtemperature variation. In a paperpublished in the WPCF Journal,Vol. 35, No. 4, pp. 445 to 455, theeffects of temperature variation ontrickling filter performance at sev-eral selected plants in Michiganwere reported. The conclusions ofthis study are:

1) There is a significant differ-ence in trickling filter efficienciesbetween the summer and wintermonths.

2) Recirculation of wastes througha filter has a marked cooling effecton the waste during the winter

months; and by lowering the tem-perature of the wastes applied tothe filter, the efficiency of the trick-ling filters is lowered.

3) Lower air temperature has amore significant effect on reducingfilter efficiencies in plants that re-circulate than those that do notpractice recirculation.

4) In plants that recirculate, thewinter filter efficiency was 21 per-cent less than the summer effi-ciency.

5) Filters having a BOD loadingof more than 10 lbs. per 1,000 cu. ft.of filter media showed no greaterseasonal variation in efficiency thanwhen loaded at less than 10 lbs.

6) Filter efficiency was affectedin those plants not practicing re-circulation when the air and sewagetemperatures were equal.

7) In the design of trickling filter

plants in the northern part of theUnited States, consideration shouldbe given to the marked change inefficiencies which occur between thesummer and winter seasons.

To obtain the predicted per-formance in an operating plant, in-telligence must be used in the selec-tion of materials and equipmentgoing into the plant and in obtain-ing a qualified operator to run it.Such items as filter media, filterunderdrains and rotary distributorsare very important and specifica-tions should leave nothing to chance.Correctly sized pumps, particularlyfor recirculated flows, are also veryimportant. Experience has demon-strated that properly designed, con-structed and operated high ratetrickling filter plants have little dif-ficulty in obtaining results equal tothe predicted performance. O D D

• DISTRIBUTOR with multiple arms is a feature of the Aero-filter ina Gaffney, S. C. installation. Below are shown the secondary units of atwo-stage filter plant that provides treatment for the Air Force Academy.

19

Trickling Filters for Industrial Waste Treatment

THE wastewaters from industryare, of course, extremely varied

in character, even within the sameindustry or manufacturer's opera-tions since much depends on thetype of processing employed. Many,however, are organic, such as foodwastes and those from distilling andpharmaceutical operations. Theseare subject to biological oxidationand consequently may be organical-ly stabilized through treatment ontrickling filters. Some may be de-ficient in elements required to sus-tain microbial growth or naturallylacking in organisms necessary tostart growth on the filter. However,these conditions can be corrected byadding the required nutrients andbiological seed. Some wastes, suchas phenols, are toxic to micro-organisms, but techniques are avail-able to handle even these wastes ontrickling filters. The selection of thetype of filter, whether high or stand-ard rate and its design, dependsgenerally on the same factors as areencountered in municipal domesticwastewater treatment plant design:The amount of BOD to be removed;preliminary treatment; the load fac-tor; the recirculation ratio; and siteconditions. One usually differentfactor is the volume of flow.

Whereas municipal sewage flowsare calculated in millions of gallonsper day, the flow of industrial wasteswill frequently be counted in thethousands of gallons per day. Thiswould perhaps lead to the belief thattrickling filters for industrial wastescan be small. This is not necessarilythe case. A normal municipal sew-age will 'have a BOD of 250 to 400mg/L or 2085 to 3336 pounds perMG., but it is not uncommon for anindustrial waste to have a BOD of4000 mg/L, and in some instances ashigh as 30,000 mg/L. This means33,600 to 250,000 lbs. of BOD perMG. The trickling niters for suchhigh organic loadings, even thoughthe daily volumes are small, mustbe very large.

Preliminary treatment may followany one of a number of acceptedpatterns including anaerobic diges-tion of very strong wastes; plantmanagement to modify operationsproducing strong and difficult totreat wastes; recovery; and blend-

ing. The method selected will de-pend on local conditions. The aimshould be to reduce very high BODcontents and to adjust the wastesso they are amenable to aerobictreatment. Without prejudice forone type or another, it may be con-ceded that the high rate tricklingfilter will often have advantages inthe treatment of industrial wastes,due to lesser space requirementsand often lower cost.

There is a tendency, where feasi-ble, to combine industrial and do-mestic wastes for treatment by anearby urban complex, with theindustries providing their share ofthe treatment costs. In designingsuch facilities, careful consideration•has to be given to the influence ofthe industrial waste load. Some in-dustries, such as canning, are sea-sonal in operation, which discour-ages the use of biological oxidationmeasures for handling the industrialwaste alone, because time is re-quired to start a trickling filter oran activated sludge plant.

The High Rate FilterThe higher the initial BOD con-

centration, the greater the loadingfactor per unit of bed volume orarea may be. Also, as the BOD con-centration increases, better resultsfrom the standpoint of percentageof reduction are obtained by in-creasing the ratio of recirculation.The volume of recirculated filter ef-fluent has usually little effect on thesize of the filter itself but principallyaffects the design of the distributorunit.

The flows from industrial plantsare usually relatively small so thatincreased recirculation rates useddo not require the use of largecapacity pumps or result in highpower demands. For instance, if thewaste flow is 100,000 gpd and it hasbeen determined that two-stagehigh rate filtration will produce asatisfactory reduction in BOD whenthe rate of recirculation in eachstage of filtration is 3 times the vol-ume of raw waste, recirculation willbe 300,000 gpd in each stage. Forthis, in each stage, a centrifugalpump will be required with a capac-ity of 208 gpm against a total headof 15 ft. This may be obtained with

a 3-in. pump, requiring a lV2-hpmotor. As these pumps run con-tinuously, the total daily usage ofelectrical power with the two pumpsin operation will be only 54.0 kwh.

To cite an example: The wastesfrom the production of alcohol fromsugar cane molasses have BOD con-centrations ranging from 22,000 to33,000 mg/L. It is common practice,when treating the effluent from thepre-digestion step used in handlingthese wastes, to employ recircula-tion rates of 5 to 7 times the volumeof digester effluent Even in a largedistillery of this type, a waste flowof 225,000 gpd is about the upperlimit, so that with a two-stage plant,using 3.5 times recirculation in eachstage, the pump size for each phasewill be 547 gpm and the pump willrequire only 3 hp to operate it.

Specific ExamplesCannery Wastes. Trickling filters

are satisfactory for these wastesbecause the BOD is exerted rapid-ly due to the high sugar contentand its quick decomposition. Forwastes averaging around 1000 mg/LBOD loadings of 5 to 6 lbs. of BODper cu. yd. may be used. Tests oncorn cannery wastes have indicatedthat loadings as high as 22 lbs. percu. yd. may be used on primaryhigh rate niters, preceded by sedi-mentation. With such loadings, re-ductions of approximately 50 per-cent of the BOD have beenachieved, with the raw wastes hav-ing an average BOD of 8400 mg/L.

Sauerkraut packing wastes havebeen treated on niters using blastfurnace slag as the medium. In thetest period, raw waste loadings onthe filters ranged from 575 to 6800lbs. BOD per acre-foot per 9-hourday. The dosing rates ranged from9 to 30 mgad, and the rates of re-circulation from 3.3 to 19. The char-acteristics of the raw wastes werepH 4.0; acidity 1100 mg/L asCaCO3; chlorides 4500 to 6500mg/L; total solids 12,000 to 13,000mg/L; BOD 2000 to 2400 mg/L. Thefilter used in the pilot plant was 10ft. in diameter and 6 ft. deep withslag media. With adjustment of thepH of the wastes to a minimum of6.6, a BOD reduction of 85 percentwas obtained. Concentrations of

20

chlorides up to 5000 to 6000 mg/Lseemed to have little effect on thebiological efficiency of the filter.There was no evident relation be-tween the ratio of recirculation andthe reduction in BOD.

Dairy Wastes. These comprise thewaste waters from creameries,where milk is separated; bottlingplants; milk condensing plants;powdered milk plants; and cheese,butter, ice cream and other milkproducts plants.

The BOD ranges from approxi-mately 100,000 mg/L for whole milkwastes to 30,000 mg/L for whey.The wastes respond to treatment ontrickling niters after preliminarytreatment to remove coarse solids.In treatment, it is common practiceto include a step of aeration to in-hibit anaerobic action during theretention period and this also aidsin BOD reduction. The screened andaerated wastes are applied from abatch holding tank to either lowrate filters or to high rate niterswith recirculation. High rate filterloading is 1.0 lb. of BOD per cu. yd.Reductions of 90 to 95 percent of theBOD have been obtained.

Single-stage recirculating typehigh rate filters are satisfactorywhere a final effluent with 70 to 100mg/L BOD is acceptable. With two-stage high rate filters operating inseries, with recirculation in eachstage, reductions of 90 to 95 percenthave been obtained with primaryBOD loadings of 1.5 to 2.0 lbs. percu. yd. and secondary filter loadingsof 0.75 lb. per cu. yd.

Fermentation Industry Wastes.These comprise the wastes frombreweries; distilleries for the pro-duction of whiskey, brandy andalcohol; and the production of alco-hol from sugar cane molasses. Suchwastes are normally high in BODand low in suspended solids, withhigh total solids content and lowPH.

Distillery Wastes. At one plant thedesign loading is 0.75 lb. of BOD percu. yd. Dosage rate is 8.0 mgad withrecirculation ratio of 4 times rawwaste volume.

At another the raw waste aver-aged 685 mg/L BOD; the high-ratefilter loading is 1.1 lb. per cu. yd.and the dosage rate is 30.0 mgad. Arecirculation ratio of 11 to 1 pro-duced an over-all BOD reductionof 77 percent.

Yeast Factory Wastes. Experi-ments on yeast wastes by Dr. Wil-lem Rudolfs and E. H. Trubnick atRutgers University, included pre-digestion, sedimentation and theapplication of digester effluent totrickling filters operated in series

with recirculation. It was found thatloadings up to 11,000 to 12,000pounds of BOD per acre-foot perday could be applied to the filterswithout material deterioration ofthe normal 50 percent removal.

Slaughterhouse and Meat PackingWastes. These wastes vary widelyin BOD, ranging from 2200 to ashigh as 9100 mg/L. Treatment ontrickling filters should be precededby screens, sedimentation andmeans for removing grease. Sedi-mentation reduces the BOD 30 to35 percent. Low rate niters may re-move up to 80 percent of the re-maining BOD. In several cases two-stage filtration, with the primaryfilter so constructed that it may bewashed, followed by sedimentationand a second step of filtration hasbeen used with an overall plant re-duction of 94 percent of the BOD.

Phenolic Wastes. Wastewaterscontaining as much as 800 mg/Lphenol can be treated on tricklingfilters to reduce the phenol contentto less than 1 mg/L through care-ful control in acclimating the bio-logical growth to the toxic natureof the wastes. Dilution from otherwaste effluents can be used to makethe filter effluent acceptable for dis-charge.

Wastewaters containing othertoxic organics such as formaldehyde,organic acids and wastes from pe-troleum refining have also beensuccessfully treated.

Many hydrocarbons, particularlythose of high molecular weight, areresistant to biodegradability, andexperimentation with specific chem-icals is advisable.

Pulp and Paper. Some degree ofsuccess has been achieved with these

Courtesy Dorr-Oliver, Inc.• DESIGNED after extensive pilot plant tests, this high-rate trickling filter plantwas built by Shell Oil Co. for treatment of refinery wastewater prior to discharge.

Textile Wastes. These are treat-able with trickling niters, thoughplastic media is preferred. Efficien-cies up to 50 percent have been at-tained, which indicates their use aspreliminary treatment devices, fol-lowed by other biological devices.

Pharmaceutical Wastes. Wastesfrom plants producing penicillin,aureomycin, streptomycin and otherantibiotics have been found to re-spond to treatment on trickling fil-ters. The usual plant comprisesstorage and equalization tanks,grease flotation unit, aeration, sedi-mentation and high rate tricklingfilters followed by final sedimenta-tion. Filters are designed to handle4300 pounds of BOD per acre-foot,with 6-foot depth of stone, and anaverage recirculation ratio of 3 to 1.

wastes, using plastic media. Withloadings as high as 800 to 1,000 lbs.per cu. ft., 45 percent reduction ofBOD has been obtained. Their usein this industry appears to be pri-marily for roughing filters. Experi-ence with rock filters has been dis-appointing with rapid clogging ofmedia a principal problem.

Metal Finishing Wastes. Wastescontaining cyanide have been ap-plied to trickling filters but due tothe extremely low final concentra-tions permitted in streams, nitershave not been used in operatingplants. In many cases regulatoryagencies require a maximum of 1.0mg/L in final effluents; others re-quire 0.1 mg/L; and in one case theallowed maximum is "substantiallynone".

21

Design Considerations for Biological FiltersEDWARD C. ARCHER

andLLOYD R. ROBINSON. JR.

Mr. Archer is a Sanitary Engineerwith TLM Associates and Dr. Robin-son, formerly Associate Professor ofSanitary Engineering at MississippiState University is with Camp, Dresserand McKee.

SEVERAL mathematical modelsexist for predicting performance

of biological filters. All, at best, givean approximation of what can beexpected from an actual installa-tion. These empirical formulae givea good indication of the perform-ance of the biological process, al-though they are by no means exact.Field data discussed later in thisarticle will demonstrate the closecorrelation between calculated andactual results.

In the design of the biologicalprocess the engineer is confrontedwith two objectives. First, the proc-ess must stabilize the waste to apoint that it can flow into the re-ceiving stream without causing anuisance, and second, the processmust be as economical as possible.Until recently very little study hadbeen given to the cost of the variousfilter units.

In a study undertaken at Missis-sippi State University, the NationalResearch Council (NRC) formulae1

were analyzed using the Interna-tional Business Machine 1620 com-puter for the purpose of developingcharts to aid the engineer in thedesign of the biological filter proc-ess. Numerous variables enter intothe process and must be consideredby the designer. Some of these,such as strength of waste, volumeof flow, and condition of receivingstream are not normally controlledby the engineer. Therefore, to pro-vide a good design the engineermust balance the remaining varia-bles in an attempt to minimize costwhile providing the treatment re-quired. Two of the major items af-fecting cost are volume of mediaand recirculation. The design chartsdeveloped by computer analysis givethe designer a comparison betweenthese two variables. Knowing powerand material costs for a specificlocale, the engineer may then mini-mize cost for the process.

To provide design charts for atwo-stage biological filter process,the effect of various volume ratiosas well as various recirculationratios for each stage had to be con-sidered. With the aid of data fromthe computer and the original NRCstudy, the optimum condition wasdetermined and used as a basis forthe charts.

Field data were collected at theMississippi State University treat-ment plant to demonstrate the ef-fectiveness of the NRC formulae inpredicting performance. Five differ-ent operating schemes for the plantwere observed and compared withcalculated results.

Development of NRC FormulaeThe study entitled "Sewage Treat-

ment at Military Installations"2 wasconducted by a subcommittee onsewage treatment formed by theCommittee on Sanitary Engineering.The net result of the study was theformulation of equations to predictthe performance of filters. The datacollected were compiled, and a graphof efficiency versus filter loadingwas plotted. A curve of best fit wasapplied, and the resulting equationbecame known as the NRC formula.The second equation, used for two-stage operation, is simply a revisionof the first with the reduced treata-bility of the organic waste takeninto account. The equations are asfollows:

(1)

First or Single

1 + 0,

Second Stage:

01 + -

F —

(1

Stage:

1

,0085J-W-

1

.0085 \-yj~

-Ex\vF

1 + R

+ R/10)2

(2)

(3)

= Fractional efficiency of BODremoval of the first-stagefilter.

W = BOD loading of settled rawsewage in pounds per day.This does not include theBOD of the recirculatedwaste.

F = A factor reflecting increasedremoval due to recirculation.

R = Recirculation ratio on vol-ume of recirculated flow di-vided by the volume of in-fluent flow.

V = Volume of filter media inacre feet.

E2 = Fractional efficiency of BODremoval of the second stagefilter.

W = BOD loading applied to sec-ond stage process in poundsper day and does not includethe BOD of the recirculatedwaste.

Development of ChartsAs mentioned previously, the en-

gineer must consider several vari-ables when designing filters. Someof these variables have a direct ef-fect on the design but normally can-not be controlled by the engineer.The required overall efficiency oftihe plant must be determined fromthe conditions of the receivingstream, the volume of flow, and thestrength of the influent waste. Ef-fluent BOD delivered to the re-ceiving stream is of a major con-cern to federal, state, and localstream pollution control officialsand acceptable permissible valuesshould be determined in consulta-tion with them. Some of 1ihe varia-bles which the engineer may con-trol are the size of the units, re-circulation and flow diagram. Thefollowing is a list of variables thatmust be determined:

1) Amount of treatment required.2) The strength of BOD of the

influent sewage.3) The quantity of flow and the

hydraulic load of the filter.4) The volume and size of the

filter media.5) Recirculation.6) Temperature of the sewage.

These factors shall be briefly dis-cussed with emphasis on their ef-fect when using the NRC formulae.

The BOD load is the parameterused to determine the efficiency ofthe process when using the NRC

22

equations. Since the NRC study in-cluded only plants with 8,000pounds per acre-foot per day orless, this load is usually consideredthe upper limit of the equations.The suspended solids are not con-sidered in the equations but areassumed to be reduced with, andapproximately as, the BOD is re-duced. This assumption is based onthe results of the NRC study andis supported by field data collectedin the Mississippi State study.

Although the actual volume offlow does not enter into the NRCequations, it is important in the de-sign of the process because it de-termines the organic and hydraulicloads applied to the filter. Maximumhydraulic loads, established by stateand local officials, are usually fourmillion gallons per acre per day(mgad) for standard-rate filters and30 mgad for high-rate filters.

The volume of the niter media,adjusted for recirculation, is thecontrolling parameter for the re-quired efficiency. The conditions ofdesign are usually such that thedesigner knows the influent BODand the efficiency desired and mustdetermine the recirculation and vol-ume of media required. The designcharts of this study are based onthis condition. The size of the mediais of importance for it determinesthe void spaces available for airflow, waste flow and slime growth.General media size is from Wz to 4inches with the upper ranges usedfor high-rate filters and the lowerranges for standard-rate filters.3

Recirculation is a very importantconsideration because it can im-prove treatment without increasingthe filter volume. Recirculation isseldom used on standard-rate filtersand almost always used on high-rate filters.4 Recirculation has sev-eral advantages. These include:

1) Return of the organic wastefor more than one pass through thefilter.

2) Dilution of incoming waste fora reduced load to the filter. Thishelps to achieve a more uniformload onto the filter.

3) Continuous seeding of the rawsewage and filter with active micro-organisms for an increased rate ofbiochemical stabilization. Gallarand Gotaas5 have conducted a sta-tistical optimization analysis thatdemonstrated recirculation above4:1 is generally uneconomical. Ifrecirculation below 4:1 saves suf-ficient volume of media to offsetthe additional cost and operation ofthe units, then recirculation shouldbe used. A factor "F" is incorpo-rated into the NRC equations for

increased efficiency due to recircu-lation. Equation (3) describes "F",and it simply indicates the averagenumber of times the waste passesthrough the filter.

Temperature is important in anybiological process. It was originallythought that air temperature had aprofound effect on the biologicalfilter, but experience has shownthat the temperature of the sewageis the influential factor.6 Shouldextreme temperature variations pre-vail in an area under consideration,the NRC equations should be usedwith caution.

In the study of the NRC equa-tions, it became apparent that thevolume of the media was varyingdirectly with the BOD and flow ofthe influent sewage when the ef-ficiency and recirculation ratio wereconstant. The linear relationshipwas utilized in the development ofunit charts for design. Because ofthis linear relationship the volumeof filter media required varies di-rectly with the load applied to thefilter when the efficiency and recir-culation are constant. As an exam-ple, if the flow remains constant andthe BOD doubles, then the load tothe filter doubles and the filter vol-ume doubles. The same would occurif the BOD remained constant andthe flow doubles. Similarly, if theflow and BOD double then the loadto the filter and the volume of thefilter increase four times.

In order to determine the opti-mum volume distribution betweenfilters in a two-stage system, theNRC equations were used to studyeffects that changing that ratiowould have on overall efficiency.Using a fixed total volume of 0.16acre-foot, a flow of 100 gpm, and

a BOD load of 100 mg/L, the effi-ciencies were calculated for a seriesof volume distributions. The results,plotted in Figure 1, show that themaximum efficiency occurs at avolume of 0.07 acre-foot in the firstfilter and 0.09 acre-foot in thesecond filter. However, it is obviousthat the efficiency at the pointwhere the volumes are equally di-vided is only very slightly belowthe maximum efficiency. Therefore,the most economical design withoptimum efficiency would be withthe volumes divided equally be-tween the two filters and the re-circulation equal for both stages.Providing recirculation is constant,a volume of media divided betweentwo stages will yield a higher ef-ficiency than a single stage withan equal total volume. Therefore,when high efficiencies are required,the two-stage filter is normally uti-lized. In the medium range of ef-ficiencies the engineer must com-pare the increase in cost of mediavolume for a single-stage filter withthe increase in cost of multiple unitsto arrive at a minimum cost.

When recirculation is proposedfor a filter system, direct recircu-lation versus recirculation throughthe clarifiers may be considered.Recirculation through a clarifier in-creases the total flow and conse-quently, will increase the size of theunit. Direct recirculation increasesonly the flow through the filtersand does not alter the required sizeof the clarifiers. Many recirculationstudies have been conducted whichindicate that the quantity and notthe system of recirculation is theimportant factor.3 Culp7 concludedfrom a study of a field installationthat direct recirculation is as bene-

FIGURE 1. Optimization curve for two-stage filter with a fixed total volume.

Uu•oV

<ra 92bmo

c 88S

I

i 84UJ

uu.li_

I

.16 .14 .12 .10 .08 .06 .04 .02 .0FIRST STAGE VOLUME - Acre Feet

.0 .02 .04 .06 .08 .10 .12 .14 .16

SECOND STAGE VOLUME - Acre Feet

23

ficial as recirculation from behindthe clarifiers.

Design ChartsIn most design, problems the in-

fluent strength is calculated fromavailable data, and the maximumeffluent strength is promulgated byfederal, state, and local authorities.With these two parameters known,the required efficiency of the plantmay be determined. Since calcu-lated results using the NRC formu-lae include the filter and subsequentsettling, the primary clarifier is theonly other unit entering into theoverall plant efficiency. Dependingupon its design, the primary clari-fier can remove 20 to 40 percent ofthe BOD applied. A general as-sumption used by many engineersis that 35 percent of the appliedBOD is removed. After assuminga value for removal in the primarystage of the process, the load to thefilter can be determined, and thevolume of media, corrected for re-circulation, can be calculated.

The design procedure describedabove was used as a format for thedevelopment of the charts. Figures2 and 3 are charts for the designof a single-stage plant and Figures4 and 5 are for a two-stage plant

Figures 2 and 4 give the filter vol-ume required for any desired effi-ciency of the filter stages and arebased on the load actually appliedto the filters. Figures 3 and 5 givethe filter volume required for anydesired efficiency of the entire plantand are based on the raw sewageload applied to the plant. Incorpo-rated in Figures 3 and 5 is the as-sumption that 35 percent BOD re-moval is accomplished by the pri-mary clarifier.

The charts are based on a BODload of 100 mg/L and a flow of 100gpm. Because of the linear relation-ship between load and volume, asimple conversion can be applied toarrive at a specific condition. If, forexample, the BOD of the influentis 200 mg/L and the flow is 500gpm, then one of the charts shouldbe used to obtain a value whichmust be multiplied by (200 -*• 100) x(500 -*-100) or 10. It should also benoted that the charts for the two-stage process are based on an equalvolume existing in each filter andequal recirculation existing for eachstage.

The charts were plotted using theoutput from computer programs.Basically the approach was verysimple. Using the NRC equations,

a set efficiency, a unit flow, a unitBOD and a very small volume in-crement, the computer made repeti-tious calculations until the total vol-ume yielded the required efficiency.The computer typed out the volumerequired for the conditions set. Asemi-log plot was then made of ef-ficiency versus total volume. Theresulting smooth curve plots arethe design charts.

Discuss/on

In the design of the biological fil-ter process using the NRC equa-tions, the only three parameters theengineer may vary to achieve therequired efficiency with minimumcost are the volume of media, therecirculation ratio and the numberof stages. The charts produced bythis study give the designer the re-quired volumes for efficienciesvarying from 75 to 95 percent forplant design and 40 to 90 percentfor the design of the stages. Therecirculation ratio was varied from0 to 4:1, and charts are providedfor both single and two-stage opera-tion. With the required efficiencyknown, the decrease in volume withthe increase in recirculation can beeasily determined using the designcharts. Knowing the construction

.05

<UJS13_lo>

o

.002

/ ,

/ /

//

—SINGLE - STAGE FILTER

Load to Filter:

BOD = 100 mg/LFlow= lOOgpm

//

/ /

2T

// ////

/

*//i/ /

//

/ /

/ / ./ /

////

/ V N 0

^-4-1 ^

Figure

I

t-ffftftJ/i

ne

•>

2

f

40 50 60 70 80EFFICIENCY - Percent of BOD Reduction

by Filter and Final Clar i f ier

90

.5

reet

V

OL

UM

E-A

cr

b

>

.01

—SINGLE - STAGE FILTER PLANT

Lood to Plant:BOD= 100 ma/L

•

Y

Flow = lOOgpm

/

/

i///4// ///

/A////"-4:1

//

/ i

r /

/ / '

/N//

3:1

V // y//t

y(/// ^

/

//

0 5:1

III'Ik'IfIII

ne

igure 3 •

75 80 85 90 95EFFICIENCY - Percent of BOD Reduction

by Entire Plant

24

and electrical power cost for theparticular locale the engineer canthen compare the cost of recircula-tion versus decrease in volume.These two parameters constitute themajor costs of the filter and aid theengineer in optimizing cost andefficiency. The charts do not give thedesigner a cut and dried solutionfor a particular condition. They sim-ply compare two of the major fac-tors affecting cost and leave theactual optimization up to the engi-neer.

In using the charts for the designof a two-stage plant, it must beunderstood that the curves arebased on the total volume beingequally divided between the two fil-ters and the recirculation ratio beingequal for both stages. However, inthe study for the development ofthe charts it became evident thatwithin practical limits the volumeratio between the units has a verylimited effect on the overall effi-ciency.

If the volume of media in the firstfilter is approximately four timesgreater than the volume in thesecond filter, the decrease in effi-ciency from the optimum is slightlygreater than one percent Similarly,if the second filter is approximately

four times larger than the first, theefficiency decreases slightly less thanone percent from the optimum. Thisis an important factor when increas-ing the capacity of an existing plantbecause the additional filter mayvary from one-fourth to four timesthe size of the original filter andstill be very close to the optimumdesign.

Field ObservationsThe field observations were per-

formed to demonstrate the effec-tiveness of the NRC formulae andthe design charts to predict the ef-ficiency of both a single-stage and atwo-stage filter plant under variousoperating conditions. This is by nomeans a proof of the NRC formulaebut rather a demonstration of thevalidity of the formulae and designcharts for the operating conditionsencountered.

The field data were collected atthe Mississippi State UniversityTreatment Plant. The plant is a two-stage filter plant with direct recir-culation around each filter. The fil-ters are 50 ft. in diameter and havean average media depth of 4 ftwith media size ranging from 2 toZVz in. The primary, intermediateand final clarifiers are designed on

the basis of the influent flow plusrecirculation because provisionswere made for recirculated flow topass through any of the clarifiers.

Three sampling periods of 24hours' duration were made on theoperation of the plant The firstrun was made with no recircula-tion. The second run was made witha 1.1 recirculation ratio. The lastrun was identical to the first ex-cept that the intermediate clarifierwas bypassed. In Run III recircu-lation was not employed, and theeffluent from the first filter passeddirectly onto the secondary filter.During the first two runs the inter-mediate clarifier was in operation;therefore, these field data could beused to study both single and two-stage filter operation.

Prior to each run the plant wasallowed to operate under the spe-cific conditions for a minimum offour weeks in an attempt to reacha steady state. Samples were col-lected every hour on the hour forthe 24-hour period with a total of25 samples being taken. The read-ing on the cumulative flowmeterof the plant was recorded at thebeginning and end of the run todetermine the total number of gal-lons passing through the plant.

.05

.01

.001

TWO-STAGE FILTER

Load to First Stage Filters.

BOD >= 100 mg/LFlow* 100 gpm M

'40 50 60 70 80 90

EFFICIENCY • Percent of BOD Reductionby Filter ond Final Clarif ier

TWO-STAGE FILTER PLANT

Lood to Plant :

BOD= I00 mg / L

Flow = I00 gpm

.00175 80 85 90 95

EFFICIENCY - Percent of BOD Reductionby Entire Plant

25

Table

Removal ByPrimary

Run ClarifierField Data

Percent Removal of BOD

1 30.0II 36.8

III 27.9

1—Comparison of

Removal By FirstFilter and Inter. Clarifier

FieldData

49.768.535.8*

Percent Removal of Suspended Solids

1 67.3 23.0II 50.0 24.8

III 51.4

*Filter only.

Calc.

63.368.760.4

Field Observations

Overall-EfficiencySingle-Stage PlantFieldData

64.980.1

75.163.3

Calc.

74.480.2

with Calculated

Removal BySecond Filter

and FinalClarifier

FieldData Calc.

44.4 51.169.3 55.161.0 49.0

19.134.2

Results

Overall EfficiencyTwo-Stage Filter PlantFieldData

80.493.982.0

83.675.986.0

Calc.

87.491.185.4

ResultsThe results of the field observa-

tions are compared, in Table 1, withthe predicted value using the NRCformulae and the design charts. Thegood coorelation demonstrated pro-vides an additional verification ofthe NRC formulae and the designcharts, as well, for the prediction offilter efficiencies.

When using the NRC formulae,it is assumed that the suspendedsolids will be reduced coincidentallywith the BOD. Except for Run IIsuspended solids removal agreedclosely with BOD removal in thefield observations. This tends todemonstrate that the assumption ofthe NRC report is reasonably sound.

Run I and Run III were com-pared to observe the effect of theintermediate clarifier on the effi-ciency of the plant. The results in-dicate that this unit is unnecessaryand supports a similar conclusion ofSorrels and Zeller.8

In

Conclusionsdesigning a biological filter

using the NRC equations, the fol-lowing can be concluded.

• The BOD of the influent, the

volume of flow and the BOD of theeffluent are the most important itemsto be determined.

*• Up to a maximum ratio of 4:1,recirculation can greatly reduce therequired filter volume for a givenefficiency.

• Within reasonable limits, thevolume ratio between the stages ofa two-stage filter plant has verylittle effect upon the efficiency.

*• When increasing the capacityof an existing plant, the additionalfilter may have as much as fourtimes the volume in the . existingfilter and remain close to an opti-mum design.

*• In addition to providing arapid and accurate method of de-termining the required volume offilter media for any set of designconditions, the unit curves presentedin this article make available animmediate visual comparison be-tween media volumes and recircula-tion ratios for any given efficiencies.

The following were demonstratedin the field observations:

• Prediction of efficiency usingthe NRC equations agreed closelywith field performance.

• Rec i rcu la t ion definitely im-proved the overall plant efficiency.

• Removal of the intermediateclarifier had no effect on the over-all efficiency of the plant.

References1. Glass, A. C. and Jenkins, K. H.,

"Statistical Summary of 1962 Inven-tory of Municipal Waste Facilities inthe United States." Public HealthService, Washington, D. C. (1964).

2 , "Sewage Treatment atMilitary Installations," National Re-search Council, Subcommittee Re-port, Sewage Works Journal, 18, 791,(1946).

3 , Sewage Treatment PlantDesign, Manual of Practice No. 8,Water Pollution Control Federation,Washington, D. C. (1959).

4. Lohmeyer, G. T., "Trickling FilterDesign and Operation," Sewage andIndustrial Wastes, 29, 1, 89, (1957).

5. Gallar, W. S. and Gotaas, H. B. "Op-timization Analysis for BiologicalFilter Design," Journal of the Sani-tary Engineering Division, ASCE, 92,SA1, 163, (Feb., 1966).

6. Eckenfelder, W. W. and O'Connor,D. J., Biological Waste Treatment,Pergamon Press, New York, (1961).

7. Culp, Gordon L., "Direct Recircula-tion of High-Rate Trickling FilterEffluent," Journal of Water Pollu-tion Federation, 35, 6, 742 (1963).

8. Sorrels, J. H. and Zeller, P. J. A.,"Two Stage Trickling Filter Per-formance." Sewage and IndustrialWastes, 28, 8, 943, (1956).

26