THE BELMONT FILTRATION | r ~ WORKS: BY JOHN W. HILL Reprinted from the JOURNAL OF THE FRANKLIN INSTITUTE, January-March, 1904. PHILADELPHIA Reprinted from the Journal of the Franklin Institute, January-March, Igo4. Meebo ANTS LN INSEE: Stated Meeting, held Wednesday, October 21, 7903. The Belmont Filtration Works. av JOELN EW ERLE: Chief Engineer Bureau of Filtration, Department of Public Works, Philadelphia. The Belmont works for filtering the water supply of West Philadelphia, embracing the 24th, 27th, 34th and goth Wards, containing at the present timea population estimated at 170,000, are located on Belmont Avenue near the city line. The tract of land taken for these works contains 60°57 acres, lying partly north and partly south of Ford Road, between Belmont and Monument Avenues. Upon the land located north of Ford Road are placed the sedimentation and clear water basins, and on the land south of Ford Road are placed the plain sand filters and preliminary filters. Enough land south of the plain sand filters is unoccupied and reserved for the construction in the future of eight more filters of the same effective sand area as those now built. The agitation of an improved water supply for West Philadelphia was a strong factor in shaping and promoting the works which are now being constructed under the ordi- r|| 2 nances of Councils providing for the improvement, exten- sion and filtration of the water supply. As early as June, 1898, Councils passed an ordinance appropriating $3,200,000 for the improvement of the water 562000. ABMS MME EE Ke 520000 | 500000 BR toes ce ee ne ee ap Oe eine | | | | Ee fa! 480000 een ES SAPS Re eS | Pane ae res i 440000|____| ina al 4 } or 4 ae! + sael) | 7 | aE : EG «o000|_| _| | Population Diagram om ii oh, ay : <00000|__| peal) West Philadelphia f 360000|__ + 4 | | | + + + see b eee HH | 360000} | jell eee le n Lauation for curve snowing rate of increase of population yn= 7X 1+3 5833 decades (550 106 ut 34Q000|. S| | | Wote fate of increase of population ml zal: le i | 4 £ taken from percentage curve ; 8 ] CaN | ail lads S Ve q B00 C00 SaaS a ee ee eee —t + f 280000, ete ub he ey ‘ | y £27056 | jal saoog_| |_| alee ka Vo bo, pA ae 7860 7870 7860 7890 7909 7910 1920 1930 1940 7950 Years Fic. 1.—Diagram showing population of West Philadelphia. supply, and mainly through the exertions of Hon. Edward W. Patton, member of Councils from the 27th Ward, a pro- vision was incorporated in the ordinance that at least $500,000 of this amount should be applied to the improve- 3 ment of the Belmont Works. The actual outlay for West Philadelphia, including land and the extension of the pipe- distribution system, will be over $4,000,000. BRORUTBASEIOING West Philadelphia, as shown by the census enumerations from 1860 to 1900, inclusive, has been growing at a greater rate than that portion of the city lying between the Dela- ware and Schuylkill Rivers, and as a preliminary step in determining the capacity of the works of filtration required for the present and future supply of this section of the city, a careful study of the past and prospective growth of popu- lation was made. By inspection of the diagram (/7g. 7) it will be seen that from a population of 23,738 in 1860, it has grown to 148,371 in 1900, and by plotting the mean annual rate of growth for each census decade, and drawing a smooth curve through the points on the diagram, the smoothed rates of growth from 1860 to 1950, stated as the -mean annual increase, are as follows: RemCent: e189 oad ye oA y oh Berga or Batt ye eae a Sa PR A eR i per: TSS JR een tee at AW ee MN ca i oS sl aa hee ‘aces Meh e s 6°03 ECE) ry ae Meet et Cee iar hts UR ik A os wgls, ee pile e ante aU 4°95 ESOC Melt cy aM aCe PC Ont vey Veer pea) CS yin MLK, eos Uk seals 4°20 LOCO MEIN Th Sante Mh ter on Ree RAS ek Maes Sere Gite, co cella 3°64 LO Omran Be ee ee Reman Pings nes ewe ere fees Mame ys ert ye ol otted 3°22 Hepler doce ten doo hes Were yl Maker led OS Se Oo aCe RM TIE Ala te 2°88 LO) CVG ee ce Meee OR WE MCU N McG Nen, y Pavia teaul age eel” es FeDc yee ote 2°63 ROAO Diets etic Coomassie ase foe teens Wd oved Po dusels ord oye bela cath aie 2°39 a Xah Soc, Bee ewe &; Caper ee ine Ete ee emt nee Sieg or eee 2°20 Corresponding to the following future populations: EQI Oar te ag CR ees) Bae Ne ed oe) La Taiaiead oe, TAs U's ay) Oren Re Uh me 203,699 TG 20 that ee ak ay Pe QB SM, DAL SORE STAT Vee ore oak RCA SL ea 270,582 Reiley cg CMe aus a aie ce BEA Cen ys areee MIO =e ec me 350, 106 1 GAO Mr MMe tetera ai gery corr! ene Poe BS cae, Ws To wee ts 443,373 C50) See em ee MMAR aren ans Oh me erat cal eR ohy os co. wahila, Pots 551,162 Upon a basis of 150 gallons of water per capita per day the consumption of water for each ten years of the next five decades should run about as follows: "SIIOAJaSoI pu Si19}[J JUOM[IG—'z “OI cyl WE rennin ae cot oot oO o “As 08M DIOS NISVS W3LVM G3U3.L14 ONY ! SUILHd 40 NOLLVTIVISNI LSUId “BIOANISIY “INVid SAS LNOWTIE 30 NYId ee ‘oo 8434 rn: SA ecto Mae OX ULEOLORR RMT RNE RI HF HS ‘ AaB EL WME eR Gg see Gallons Per Day. EMSRS! CO A meade ao bout tg av. BS ip Greet Bae enn a oe 30,554,800 TO 2G teem ecient del a th ae A alias! vagal 2, « 40,637,300 TORO Pears wee Fe eases oe fat ee es torr, Ketan ees ot Aes Rie 52,515,900 TA Ogee ema ener eer ae ci ln terse ters oo. Wicea a oi sty nN e al ey 6 66,506,c00 ee Gee Se fa EY, Tg Ue Ga es de ek eet oar 82,674,300 The present consumption of West Philadelphia is, as an average, 185 gallons per capita per day, and if steps are not taken to reduce the unnecessary and avoidable waste of water, the consumption by 1950 will reach over 100,000,000 gallons per day, or about the ultimate capacity of the Bel- mont works. In the report of the Expert Commission appointed by Mayor Samuel H. Ashbridge, May, 1899, it is stated that the least flow of the Schuylkill River has been shown to be about 150,000,c0o0 gallons in twenty-four hours, and that it was thought prudent to not exceed this quantity as the maximum daily consumption of water from this source. Apportioning this consumption of Schuylkill River water between West Philadelphia and Roxborough, it has been assumed that if the present rate of consumption is main- Baie Cm rReSeu CWOeMGisttiCts-wwille Treqitire respectively 100,000,000 gallons and 50,000,000 gallons per day within the next fifty years—the entire allowable quantity which, accord- ing to the experts’ report, should be drawn daily from the Schuylkill River. PRINCIPAL, DETAILS. The Belmont Works consist of two subsiding basins, each containing at flow line about 36,000,000 gallons of water, which, at the present rate of consumption, represents 2°40 days’ sedimentation of the water before it is drawn from the basins; eighteen plain sand filters, in part modeled after the filters at Berlin, Warsaw and St. Petersburg, and in part after the newer filters at Hamburg; a system of prelimi- nary filters adapted at a rate of 80,000,000 gallons per acre per day, to deal with 40,000,000 gallons of water daily, and a clear water basin, 15 feet in depth, with a capacity of 16,500,000 gallons; eight hopper sand washers, patterned after the washers in use at the Hamburg filters; an admin- ‘SIIOAIOSOI UOl}VJUSMIpes jo JUOMTYULqMIS YSsno1y} UoIWeSGS—'e “Oy aos so. te ¢ a ie ioe LEROWTHT 9 ONIN? DMIMOHS SHORTS Be es YOUICTY INOWTIG £ S297 yy > oe. A eX. 2S Bie t BE Le mean RIS Pape - Sp na Same 40 fuse engny Prt) Oe ‘HOISIAIp }S¥Y“AJOAIOS9I U01}BjDOUTIpeaS— ‘Ph ‘oT 8 istration building and pumping station; centrifugal pump- ing machinery to supply wash water to the preliminary fil- ters; direct acting plunger pumps to supply water under pressure to the sand washers; steam boilers for power and heating the buildings, and an electric lighting equipment. Considering the chief features of the works, zg. 2 shows the location of the sedimentation reservoir, preliminary fil- ters, plain sand filters, the clear water basin and the admin- istration building. Belmont Avenue, as will be noticed from the meridian symbol, has a nearly north and south direction. With this explanation it will be seen that the sedimentation reservoir occupies the land bounded on the west by Belmont Avenue, on the north by City Avenue and Overbrook Avenue, and on the south by Ford Road. The preliminary filters are located at the intersection of Belmont Avenue and Ford Road. The plain sand filters are located along the east side of Belmont Avenue, the south side of Ford Road, and the west side of Monument Avenue, and the clear water basin at the intersection of Ford Road and Monument Avenue. The unoccupied land, which is shown with contours south of the plain sand filters, is reserved for an addition of eight filters of the same general dimensions as those shown. The small rectangles, of which there are eight, shown in the court between the filters, indicate the location of the sand washers. Between the preliminary filters and the group of six filters on Belmont Avenue is placed the administration building and boiler and pump rooms. SEDIMENTATION RESERVOIR. (Contract No. 16.) The sedimentation reservoir consists of two divisions, or basins, each 25 feet deep, measured at the flow line, eleva- tion, 279'00 C. D., and 29 feet deep from the top of embank- ment, elevation 283:°00 C. D. The area of the east division at the flow line is 5°43 acres, and at the floor line 3°47 acres. The area of the west division at the flow line is 5°33 acres, and at.the floorsline 3:0272acres.. “hhesinsidesandmoutside slopes are, at all points around the basins, two horizontal to one vertical. 9 Top width of embankment, 18 feet; width at toe of inner slope, 134 feet. The east division was constructed about equally of cut and embankment, while the west division, excepting the embankment along Ford Road, was nearly all in excavation, some of which was quartzlike trap rock. The materials of excavation consisted of clay, micaceous rock, sand, gravel and hard rock, the latter requiring drilling and blasting for its removal. In making embankments, the best materials of excava- tion were placed next the inner slope, and all materials were rolled in thin layers. When the excavation in rock discov- ered fissures, these were filled with grout, and the irregular surfaces of the rock in the floors, and slopes leveled up to sub-grade partly with concrete and partly with clay puddle. In preparing the ground for rolled embankment for the sedimentation reservoir the top soil was first thoroughly stripped off and the inclined ground stepped in horizontal terraces. The reservoir embankments and the fill under the filters on Monument Avenue were rolled with four 25-horse-power traction engines weighing 10 tons each, or 3,300 pounds per foot width of roller wheels, two 18-horse-power traction engines weighing approximately 8 tons each, or 3,000 pounds per foot width of roller wheels, and one 10-ton traction roller, manufactured by the Julius Scholl Company of New York. All rollers were grooved. In the original plan of the reservoir, as shown in NN oe the inner slopes were constructed with a berm at an eleva- tion of 12°7 teet above the finished floors of the basins, and the upper slope set back a distance of five feet to receive a granite block paving to be set dry, and backed with broken stone. This was intended asa protection to the puddle lin- ing from frost, but further consideration of this feature of the construction raised doubt of its utility, and it was omitted in finishing the basins. Each division of the reser- voir is lined on the floor and slopes with 18 inches of clay puddle, on which is placed a 6-inch course of concrete. On the concrete to within a vertical depth of 10 feet from the water line is placed a 3-inch mixture of asphalt, asphaltic IO mastic and grit. /zg. 5 shows the height on the slopes of the asphalt lining. In the division wall between the two basins, which is constructed partly in excavation and partly of embankment, are placed two equalizing or pass pipes, which, when the valves are opened, maintain a uniform elevation of water surface in both divisions of the reservoir. Each of these pipes is supplied with a floating inlet pipe, which receives the water at the surface of the reservoir and conducts it to the pass pipe and thence to the opposite side of the division wall, where itis discharged at the bottom of the basin. The equalizing pipes, and the influent and effluent pipes, are placed in trenches provided with concrete cut-off walls, the Spaces between which, under, around and over the pipes, are packed with well-rammed puddle. fig. 4 shows the 48-inch diameter influent pipe in the east division of the reservoir, into which is connected, at theangle near the centerof the basin, the equalizing pipe from the west division of the reservoir. The influent pipe crosses the basin diagonally from the southwest to the northeast cor- ner, and terminates in a set of fourteen 48-inch tees and one 48-inch elbow, with the branches turned slightly upward from the horizontal, as shown. ‘The line of pipe and special castings are supported on low brick piers set on the asphalt lining of the floor. By means of the arrangement of shown pipes, the water from the Schuylkill River at the Belmont Pumping Station is passed diagonally from end to end and from the bottom to the surface of the water in each division of the sedimentation reservoir. It is thus hoped that currents having a tendency to carry the heavier suspended matter out of the reservoir, will be avoided. In the line of each influent pipe is placed a 48-inch check valve, made by the Ludlow Valve Company of Troy, New York. Each of these check-valves weighs about 32,000 pounds. The subsided water is conducted from either division of the reservoir to the screen-chamber, and thence to the pre- liminary filtersthrough floating pipes, one of which is shown lying on the floor tothe left of the effluent chamber in ‘sodid Suljeoy pure 1squieyqo use190G—'S “oy Stl thsumandlitcarcatiod LAZO ihnm_,démuaeeen ee amanemenneens. 9 12 fig. 5. This pipe, 48 inches diameter, is made of galvan- ized sheet-iron, attached at the lower end to a 48-inch swiv- eling tee, mounted in bronze bearings, and at the upper end by flexible connections to a float which consists of a galvanized sheet-iron, water-tight cylindrical barrel, which allows the mouth of the pipe to be submerged a few feet below the surface of the water. Should the floating pipe for any reason fail to act, water can be drawn through either of the three sluice gates shown Fic. 6.—Sedinientation reservoir. Gate-house. on the inside face of the effluent chamber, each being pro- vided with a compound geared gate-stand on the top of the effluent chamber. In the gate-house, /zg. 6, are placed eleven 48-inch diam- eter Ludlow stop-valves, which control the flow of water into and out of the reservoir, each of which is provided with a compound geared gate-stand and extension stem to operate from the floor of the house. The gate-house, when the works are completed, will have a frontage on the north property line of Ford Road sufficiently elevated above the 13 grade of the street to make a neat and pleasing feature of the works. The plan of the influent and effluent pipes where they pass through the gate-house, is such that the raw water may be delivered entirely into one division of the reservoir and drawn from the other; or may be delivered into and drawn from each division at the same time; or one basin can be cut out of service entirely; or one of the two supply pipes entering the gate-house from the left, will supply into both or either division of the reservoir, and likewise the sub- sided water can be directed into either of the two lines of 48-inch effluent pipes shown at the bottom of /7g. 7. The two lines of rising pipe which bring the water from the Belmont Pumping Station to the gate-house are each 36 inches in diameter. In the upper part of the gate-house is constructed a room for the reservoir watchman, access to which is had by means of a spiral iron staircase. All the joints in the gate-house, excepting the last joint in each line of pipe, are made with flanges and bolts and gaskets. PRELIMINARY FILTERS. (Contract No. 38.) Fig. § shows in place the first installation of prelimi- nary filters, placed in the center of the ground lying between Ford Road on the north and plain sand filter No. 6 on the south. South of the first installation of filters, ground is reserved for an addition of 25,000,000 gallons daily capacity, and NOLL MOM Le: tirst set OLehiters, space 31s left, fora further addition of 30,000,000 gallons daily capacity, or an ultimate aggregate daily capacity of 95,000,000 gallons. In the arrangement of the 48-inch main pipes to conduct the water to and from the preliminary filters, provision has been made for the necessary influent and effluent connec- tions to the future installations, and to admit of these con- nections being made without interrupting the service of the first set of preliminary filters. Connection of the two 36-inch rising mains in Belmont “OSNOY-3}88 YSNOAY} UOIDIG “4LOAJOSad UOLWeJUaMIpasS—"L “oy MPINIONT LFIMI~TTIL M NNO Pod [= . NOMLVELUY SO NVIFNG LATIF -TIVIS YIGNVHI FLVO YIOAYISTY LNOW IIE sem gets I bee ‘ f we BPA, PUES LS BE . yore lee Bauls 9 baAEE prea Ti NOILITS ; 3 ¢ ¢ FE eb mo DOR GS BUI US RI Hodge ‘siaz[y Areurmtiyperg ‘sedid juanpje pue yuenguy—'sg ‘O14 SS eee ae i c ° / — © PBNIT TIDE NOILYALEINLIVOE ONPSHILTUS TUNES ONG AXBNM TILES NWSW NYGLLBLNFIIOIS YO CN OLITNNODI AA \L SMBLTS LNONTIE ) K PPS S TE IS ste aaa aoa PIS OME A AeA, Cb Pe Se ersten incer ns ig CMOS Hed PPE Ee é Pd ; q i 7 : { i {enact y 8 H i é f : % \ i / f ‘ ‘ | x é og 4 Ris ‘ & *. Hos / ATWO SNOT IPD NUTT O8 ALLA MI, eS ee ; n FEATS HPLIM ATIF «=| FON ee hey / | 3 N | KUBO SROTIYO NOITUNGE ALIBad | : | é “ $3 ig : yl x 3 it / Y Sik ¢ CHILI S : ! BON HILTIS LET Ses N x = ft . ee 4 SHILIAS i \" i 3 RY k % : t / x | Ri ASUNINITI MA ‘ | fs s eT } AYUNIWS TF cf ‘ i ve 8 i : 5 ; ‘ t eo a i i / PHILOS ro 4 IX ALA 4 Vs : Pe : ; } x j / / 3 j & e i 7 i q : j ” a 7 é { J ; — Pssst OO acBlEC f ee eee ees Se Dottincsessesse sta tausuasteutastaces otamamasees wil cos RRR PONE Ve . ses euesias se EP See omens ua NaN Ye S ra { 3 7 $ £ $ i cnn | ORNS PLOT ey CU OMEES OLS ELE OLY VB HODES? BL 43 BARD POLE BE ee ne x c sie Fodelisindnsi SPesSgeoe eS Se eet —_ fs Ro BYE OEE LUM GY FLOUR Od gf HERS ONL OPI PUNE Bia TANTAY ‘sioyy Areutuntpeid jo uetq—'6 ‘oI Aa i KOR é PP acon cto scan. isi hase 04 ORLA S LBYS SLE AP RPS i PUSLES BO TRL AE #4 SAQLL) DAOOMLG RLS LOO RS ey ity Ct agen ee oe Os SOLIS. 17 Avenue with the main influent pipe of the preliminary filters, is also made to admit of the pumping of water direct to the filters at any time when it might be necessary to take the sedimentation reservoir temporarily out of service Likewise, if for any reason it might become necessary to take the preliminary filters temporarily out of service pro- vision has been made to draw water direct from the subsiding basins to the plain sand filters. As a measure of economy in the operation of the works, and to attain the highest possible efficiency in filtration of the water, the system of operation should consist of limited subsidence in the sedimentation reservoir, preliminary fil- tration, and final filtration. The preliminary filters, as planned by the Bureau of Filtration, /zgs. 9 and so, consist of twenty concrete tanks each 60 feet long, 20 feet wide, 8 feet 6 inches deep. In the bottom -ofecachatank-isefirsty placed 12 inches of: oravel; ranging in size from 24 inches to + inch in diameters, and above this a layer of 30 inches of coarse sand, consisting of erains which will pass a No. 6 sieve and be retained on a No. 30 sieve. Such sand will be coarser and of more nearly uniform grain than that used in the plain sand filters. The water from the subsiding basins is introduced at the top of the filter and percolates at a high rate downwards through the bed of coarse sand to the layer of underdrain gravel at the bottom, through which it flows in many streams tora mainecollector placed) in» the center of the filter, and thence out of the filter over a measuring and regulating weir to an open duct built of concrete, which in turn conducts the rough filtered water to the 48-inch cast-iron supply pipe of the plain sand filters. At the level of the gravel under-drains is placed a system of wash pipes consisting of two lines of 12-inch main pipe, from each of which 2-inch branch pipes placed on 8-inch cen- ters extend right and left under the sand bed to the opposite main pipe and side walls of the tank. The main pipes and branch pipes are perforated on the upper side. Water under a head of fifty feet will be used to wash the sand bed by reverse current, or rather by forcing the wash water from 2| ‘s19}[y Areururpead yYsno1y} UOlJVeS—O! “OM SOD LBL JES 1G OE LIBERALS Gf GO OIL cee ene ed pepe XE opps Su 28g SBR g PMMAY Bhd LATA ie we se ; DRG Abit, DOGRGS WAGIIG: & sapsioomas RIGS ics PF Bibs > =} 2 a 2 s > LLOQ SABER LOLOL OGRE PBB LE SINE SSE) pe PIR AOT SR GIRS OSG BO GIOS GEO IIE ORAS PMOL. een re eee GROOK [9 below upwards in the same manner as in several types of the mechanical filter, excepting that the wash-water pipes are not effluent pipes, as they are in the mechanical filter when operating under normal conditions. Overflow troughs are placed around the filter tank at an elevation about 12 inches above the level of the sand bed, adjustable as to height. These troughs receive the wash water and conduct it out of the filter to a waste gullet of concrete placed in front of the filters. From an experience of eighteen months with an expert- Menta lminecwaniCaleiuitermrortecem diameters aratic, spring Garden Testing Station, in use for part of the time without agitation of the sand bed with rakes or by other means, it is thought that mechanical agitation of the sand bed of the preliminary filters while washing is not essential, and that a washing daily, or as often as the loss of head in the op- eration of the filter may show to be necessary, with the occa- sional removal from the tank and washing of the sand in hopper washers, will restore the sand bed to its normal con- dition at less cost of operation and with less chance of inter- rupting the work than by the use of mechanical agitating devices, or with agitation by means of compressed air. The principles of construction and operation of the pre- liminary filters are essentially the same as that of the plain sand filters, excepting the filtering materials are coarser and the sand bed will not be scraped periodically, but washed bya current of water from below upwards daily, or as often as may be required, and the rate of percolation will be main- tained at about fourteen times the rate af percolation adopted for the plain sand filters. The purpose of the preliminary filters at Belmont, and at the other works forming part of the improvement of the water supply, is three-fold: Primarily to enable the plain sand fil- ters to operate at a higher rate than has heretofore been em- ployed and correspondingly reduce the acreage of filter sur- face required to treat the whole water supply. Next, to pro- long the life or rather to increase the yield of the plain sand filter between scrapings from 60,000,000 or 70,000,000 gallons per acre to from 90,000,000 to 150,000,000 gallons per acre, 20 and finally to maintain a more regular effluent than is pos- sible with the plain sand filter when supplied with water which has been under either quiescent or continuous substi- dence for a day, or fora few days. The preliminary filters are intended to perform in a short time what could be accomplished only in a very long time by sedimentation reservoirs. The capacity of the Belmont preliminary filters is based upon arate of percolation equivalent to 80,000,000 gallons per acre per day, and at this rate nineteen of the twenty tanks will each furnish 2,200,000 gallons of rough filtered water, or pre filtered water, per day. In the operation of the preliminary filters, one filter con- taining only the gravel and underdrains will always be out of service,so that, when either of the filters becomes clogged with “mud balls” so as to demand a washing of the filter- sand, this sand will be thrown out of the filters by means of the ejector (presently to be described in connection with the removal of sand from the plain sand-filters) into a _ three- hopper ejector-washer and delivered from the washer into the. empty filter. Thus it will be-seen that the sand in transportation and washing will be worked from filter to filter, and, excepting at such times when one filter will be out of service for washing, and one filter receiving the washed material, there will always be nineteen filters in service, -1f1s sestimated that 1¢ willsrequire aboutectont hours to remove and wash the sand and replace it in the empty filter. When the operation of washing by the hop- per-washers has been performed, the sand will then be in the same condition it was when the filter was first started in service. Experience at the Spring Garden Testing Station has shown that even with the addition of the revolving rake agitator the reverse current of water is not capable of thoroughly washing and preventing the formation of some “mud balls” in the body of the sand, excepting with a wasteful expenditure of water, and it is anticipated that the cost of prefiltering the water, including the washing and transportation of the sand by means of the sand-ejector and lcs ot eee tert eter ceeaneaeomeeereememrerieeereeee ee ARES ‘S19}[J pues uleld jo ue[q—‘II ‘Oly $6 ¥20 == ae en a aN eer as ¢ ~ ant APRS LEY Med AR ce : cnedosprrnn, m $ ae te ee ee a Gre 48 OS Mtb BIWOR, 5 le is Pei Pa % x AO HDR AO PRS IC Bey, Beg, 40 ea _ a ary SON ADaHS SABGHS te oe ONIdid JO NV Id ey edakicsy 99 neobIRy WHS fay ANWId YSL1d LNOW Sd LCP SAE OY pe plog Lape SON LOVYLNOD Yr He Fel RS BUNT Rory panos po ie ~ io) 4 Pee f / “ay iff \ é F / ‘ Lf Lif f ; ji fe iy 3 “4 F ml ff fa , , > & £ ¢ - aN ay DM L970? AS Pht Deg 9 pis shen ahs G rveccer ** < eee ete ant saw: ee ae am eee 3 bananas 8 OO BS eeu? J ies eS Timer ri smshanie : Ee aera sewer © Ss * a , a 3 , t oF : gh exh % Reon sie & y Pd 3 & 7 ¢s a & ss x | ‘ é * : ip. ~ ‘ SRF Topmeray & bic decreas 8 ‘ Mops ee aed “7 ‘ et Sears ae * aide fen warna | ees ad Po aN Tat % : ; Spe Eh ak é ae : Agel 3 = 2 aoe ar H ~ 5‘ *, ; : = ~¥ ett Paes: : i : i . a ef * Hi i e : é ¢ & 3 ° : Hecatanens z. Rae: 5 cs os & SNNBAY : we SFR LNOWN38 EE Etat aE ee 3 3 é on —~ ore gv? | | etre ceri sideom Tra U if » | set aah : 4 : 1 h i | 1 | ee I 1 ' i | 22 ejector-washer, will be less than the cost of wash-water power and repairs to operate a filter with a mechanical agitator. PLAIN SAND FILTERS. (Contract No. 16.) fig. rz shows the eighteen plain sand filters now rapidly nearing completion under Contract No. 16. Six of these filters lie along Belmont Avenue, six along Ford Road, and six along Monument Avenue. The plan of the land taken by the city for the works rendered it impossible to maintain a standard of length and width of filter, as at Upper and Lower Roxborough, and generally at Torresdale, and, while the effective sand area is about the same for all filters, in planning the works it was found convenient to vary the relation of length and width to suit the ground. The horizontal. dimensions of the eighteen filters are given in the following table: DIMENSIONS OF FILTERS. Fei iters atz.— sae ee 242 feet 2 inches long, 135 feet 5 inches wide 8 SS ek Lees ee 272 66 8 6 66 I20 ac 2 ce 6 2 SL dre Shae ee 196 ce 5 66é ¢ 165 “6 Tel ae ce The contours of the land also was such as to make it inconvenient to locate all of the filters at common elevation, as at Upper Roxborough and Torresdale, and the terraced arrangement, through the six filters on Belmont Avenue, shown by /7g. 72, was resorted to. This involved a slight loss of head between the level of water in the sedimentation reservoir and in the clear-water basin, which was diminished as much as possible by rolling into fills on the ground at lower elevation, the excavation taken from the filters built on the ground at higher eleva- tion. The elevation of the normal flow-lines, or of water above the sand beds, is shown in the following tables: 5 Filters, elevation..\.0.0. os lsh cet han a ee eee 248 22001), 2m is a (eee me *o5; Wein OY, US Mon fond Aen 25034 on ere * Me Lu Pa yee ee A eae 25103 500s Dimers tC) Vic eMciaG! SiCene 307 clel, Ses am Gere Ges Bh es2 eae 4 a Ey ar Torun mins She. ey, har eee eee he AR Ae8 Smee ‘sId}[J JO JUIMISUvLIe Bde119} SUIMOYS TONI3G— ‘ZI ‘OI BEDS Ores rs yom Bowing ybaoigs waig2ee useubag gety4d i I eyep Uoyelyts 40 neBing Slip gO fUBLIDEL LL PISOLIB RIB {UOLY HF a HEM RE EEE : os senescence sane eccsstar 4h, BIN 184819. poset ihe Ee LEN PHS 3 *Iayy peo1dd4} jo uorjoas pue uel,q—Cl “OI 330-084 HO NOWORS TENIGNLIONOT vol vz Sst FA katie ailiee SA : Same, eee we 78 12. ‘ 108 8 ok OH Oe CO lane : F: ag s-mi Fo ayes shee se Aastg mawhoz yawn WON ASRs ALIS 3S SSEFLIEE 2 'ON MAL Tid 40 NOLLOZS GNY NYT re tae = INYid NILTld LNOWTEa PS he BOA ee ae eRe Dad 2 ne ere VET IS £0 8202 PRK III IA AY EKARASV ARE BRK BN RM ODP ow? PRGkO SF e7ANCRERD ae See “AISA SBA 20 WORLVEL TY HORRRIRS “LEYS 25 Difference of water surface between the highest and lowest filter, six feet. Since the level of water in the clear-water basin (for any system of terraced filters) must be adjusted to the flow-line of the lowest filters of the series, the apparent loss of head at Belmont is six feet; but to attempt the placing of all filters at a common elevation would have required the rais- ing of the fills under the filters at the lower levels, the materials for which could conveniently be obtained only from the excavation for filters at the higher levels, and the resulting common level of water surface in all filters would have been only about three feet higher than the water-level in the lowest level filter of the present series. ‘This, there- fore,is the true loss of head, against which the water is pumped from the Schuylkill River by reason of the terraced arrangement of filters at Belmont. That is to say, if all filters had been arranged at common water elevation, 251°33 CrOrethe= Noweline sof the Sedimentation reservoir might hav enpecniinades276:00.8,)),..instead.of-270°00'C.D:; but the fixed charges on the increased cost of construction to accom- plish this were found to be much greater than the annual cost of pumping against the additional three feet head. The plain sand filter at Belmont, as shown by /7g. 13, is in plan a rectangle with a net area at the sand line of 0°735 acre. By the longitudinal section through the filter it will be seen that the roof arches are carried on monolithic concrete piers 30 inches square at the base and 22 inches square at the top.of=the battered portion of the pier. “he battered section reaches to the proposed normal sand line. ‘The side and end walls have a width at base of 4 feet 2 inches, and at spring line of arch a width of 1 foot 1o inches. Division walls have the same dimensions in thickness as the piers. The floors consist of concrete inverts 15 feet 3 inches square, having a thickness of 6 inches at the center and of PAnuchbeswiudersthe piers, “lhe vaulting. ofzconcrete 1876 inches thick at the crown, and normal to the arch curve about 15 inches thick at the spring line. The piers are 9 feet 1 inch high, the rise or depth of arch of floor inverts at 26 center 8 inches, and rise of roof-arch at center 36 inches, making a total height from center of invert to center of soffit of arch 12 feet 9 inches. The clear span of roof-arches is 15 feet 5:inches: Under the floors and against the side and end walls of the filters, taken in groups of two, four, and six, clay puddle is placed as shown. The floor puddle was placed and rolled in two separate layers, each of 6 inches thickness when rolled in place. The puddle around the filters was carried 1 foot higher than the nominal water line in the filters, and rammed in layers of from 4 to 6 inches in thickness. (The value of the puddle as a means of insuring watertightness of the filters will be shown hereafter.) The end walls of all filters, and side walls of the end fil- ters, were designed as abutments to transmit the arch thrust to the foundations. Each alternate bay of the filter is provided in the roof- arch with a round ventilator opening 36 inches in diameter. This is closed with a double iron plate cover, forming an air space between the two plates to prevent the frost from acting on the water or on the bared sandbed while it is being scraped in winter. In summer time it is desirable to have some of the ventilators open to prevent the air in the filters from becoming musty by the decomposing sewage solids intercepted at the surface of the sandbed. With open filters the musty condition of the air over and around the filters is not noticed, but with closed filters, excepting ven- tilation is provided, the air over the water becomes saturated with the gases of decomposition of sewage solids intercepted at the surface of the sandbeds, and ventilation is a necessity for the comfort and possibly for the health of the men required to enter the filters for scraping and removal of the fouled sand. Over ithe arches ofthe nlters at theccrowneaslaversole2d inches of earth is placed to protect the water and sandbed, when laid bare, from frost in the winter and to protect the water from the sun’s rays in summer. At the depressions in the arches over the piers is placed gravel or ballast to act asa “French” drain and collect the water from rainfall which alias mip may percolate through the earth filling over the vaulting and pass it down to the filter through the opening left in THemUaUNCHIeOlUiemat Cine | Ne oper surlace-of the earth filling is finished with a dressing of top soil and seeded. In due time the surface of the fill will have a turf which, if kept properly watered and trimmed, will make a pleasing feature of the landscape. Six lines of cast-iron waterpipe are shown to the left of the filter, arranged as follows: Influent pipe to supply water to the filters. Effluent pipe to conduct away the filtered water. Refill pipe to refill filters from below with filtered water after scraping the sandbed. Raw-water drain pipe to draw off the water figs above the sand line. Drain pipe to draw off the watery from below the sandbed and conduct it to the sewer. High-pressure pipe to supply water to the sand ejectors, and also to the sand washers in the courts. In addition to the cast-iron pipes enumerated a 36-inch circular brick sewer, for waste water from the filters, and from the sand washers, and storm water collected on the courts, is also Shown. At Belmont (with the exception of two filters) the regu- lator houses are arranged each to serve two adjacent filters. (See /2zg. 74.) The well or chamber under the regulator house is divided into a middle dry chamber and two side wet chambers. Inthe middle chamber are placed the pipes, stop-valves, and special castings, which connect the main influent and high-pressure pipes with the filters, the raw water drain to remove the water from above the sand line, and, as shown, near the bottom of the chamber, the main effluent pipes, which originate in the side chambers, and are brought togetherinto a tee at the center of the dry chamber. The influent and effluent pipes for each filter are provided with stop-valves to cut either out of service without inter- fering with the adjacent filter. In the wet chambers are placed the effluent pipes, drain- pipes to remove the water from below the sandbed and L6 WZ BO Pe ee en Avy eo NI SSS RON A2FIHS ~ GLdTHS iP LES OFG6 SL 9CFETI SON SHA EHS SYIGWVHD ASIN) ONY SONLLYINDAY INYid Yai = INOW13E SHON LOVHLNO: Medi BMLVB. Hs 4 SOIL DEPOTS 88 y Oey Ler Pe Pease roMMi Ee bw Seg Segrcie? O pod ee ea : ate ge eae g yas LES ag oe CFR RE PAL CORRS 6 BIAS ey powaiebirneet Mace ted ge Le bos we MY BP Fhethes Deity 2 30h SOK PE Dailfe BS PRIS OK Oa y s ccseansaessanenneslilicsndiliiessedl at + . “ad te nom awos NOILLQANNOS ¥ - WAS WY38 40 TivL3d- $ * Begay 2u) 19 AE j 3 : ieeceeaccs: ie came net ae cae ERR SA WERE Speat PHB “Soq(meyo JUON_yS pue JuUonyYuy— US NOMS Vi ° OLS LORE LIP et a. RAT aE ¥v NO! hb eet Tes # > o he, 29 waste it to the sewer, and the re-fill pipe to supply water to the filter from below upon starting after a filter has been scraped. Upon the main effluent pipe where it starts in the wet chamber, is placed an automatic telescoping circular weir, controlled in position with reference to the surface of the water in the chamber, by an annular copper float. This self-adjusting weir (which has its counterpart in the effluent chambers of the filters at Warsaw, Zurich, Bremen, and other places abroad) has its lip fixed at a definite depth below the surface of the water, and is not only an effective device to maintain a constant rate of flow of water from the ‘filters, but when properly standardized becomes a reason- ably accurate meter for the measurement of the flow. The relation of the lip of the weir to the position of the float is adjustable, and it thus can be varied to produce any desired rate of percolation through the filter. Ingthe: early torm) of) the effluent weir: the Jtelescoping joint was made with a water-packed bronze gland; but experience at Roxborough has shown that it is impos- sible to produce drawn-brass pipes, 20 inches in diameter, so round or uniform in diameter as to admit of the use of the packing shown, and a leather packing, which will TeACil yma Usieitceli tO mtne: eccentricity -Ofs the sliding tube and to the inequalities of diameter, has been substi- tuted for the internally grooved bronze ring. The leather packings have worked admirably on the effluent regulators of the Roxborough filters, and, with the weir floating upon a constantly declining water surface, maintain, when adjusted, a practically unvarying rate of discharge from the filters. Some trouble has been experienced from the air whichis drawn into the column of water descending the telescoping pipe, and to remedy this defect in the construction of the feo uintorsmt lemieckSeatetnestOpran the txed barrels:or the regulators are being tapped and provided with pipes reach- ing to a few inches above the highest possible water level in the effluent chamber, to vent the air from the regulator barrel. A test of one of these air-vents at Upper Roxborough “1OJE[NSaI JUSNYU oyewMoynY—'S1 “OI BALLIA JO UII fe a oe Z WIIMONT BRD ai ° ; “THM NHOP * : NOLLWMIT4 JO AVANOE 2 AOLYVINGIY LNANTINI SYaLid : Z NOW qd g $OUY DUB UOIG BROEOAA f? Bek cS UO OABIF PRPUPED,, t seretann TG fO2s f OUE MORAG OULYIB AS pe = % Ais pie ts os af * & PEO E AEE HEY PRY ; st See BESS gS GCE HO 808) BIR Pian Sewn ases actor i i ce ee ne rere ane ni ig Ce ee « a : . 4 * be. bey Sig SAORI > ~ AD 4 OIE AE Aik seems to indicate that the trouble due to induction of air by the annular column of water flowing over the lip of the weir and down the telescoping pipe will be avoided in this manner. The sleeve, or jacket-pipe, outside the telescoping pipe is used to prevent the current of water flowing from the regu- lator from having a tendency to draw the telescoping pipe from its central position, and cause it to bind in the packed joint. Inside the filter on the end of the branch from the influ- ent main is placed an influent regulator (/7g. 15) to auto- matically maintain a constant depth of water over the sand- bed. ‘This consists of a plain double-seat disc-valve, nearly balanced, actuated by a cylindrical copper float. By the rise and fall of the float on the water in the filter-the valve is closed or opened and the rate of inflow of raw water main- tained nearly constant. This valve is in all material respects a copy of the influent regulators used on the Hamburg filters. WATER COLLECTORS AND FILTERING MATERIALS. (Contract No. 49.) In the center of each filter, at the bottom, as shown in Fig. 16, 1s placed a main water collector into which, at inter- vals of about fifteen feet, are connected the lateral collectors, to conduct the filtered water from the respective bays to the main collectors. The main collectors in filters Nos. 1 and 2 consist of a line of 30-inch diameter vitrified sewer pipe, provided at the center of each filter bay with a double 8-inch branch, into which the laterals are connected. ‘The main collectors in all other filters at Belmont were formed by placing concrete- steel reinforced slabs, 5 inches thick, over two low concrete Wiilcmiliiiettiethescenter Olfstierniter oor, Che cwalls are 16 inches high, and average 134 inches thick, plumb on the inside and battered 3 inches on the outside, spaced 4 feet apart; the space between the walls and under the slabs con- stitutes the filtered water channel. These collectors were built in the filters after all other work, previous to placing the filtering materials, was done. ‘S[BIIOJVUI SULIO}[Y pue si9q[y YS8no1y} uorjoas—'9i “OY : SU2d HONOWHL NGILIGS YNIONUIONGT : coe ee ieee sf (eno BROS 80: RAE YRS mR PE EA ER, ¢ : x ene SH : v auras ahesny Bo wane 8 Oe es ee WeROM CFIA 4h TeaKEIeED es ee Kets oat RWUNIED. Ki ELON RID AS BRREDLNY "AWING BELEN 40 WO YRL Ya GW KoteUIL S2 LETHAL TEER! HOMVHLTia fo Hywe Ash w f ai¥oR ccnionsennnnicennns “7 ON L33HS SiTsHs 2@ SIIWSLYW ONINAL Wd HONOYHL SNOWD3S ANW1Id Y3LUs LNOW128 6V ON LOVYELNOD pn fees HOLOFTION WHBLYT $8 40 ST¥lad vos BPE LIBS é "apace sasgar Won f Wer Pes ‘HOUY 40 WAKOUD HORGUHL NOILIGS JeuzASHYHL PAT RIMES FE LOBE AIRHRS BOE? Ke whi Sake aye The lateral collectors, as shown by fig. 77, consist of a line of 8-inch diameter vitrified pipe, perforated all around from end to end, and plugged at the end of the line remote from the main collector. Around the collectors and for a height of 6 inches from the floor, is placed gravel, ranging in size from 3 inches to 12 inches in diameter. Above this is placed a 4-inch layer of gravel, ranging in size from 12 inches to 2 inch in diameter. Above this is placed a 3-inch layer of gravel, ranging in size from 2 inch to # inch in diameter. Above this is placed a 2-inch layer of gravel, ranging in size from 4-inch diameter to material which would be re. tained on a sieve having fourteen meshes to the linear inch, and above this a final layer, one inch thick, of coarse sand, which would pass a No. 14 sieve and be retained on a No. 20 sieve. The whole depth of underdrain gravel is therefore 16 inches, measured from the center of the floor inverts. Three plans have been tested for the distribution of the underdrain materials, as follows: Plan “A,” fzgs. 16 and 78, shows the gravel everywhere kept from 20 to 24 inches clear of the masonry side and end walls, and piers of the filters. Plan “ B,” fzg. 76, in which the gravel is kept 24 inches clear of the side and end walls, but impinges against the Die re Piatee Cy diz ane which the oravel is spread shorizon- tally from wall to wall, and impinges against the piers. Plan “A” requires the least amount of gravel, but is most expensive for labor of placing, while plan “C” requires the largest amount of gravel, but is least expensive for labor of placing. So far as our experience has gone in the operation of the filters at Roxborough, neither plan has any preference over the other, although plan “‘A” appears to be the ideal system for the underdrains in covered filters, because water which might pass down between the bed of sand and the masonry of the walls or piers, could not possibly escape from the fil- 3]| ‘|BII9}VUL UTeIpPIOpUN puke S10}D9T[O9 [vId}LVT SuIMOYS ‘s19}]y Jo 101Ia}UJ—"ZI “OI "pues jo uorjdeoar 10j Apes [elsio}emPUleIpIopungsuimoys 19}[y Jo 10119jU[— "gr ‘OTT ; 36 ter without first passing laterally through the lower part of the sandbed. With open filters plan “B” will produce the same results. In a locality where materials are easily procured and cheap, a system of underdrains by plan “ C,” with carein the placing of the layers of sand in the sandbed, should meet every practical requirement. Above tlie gravel underdrains to a depth of 36 inches is placed the bed of filter sand. In placing the sand in a series of filters, some are filled to a depth of more than 36 inches, and some to a depth less than 36 inches, in order that the time of going out of service for resanding may not occur to more than one or two filters of a system at the same time. Thus, at Belmont, 3 filters will receive the sand for a depth of 28 inches. 3 filters for a depth of 31 inches. syolters fora depth ofsqunches. 3 filters for a depth of 37 inches. 3 filters for a depth of 4o inches. sail terest Otascepta Olena nchec An average depth for all filters of 35°5 inches. The original depths of sand before the water is intro- duced and the sandbed settled, is about 6 per cent. more than the depths given. In the thin beds the sand is placed in two layers, and in the thicker beds it is placed in three layers. After the sand has been placed to the proper depth, the filter is slowly filled with water from below until it fairly covers the sand, and the bed allowed to settle for a period of ten days or two weeks. Afterwards the water is drawn down, and the depth of the bed then taken. Care in placing the sand has reduced the shrinkage by water settlement to as low as 4 percent. That is, a bed originally 40 inches in thickness will shrink upon settling with water to slightly more than 38 inches, although the usual shrinkage amounts to about 6 per cent. of the original depth. The sand used may be river or bank sand, provided it is well washed, and complies with the following physical requirements: 37 No particles should be intercepted by a No. 6 sieve, and but few particles should pass a No, 60 sieve. Such sand, by Massachusetts State Board of Health standard, will have an effective size of about 0°35 millimeter, and a uniformity co-efficient of about 2°50. | Experience with the Roxborough filters indicates no material difference in the work of the filters supplied with river or bank sand, excepting that the river sand filters earliest give the best clarification of the water, and the Fic. 19 —Filter entrance and regulator house. bank sand filters earliest give the best bacterial results. After a few weeks of operation, however, it is difficult to detect any difference of performance which might be attributed to the source of the sand. The cleanliness of the filter sand when first placed must comply with the following requirements: When 100 grams of river sand are thoroughly shaken up in a beaker containing 1 liter of distilled water (or filtered water showing o + turbidity), the resulting turbidity of the 38 water shall not exceed 400 parts per million by the silica standard, and, when 100 grams of bank sand are similarly tested, it shall not show a turbidity of more than 200 parts per million by the silica standard. Sand shall show not less than 95 per cent. silica, calculated as oxide of silica, and not more than 1 per cent. of lime and magnesia taken together and calculated as carbonates. Bank sand being always lighter in color than river sand, has an advantage, when the filters are scraped, in striking a sharp line of division between the clean and dirty sand. Experience at Roxborough shows the sand scrapings to average about 1 inch in thickness, and the theory that a thin layer of sand at the surface of the bed does nearly, if not quite, the whole work is abundantly proven by the scrapings of these filters. Fig. 19 shows one of the regulator houses built over the influent and effluent chambers of the filters, and the entrance to a filter. The houses and filter entrances are uniform in design for all the filter works. Roman size brick has been used in all face work, and the best quality stretcher brick for inside linings. Cut stone, Eastern grey granite; terra- cotta belt courses and architraves; copper cornices, gutters and roof flashings; copper ridge and hip rolls and slate roof covering. ‘The floors of all regulator houses are made up of iron plates laid on rolled “I” beams, perforated plates being used over the dry or influent chambers, and solid plates over the wet: or -efiluentechampbers> =) The losseor beads gauges, which show automatically the difference of water levels over the sandbeds and in the effluent chambers, are mounted in the regulator houses. THE CLEAR-WATER BASIN. (Contract No. 16.) The clear-water basin, which receives the effluents from all the filters, is located at the southeast corner of Monu- ment Avenue and Ford Road. The basin measures inside on neat lines, 396 feet long by 382 feet 2 inches wide. This detail is constructed like the filters, with a concrete floor, 39 concrete piers to support the roof, and concrete groined arched vaulting, above which is placed a covering of earth about 24 inches thick at the crown of the arch. The puddle layer under the floor is 12 inches in thickness. Floor inverts, 6 inches thick at the center and 14 inches thick under the pillars. The piers are plumb 22 inches square and 14 feet 4 inches high. Depth of water, center of inverts to spring of arches, is 15 feet. Clear span of roof arches, 14 feet. Rise of arch, 3 feet. The filtered water is conducted to the clear-water basin through .a line of 48-inch cast-iron pipe in Monument Avenue, which enters the basin at the northeast corner. In the chamber where the influent pipe terminates, a 30-inch overflow pipe is located which will limit the depth of water in the basin to 15 feet r1oinches. The elevation of normal flow line in the basin is 23900 C.D., but this at times may be temporarily increased to 239°83 C.D. The flow line of the present George’s Hill Reservoir in the West Park is 21200 C.D., so that the clear-water basin which in the future will furnish the head for the West Philadelphia lower service, is 27 feet higher than the old distributing reservoir. The Belmont clear-water basin will have a capacity at 15 feet depth of water of 16,500,000 gallons. In the plans provision has been made for the construction of another basin of the same capacity directly north of the basin shown, but it is possible that in the future operation of the works the additional basin may not be required for many years, certainly not until after the consumption of water from this station exceeds 60,000,000 gallons per day. The effluent from the basin is taken through a 48-inch cast-iron pipe located at the southwest corner, which leads into Monument Avenue, and is connected with a 48-inch line of pipe, known as line ‘*‘ kK,” and placed under Contract No. 19, which leads southward on Monument Avenue to Belmont Avenue, and on Belmont Avenue to Montgomery Avenue, where it is connected into the present rising pipes from Belmont Pumping Station to the George’s Hill Reser- voir. When the Belmont works are started, and the whole ‘UISBG 19JVM-1IVI[D YSNOIG} SUOTIIIG—'O% “OI Oll We SCS Higents 1 ey SU3id HONOWHL NOILI3S 108) Be Cups ¢ : z ; ; com id A aoe at al oc ae vwos gs wiz OU GN AW3KE SAIIHS Ie NISVB Y3LVM G3Y3L14A 40 SNOILDIS ANY Id YSLWd LNOW ISS 91sN LOWYLNOD = uN Evgnn@ i + = eo 2x o. dof - 3 Sec 2 MELE REL PME A ADE so REE BORE PNR Up WO BE 6AM BARS ABCD Fe Racy ‘ : BRE BREE ES ESP Bo? ~ ROY AI supply of West Philadelphia is coming from the filters, the connection of the rising mains with the reservoir at George’s Hill will be cut out, and these pipes thereafter will form a part of the distribution system. Ample provision has been made at the intersection of the new mains from the clear- water basin with the old rising mains in Belmont Avenue at Montgomery Avenue, to prevent unfiltered water from the Schuylkill River mixing with the filtered water coming from the filter station. The effluent pipe starts in the floor of the clear-water basin with a bell-mouth casting to reduce the resistance of entry. Drains connected with the sewer in Monument Avenue are provided to permit of emptying and examining the basin. At such times a by-pass connection from the main effluent pipes in the court lying between the group of filters 7 to 12 and 13 to 18, to the 48-inch distributing main in Monument Avenue, enables the filters to temporarily discharge their effluents directly into the distribution system. The only object in cutting out of service the clear-water basin will be to repair some damage, because it will only receive filtered water, and of course will never require cleaning. It is thought that this basin, after it is properly started in service, may never be taken out of service, excepting tests should be desired to prove its continued watertightness. The puddle lining under the floor and against the side and end walls to the extreme water line is sharply shown by /zg. 20. It is not admissible to pass the water which may percolate through the earth Alling over the clear-water basin into the basin to mix with the filtered water, and sub- soil drainage was therefore provided to remove this water and discharge it into the sewers. The spaces between the arches and over the piers were solidly rammed with a rich puddle to prevent water from flowing through the material and collecting over the haunches of the arches, and the sub- soil drains, shown, will remove the water of percolation through the earth fill down to the upper surface of the puddle and crowns of the arches. The upper surface of the fill is finished with a dressing 42 of topsoil and seeded to produce eventually a turf over the roof. All slopes of the clear-water basin and filters are sodded. A few ventilator openings are provided in the roof of the clear-water basin to furnish light and ventilation when required, the covers of which are securely locked to prevent intrusion by mischievous or curiously inclined persons. CLAY PUDDLE. (Contract No. 16.) The chief reliance for watertightness of the structures is placed in the clay puddle, which consisted, as manufactured for the Belmont works, of 50 per cent. by volume of clay and 50 per cent. by volume of ballast, which may be clean gravel or broken stone. The materials were mixed in hori- zontal screw-paddle pug mills, such as are used for temper- ing brick and tileclay. During the busiest part of the work, five of these machines were kept constantly in motion pre- paring puddle for the water tight linings of the reservoirs and filters, and for packing around such lines of pipes as were placed in embankment. The clay consisted of equal parts of a strong, heavy ferruginous clay obtained from Perth Amboy and Bruns- wick, in New Jersey, or from Charlestown, Maryland, and of a weak clay combined with small gravel obtained from Swedeland, Montgomery County, Pennsylvania. The heavy clay could not successfully be worked alone in the pug mill, and the weaker clay was added to assist in breaking up and tempering the heavy clay. The ballast generally was broken stone, which varied in size from } inch to 14 inch in diame- ters. Mixtures were sometimes made, consisting of I part by volume of the strong clay, 1 part of the weak clay mixed with gravel, and 1 part of ballast. The strong clay by rational analysis (Ulzer’s method), which wrought the separation of the silicate of alumina and iron from the insoluble silica and other substances, showed by weight from 60 to 70 per cent. silicate of iron and alumina, which was regarded as the clay constituent of the material, while the weaker clay showed by the same method 43 of test from 35 to 40 per cent. of silicate of alumina and iron as the clay constituent of the material. The mixture of clays usually yielded about 50 per cent. of the silicates of alumina and iron and 50 per cent. of silica and other insolu- ble substances. (A natural clay showing by weight about 50 per cent. of the silicate of alumina and iron could be used alone in the manufacture of puddle.) Using equal parts of clay and ballast, gave a matrix for the ballast about 25 per cent. in excess of the voids in the mass of hard material. Many clays were examined for use in the puddle, varying from a micaceous loam to strong clays, and showing by uni- form method of rational test from 20 to 70 per cent. silicate of alumina andiron. In all cases it was assumed that the iron present in the clay was a valuable constituent of water- tight puddle. While the clay and ballast was being worked through the pug mill, water was added in quantity sufficient to make a plastic mixture. The puddle was placed in two or three layers of from 6 to g inches and rolled in place to a thickness of 4 to 6 inches. Puddle linings varied from 12 to 18 inches in thickness, and were always placed in the workin from two to three separate layers of uniform thickness, each layer being rolled to a solid, dense mass having great sustaining power before the next layer was spread. The least thickness of puddle lining, when rolled or rammed to proper elevation, was 12 inches. When the rolling of any layer of floor puddle under the basins or filters was finally completed, the puddle was as solid and almost as hard as new concrete. When rolling was inadmissible, as, for example, around the walls of the filter and the effluent and influent chambers and gate chambers, the puddle lining was used sometimes 24 inches in horizontal thickness, and solidly rammed in thin layers. The value of the puddle asa water-tight lining for thin concrete sections, is shown by the following table contain- ing the leakage undera head of g feet of the accepted filters at Belmont: ‘sadojs 10Asaser uo afppnd Suljoy— iz ‘oy - 45 Filter No. 1 . Leakage 970 gallons per day of twenty-four hours ce ac 3 ce 12, oe ¢ (as ec c¢ c¢ oe ¢ 4 é ; ‘ im 970 oe oe ce «¢ a c¢ Las ae ie 610 - cae ie te + de eG “ 243 i oS Bee ate ee Y oe ‘ S «¢ 566 a¢ ee ce ims oe oe YF Fda CYA. agi a 728 4 Sear tees ‘ * s Pi te ee Oe : 776 a ae Pe aes “ . ae Seer To ee ee “s 582 up aN Seg Die vy es os Sh ted ae ‘* 475 $ ae EMD gf Bs 2s a TLL atures ye 868 Be me! ayaa es ye oe ws me yuige epa ee sf 849 ty 0 es eee i ve The standard of waterttghtness for the filters was a leak- age of not more than 1,000 gallons in twenty-four hours, corresponding to a loss, based on the daily capacity of the filter, of 0'0228 (or =) of I per cent. The puddle on the slopes of the sedimentation reservoir, which could not be rolled with ordinary grooved horse- rollers, or with the steam rollers used on the floors of the reservoir and filters, was rolled by the ingenious single horse-roller (Fzg. 27) improvised by Mr. Lawrence O'Toole, the foreman for the contractors, Messrs. Ryan & Kelley. This consisted of 36 inches of a 20-inch cast-iron water pipe filled with concrete, to give weight, and convert the pipe into aroller. Inthecenterof theconcrete an iron axle was fixed which turned on bearings provided in the lower ends of two standards bolted to the under side of the shafts at the rear. Wrought-iron bands were shrunk on the pipe to make a grooved roller. This apparatus, with the addition of a mule, constituted a roller which was worked around the slopes. The puddle was placed on the slopes in thin horizontal layers, and three of the “O'Toole” rollers were constantly worked over it. The puddle was rolled by six horse-rollers, varying from I to 5 tons in weight, and two steam rollers weighing 6 and 74 tons respectively. The horse-rollers weighed each 1,000 pounds per linear foot of roller and upwards, and the steam rollers weighed about 2,000 pounds per linear foot of roller. 46 CONCRETE. (Contract No. 16.) Nearly the entire masonry, including the reservoir floor and slope paving, and the paving in the courts, foundations of the buildings and sand washers, was built of Portland cement concrete. Concrete was used in the floors, piers and vaulting of the filters and in the clear-water basin; in the influent and effluent chambers, and everywhere within its adaptability, partly to facilitate the construction of the work, partly to economize in the cost, and partly to reduce the number of joints in the structures. If other material than concrete had been used in much of the work, the time and cost of construction would have been largely increased, with no corresponding advan- tage to the works. Seventy-four thousand barrels of Amer- ican Portland cement from the Star Bonneville, Lehigh and Atlas factories were used in the manufacture of concrete for Contract No. 16 alone, after the following proportions: Cement Dy volume sag coasts oa Bore ceo cal ore ee I part, Sand:-by: volonie-47.9—. oc egt seeas ue eres ae con Oe eee 3 parts, Ballast Dy;VoliMe ei a Rete. oe ae oe eee eee 5 parts. Six-inch concrete cubes were molded from day to day, and crushed at the end of 30, 60, go, 120, and 180 days, with the following results: Average of all the cubes at the end of 30 days, 1,781 Ibs. per sq. in. a a io 6007 8a Score. = - ce Hy OCT were ek as ue e ay 1204 2306205 he Be _ + In0..),. «2.-210—- Considering the go- and 120-day cubes, which best repre- sent the concrete before being subjected to external stresses, the average is over 2,000 pounds per square inch. Very rarely did the cubes fall below 1,200 pounds per square inch. Each cube was numbered and its location in the work entered in the records, and whenever any cube after ninety days’ time showed a crushing strength of less than 47 1,400 or 1,500 pounds per square inch, the concrete in the structures of which it was a sample was drilled or cut into to determine its hardness and density, and in no instance was the concrete found in such a condition as to raise a doubt of its quality. Excepting the arches in the vaulting of the filters and clear-water basin, the concrete is nowhere severely stressed, and concrete much weaker than that usu- ally employed in the construction of fire-proof floors, and steel-reinforced concrete beams, would meet all requirements of such work as that under consideration. The cement used in the manufacture of concrete was fur- nished under the following conditions: MPeciuc PlAvity NOmless (UAT wr s sea wea ne Sete per cent irenicssmetaimed sO IN O7150.StCVe qi-m = mara 8 ante - . Oper.cent. Ss x DE ESOL ENE. Oe 60 rd ey he ey a ee ae (ee : e io ZOOL erie CMe ad he a aa a Ae ye Pa Initial set (determined with a Vicat needle), not less than 20 minutes. Tensile strength of briquettes, consisting of one part cement, three parts standard quartz sand, one day in air and six days in water, 170 pounds per square inch ; one day in air and twenty-seven days in water, 240 pounds per square inch. The average of tensile strength for briquettes made of one part cement and three parts of standard quartz sand, as stated above, during the two years of construction work at Belmont, is about 200 pounds per square inch for seven days —one day in air and six days in water, and 300 pounds per square inch for twenty-eight days—one day in air and twen- ty-seven days in water. When the boiling test of cement was applied it was ex- pected to show no disintegration of the egg-shaped sample. No cement that failed to give the required strength at the end of seven days was allowed to go into the work for twenty-eight days, and if it failed to show the required strength at the end of twenty-eight days, it was rejected en- tirely, or its use was occasionally permitted at some point in the work where strength of concrete or mortar was not par- ticularly desired. The sand used in the manufacture of concrete was clean New Jersey bank sand; the ballast was broken limestone, 48 ranging in any dimension from 14 to + inch, thoroughly screened of finer materials. In the granolithic finish of con- crete work, the proportions were one part cement, one and one-half part clean New Jersey bank sand, and one and one- half part (liumestone screeninos: experiments mimi neue tty Laboratory having shown better results in point of strength for a mortar from limestone screenings than from quartz sand or New Jersey bank sand. The percentage of water used in mixing concrete varied {rom \15° Per Cents tOpLompch Cents OlsLiemCGilicH (sent iic mii ballast by volume, In the cubical box mixers used in this work, after the sand and ballast were introduced into the box, the materials were turned from four to six times dry, and after the water was added the mixers were again turned from eighteen to twenty times. A small portion of the concrete was mixed in a horizontal knife mixer, the knives or paddles being so arranged that they worked the materials from the ends of the box to the center, where it was forced up and over the knives at the center to the ends of the box, and again forced from the ends to the center, and so on, until properly mixed. The capacity of. the concrete mixing machinery was limited, 16450, cubic, yards pers day, butsthevactwaliaterat which concrete was mixed and placed in the floors and slopes of the reservoir and in the floors of filters, and clear- water basin, was usually determined by the rate at which puddle could be, mixed, placedvand= rolled Wirectlysine puddle was finished at any point, concrete was immediately placed over it. In the floor sections of the filters on fill, expanded metal was freely used, so placed as to bond adjacent sections of concrete, and generally to strengthen that portion of the floor which it was assumed would be stressed by the loads the bases of the piers. Each pier in the clear-water basin sustains at its base a load of about 64 tons, spread over an assumed surface on the puddle lining under the concrete of about 17 square feet, or the unit load on so much of the floor inverts as are 49 supposed to resist and distribute the pier loads, is nearly four tons per square foot. The rolling of the subsoil or earth-fill preparatory to placing the puddle lining, and the subsequent rolling of the puddle, rendered this in all instances a good foundation for what are comparatively light loads. When the foundation (in the clear-water basin) was not satisfactory, concrete sub-piers three feet square, and from six to eight feet deep or high, carried up from solid ground or from a grillage to the under side of the puddle, were built; the whole load on the pier was thus transmitted through the puddle to the sub-pier below. The foundation under the concrete floors and the puddle lining was rolled with grooved rollers, weighing about 3,300 pounds per foot width of roll. With the exception of the piers in one filter, all concrete was rammed and finished in place. In Filter No. 1 the piers were fabricated as monoliths and set by a derrick. ASPHALT RESERVOIR LINING. (Contract No. 16.) Over the concrete floor and on the slopes to a height 10 feet vertical below the water line, as shown by /7gs. 22 and 23, a lining of asphalt 2inch thick was placed in two layers, each uniformly inch thick. Forseveral reasonsitis desirable that the subsidence basins be as nearly water-tight as such struc- tures, constructed partly in embankment, can be made, and in addition to the outer lining of 18 inches of puddle and 6 inches ot concrete floor and slope paving, it was deemed ad- visable to line the floor and slopes with an impervious coat of asphalt. Two mixtures of asphalt were used, one containing the larger percentage of bitumen on the floor and first coat on the slopes, and the other, slightly lower in bitumen, ir the second or finishing coat on the slopes. The mixture of Neufchatel or Seyssel asphalt, Bermudez asphalt and grit, as specified in the contract, should contain about 18 per cent. of pure bitumen, but tests early indicated All ‘S11OAdaS01 JO SIOOY UO Buruty yeydse Surse[g—ez “iy ‘WOAIOHOL JO sedole HO Saray ypuydey Aupoup_p "Le ons BAe RE Ps a 52 that such a mixture was too soft for use on the slopes, and probably not superior as a water-tight lining on the floors of the basins, and, under the provision of the specification, the percentage of bitumen was accordingly decreased. The several mixtures tested contained as an average the following weight of materials for each batch that went into the kettle. ASPHALT MIXTURE USED ON THE FLOOR OF RESERVOIR AND IN THE FIRST LAYER ON SLOPE. 585 pounds Seyssel mastic. 315 pounds grit. 50 pounds refined Trinidad asphalt. 50 pounds refined Bermudez asphalt. ASPHALT MIXTURE USED IN SECOND LAYER ON SLOPES OF RESERVOIR. 5.8 pounds Seyssel mastic. 332 pounds grit. 33 pounds refined Trinidad asphalt. 37 pounds refined Bermudez asphalt. The above mixture gave an average of 15°5 percent. of bitu- men for the lining on the floor and first layer on the slopes, and 13°2 per cent. of bitumen for the second layer on the slopes, the latter, of course, requiring a stiffer mixure to pre- vent or limit the creeping by action of the sun’s rays, and likewise, of course, to avoid cracking due to the influence of the frost. On the floor of the basins, the asphalt was laid on the smooth concrete, but on the slopes the concrete was rough- ened by indenting grooves 4 inch deep and 2inch wide, spaced about 4 inches center from the toe to ate top of the asphalt line, to secure the asphalt against slipping or creep- ing on the concrete. The watertightness of the asphalt, concrete na puddle in the reservoirs is still to be tested, but I have no doubt that they will be as near water-tight as structures as large as these can very well be made. Generally, of course, we cannot expect earthen embankments, however carefully they may be lined with impervious materials, to be absolutely water-tight, but I anticipate that the measured leakage of 53 these basins will show such a small percentage of loss as to indicate practical watertightness. It is possible that the asphalt lining might have been omitted without seriously affecting the watertightness of the reservoir; but, considering the height to which the water is pumped from the Schuylkill River, and the nature of the surroundings, it was thought wise to omit no precautions to insure the nearest approach to absolute watertightness of the structure. ADMINISTRATION BUILDING. (Contract No. 42.) The Administration Building, Figs. 24 and 25, is located on the west side of the main court, between Filter No. 6 and the first installation of preliminary filters, and contains on the ground fioor a boiler-room, engine- or pump-room, an office and a shelter and locker-room for the men employed about the station. The second floor over the office will be fitted up fora laboratory for the technical examination of water samples from these and the Roxborough filters, and the basement under the office end of the building will be used as a store- toom for tools and supplies. In the pump-room are placed (under Contract No. 40-A)a set of centrifugal pumps to supply subsided water to the preliminary filter-house to wash the sandbed, and, if desired, also to draw off the water from above the sandbeds of low- level filters, and pump it into the supply pipes of high-level filters (thus avoiding the waste to the sewers of pre-filtered water); a set of duplex direct-acting pumps (furnished under Contract No. 40-B) to take water from the main effluent pipes of the filters, and pump under pressure of 80 to 90 pounds to the sand ejectors and sand washers; and the driving engines, electrical generators and main switchboard (furnished under Contract No. 46) for the electric lighting equipment of the works. The boiler-room contains four internally-fired marine boilers, each of 200 commercial horse-power capacity. The steam power developed will be used partly to pump wash- ‘uor}ejs Surdmind pur Suipiing vorzesysIUIMIpY Jo ue[G— Pz ‘OI hes Spe prt eR FS ES Me oR pues oe nore ee 808 uegvucgesrer pgby % ~ BHD ®: 2 ~ PREGA Ret ed poh dang sagen, bap ra 3 RRO OFTERE 40 emer Wiha Ry ms, pace eee : is x : : : 5: ‘ : : i : *: SUES 45 CIOMINBBS 4 SER ROG FAR : ; oe ae : : : : ee oor ‘Rubst0 BBLEA, $0 : Epes 4 reat anaes : See { 2 Sees = Stee te ot Pot PRL BE SPRUCE : : é Sees Scoteac ar a ge go> : RALVNE US 4S PEAS BRP ORLY 2608 f : Rk i dad a iets Hench came Nees, nes Pecan RUS UE ae ee eke Sas pe ed ges “319 ¥ : : ‘ a ‘ : Bae ee 35 : bryybry fo bi MOLUMG GALIL) LOLS TI gam -@ sump pososktey _— eee puanainng MolLassiidory —rrr—t—‘a_——O—si—r—S—~—~s~—<