STUDY ON THE EROSION RATE OF POWDERED ACTIVATED CARBON IN RAW WATER AQUEDUCT

442Available online at www scicncedirect comScienceDirectJounat of HydrodynamicsFL SEVIER2012,24(3):442-449“wu sciencedirecl.com:DOI: 10.1016/S1001-6058(11)60265-5scrence/jourmal: J016058STUDY ON THE EROSION RATE OF POWDERED ACTIVATED CARBONIN RAW WATER AQUEDUCT*YIN Hai-longState Key Laboratory of Pollution Control and Resource Reuse, Tongji University, Shanghai 200092, China,E-mail: yinhailong@tongji.edu.cnYANG YangKey Laboratory of Yangtze River Water Environment, Ministry of Education, Shanghai 200092, ChinaXU Zu-xinState Key Laboratory of Pollution Control and Resource Reuse, Tongji University, Shanghai 200092, China(Received November 3, 2011, Revised December 20, 2011)Abstract: Growing interest in using Powdered Activated Carbon (PAC) in raw water aqueduct, as a method of polluted surface watertreatment, raises the question of transport of PAC in the aqueduct, which is related to the potential PAC erosion along the aqueduct.By means of a recently developed re-circulating flume, erosion rates of PAC with the grain size of 230 meshes (less than 62 μm)depending on shear stress and bulk density were the discussed with real-time measurement of suspended PAC conccntration. LatcralCross sectional averaging shear stress was decided by the actual value in the raw water conveying aqueduct of upstream HuangpuRiver, Shanghai, China, smaller than 1.8 Nm 2. As for the bulk density. it was measured with compacting times varying from 1 d toI5 d, equivalent to I 550 kg/m2-1 800 kg/m'. Expcriments werc conducted for the shear stress and bulk density separately, so as toisolate and quantify the effects of one of the parameters. The results demonstrate that, for a particular PAC particle, the erosion ratcincreases with shear stress and decreases with bulk density as a function of power form. A product of powers of the lateral crosssectional averaging shear stress and bulk density to estimate PAC erosion rate is presented by approximating experimental data sets.Key words: Powdered Activated Carbon (PAC), raw water aqueduct, erosion rale, shear stress, bulk densiryIntroductionyears, the raw water aqueduct as the water movingPowdered Activated Carbon (PAC) has been acarrier from drinking water source area to drinkingwidely used and accepted method for the removal ofwater plant has been focused on. The reason is that, ifwater pollutants, owing to its efficiency at removingPAC is used at the starting point of water aqueduct ondissolved micro organic pollutants. However, PACthe magnitude of kilometers or more in general, theadsorption may be not effective if the application isHRT of PAC in water can be prolonged, helpful foronly conducted in drinking water plant, with no suffi-improving PAC efficiency. For cxamplc, in China,cient Hydraulic Retention Time (HRT). In recentapparatus feeding PAC into raw water aqueduct werebuilt in Beijing in 2008, for the purpose of treatingpotentially polluted raw waler from the Yangtze River* Project supported by the Narional Natural Science Foun-via China's Middle Route South-to-North W aterdation of China (Grant No. I1 102137), State Key Laboratory ofDiversion Project. Another apparatus were built inPollution Control and Resource Reuse Foundation, Tongjidrinking water source area in the upstream HuangpuUniversity (Grant No. PCRRY1 1009) and the National HighRiver in Shanghai in 2009, for treating potentiallyTechnology Research and Development Program of China (863polluted water from Huangpu River.Program, Grant No.2008AA06A412).Biography: YIN Hai-long(1976-), Male, Ph. D,中国煤化工PAC feeding facili-Associate Professorties,he transport processCorresponding author: XU Zu-xin,of P/MHC N M H G. Typical interactiveE-mail: xzx@stcsm.gov .cntransport processes comprise PAC transport by advc-ction and turbulent diffusion, with sediment deposi-4441.2 Experimental setupfor a trunk of channel, when the flow is consideredAs an experimental apparatus, usually flume cansteady and uniform energy slopebe subdivided into two types: re-circulating flume andstraight flow through flume. Straight flumes were theTm =PngRJc(1)earliest devices to quantify sediment transport andgenerally have been used to measure the bed-load ofwhere τm is the mean bed shear stress, Pw thenon-cohesive particlesh. A trial experiment on strai-ght flume shows that PAC with layer depth of0.15 mwater density, g the accelaration due to gravity, R，is fully eroded in a short time due to its fine particles,the hydraulic radius, and J the energy slope.as a result it is hard to measure the PAC erosion pro-For steady flow, J。can be calculated from thecess. So, more recently, many other devices of re-cir-Manning Strickler formulaculating type have been developed, primary to mea-sure sediment re-suspension or erosion. In this study,PAC erosion is studied using a re-circulating annularU=-R°'J!2(2)flume.The deposition characteristics of the PAC werestudied using a rotating circular flume at Tongjiwhere U is cross-sectional averaging velocity, andUniversity at Shanghai, China. The rotating circularn is the Manning's coefficient, n = 0.012.flume is with 2.0 m in mean diameter, 0.30 m in wideSubstuting Eq.(2) into Eq.(1) givesand 0.45 m in deep. A counter rotating top cover (ring)fits inside the flume with close tolerance, and makes_∪2ncontact with the water surface in the flume, so as totm= pg(3)5 R,reduce secondary circulation (Fig.2).To study the erosion rate of PAC in raw wateraqueduct by the re-circulating flume experiment, it isComputerStuddlerequired that the mean bed shear stress in the flume isequivalent to that in the actual raw water aqueduct.Y ADVThe water flowing in the actual raw water aqueduct isPristalie pump IntineHanging wall。Chifull-pipe flow, with a hydraulic radius of 1.19 m. Incontrast, the PAC erosion experiment is performed inEktexineo3mmgthe overlying water of 0.20 m depth, equivalent to a .hydraulic radius of 0.07 m. So, to achieve the maxi-lLying wallmum mean bed stress in the raw water aqueduct withSampling pointthe flowing velocity of 1.28 m/s, the correspondingMotor) Moorcross-sectional averaging velocity in the flume shouldreach 0.65 m/s.Fig.2 Schematic of experimental set-up1.4 Flume experimentPAC material is the bamboo charcoal used in theBefore the experiment, the water flowing velo-raw water aqueduct of study area, with the grain sizecity in the flume was verified by Nivus PCM Pro, anof 230 meshes (smaller than 62 μum). Flume experi-ultrasonic device to measure the profile velocity distri-ment was conducted in the following way. At thebution profile and cross-sectional averaging velocity.beginning of the experiment, bamboo charcoal wasFour outlets are opened in the flume side along theput at the bottom of flume apparatus. Afterwards thevertical direction. In this experiment, samples fromflume rotated at a constant rate, producing the desiredthe second point from the bottom are taken out conti-shear stress. The bamboo charcoal concentration innuously, feeding into a small glass vessel, and thenthe overlying water was measured as a function ofpumping into the flume by a re-circulating route. Sus-time.pended PAC concentration in the vessel is measuredAt any moment, the time-dependant variation ofby an Acoustic Doppler Velocimeter (ADV) in real-the suspended bamboo charcoal concentration can betime, which will be discussed in detail in Section 1.5.determined from the mass balance equation for thesusp中国煤化工increase in the suS-1.3 Experimental conditionpendfYHJlue to the differenceBed shear stress is the dominant factor to deter-betwlC N M H Gi the deposition ratemine erosion of sediments. In general, the followingas followsequation is applied to calculate mean bed shear stress445BSI = 10- x1003xampL(7)ndC=Eo- w,C(4)where ampl. is the average signal amplitude datawhere h is the depth of the water, C the suspe-measured by ADV system. The coefficient 10~ hasnded bamboo charcoal concentration at any moment,been introduced to avoid large values of BSI2TEquation (8) was calibraled with the field datadC/dr the slope of the C(t) graph at any mo-sets conducted with the samc circular flume instru-ment, E。 the initial re-suspension rate， w。 thement, as shown in Fig.3. lt is concluded that for thesetling velocity of bamboo charcoal paricles.laboratory tests, the best fit relationships is given asHowever, at the initial time, there is only crosioneffect in the flume, so the erosion rate can be obtainedSSC= 2.75x BSI(8)E。-n(dC)(5)withdt )。R' =0.971where (dC/d)。 is the initial slope of the C(1)1.6 Measurement of PAC bulk densitygraph.Bottom sediments are a mixture of solid particles,1.5 Measurement of PAC concenrationwaters and gas. The bulk density can be defined asA key issue in experimental monitoring is thechoice of a measuring technique of Suspended Sedi-ρ=x,P; +x.Pw +xxPg(9)ment Concentration (SSC). Several studies suggestedthat Signal-to-Noise Ratio (SNR) and backscatter in-where ρ, ρ,，p。and Pg are the true densities oftensity (BSI) may be suitable surogate measures forthe bottom sediments, solid particles, water and gas,SSC. Fugate and Friedichs20) claimed that the ADVrespectively and x，x。and x。are the volumeinstrument's acoustic backscatter amplitude may berelated to the instantaneous SSC with proper calibra-fraction of the solid particles, water and gas. respe-tion. The study was extended to the investigation ofctively.relationship between ADV backscatter intensity andConsidering P。is much smaller than P, andSSC in a subtropical system, proving the ability of anPw, the bulk density is simplified to be in the formADV to measure accurately instantaneous suspendedfine sediments for long durationsCorrelation between acoustic backscatter inten-ρ=xP +x.pPw(10)sity and suspended sediment concentration can beexpressed ast23For the water content, it can be defined as themass of the water contained in the sediments dividedSSC= ax BSIby the total mass of the sediment, that iswhere SsC is PAC concentration, BSI is ADV back-W=.(11)scatter intensity, and a is linear regression constant.m. +m, xPw +x,p。300pwhere m. and m, are the masses of water andsolids in the sediment. By solving the equation for x。 ，00I+one obtainsρ,W勇100x=1+(p,-1)w(12)2010From the definition of x，it follows that品P.= QW andthereforeFig.3 Relationship of PAC concentration and BSI中国煤化工The ADV does not output BSI directly. The BSICHCNMHG(13)is related to the signal amplitude byI +(P,-T)m446120 [240 r180，80t120 t苏40虽60 t010015020050(a) 0.10 m/s(b)0.20 mn/s240 100 r、18040 t80 t言12065(0050 2001015/s(c)0.30 m/s(d)0.45 mvs3000240E 12员I20|150 200(e)0.55 m/s(团) 0.65 mn/sFig.4 Process of suspended PAC concentration with flowing velocitywhere P。 is the true density of PAC， ρ, =was taken out to measure its water content, hereby thePAC bulk density with different compacting time2 000 kg/m3.could be determined based on Eq.(13).1.0 |.8 t1.75.6 t1.700.41.65.221.600.00.81.5552018Fig.5 Initial erosion rate under various flow velocitiesFig.6 Bulk density of PAC varying with compact timeExperiments for measuring bulk densities aredescribed as follows. PAC with the thickness of2. Result and discussions0.04 m was flled into the glass column firsly, andthen water was flled up to the level of 0.20 m depth.2.中国煤化工erosion raesThe PAC was compacted in different columns withHCNMHGonasafunctionoftime varying from I dto 15 d. After the compactingtime Is shown 1n r1g.5. I[ can De concluded that:time, water was drained from the column, and PAC44730025020000150100605015(a)1d(b)3d; 1060t0 +.150200150 200t/(e)5d(d)7d0t80t:曾6105S018(e)l1d015dFig.7 Scouring process of PAC with compacting timne ranging fromldto 15d(1) Suspended PAC concentration in water in-A, n are constants that depending on the type ofcreases rapidly at first, then increases slowly until asediment.steady state is reached. This is due to the differenceFitting curve gives the equation parameters asbetween the re-suspension rate and the deposition rate.Specifially, at initial time there is only scouringE = 0.0005t"(15)effect which makes PAC concentration increase sha-rply, afterwards co-existence of deposition and re-sus-2.2 Effects of "builk density on erosion rates.pension makes PaC increase slowly and finally reachAn experimental result of PAC bulk density withequilibrium.compaction times varying from I dto I5dis prese-(2) Steady suspended PAC concentration is rela-nted in Fig.6. It can be demonstrated that bulk densityted to water flowing velocity or mean bed shear stress.inereases with the PAC deposition, and after 10 d theThe steady PAC concentration increases with increa-compaction lowers. The results for 15 d do not differsing mean bed shear stree.gratly from those for 10 d. This is due to the increaseWith experimental results, erosion rate can bein bed compaction with PAC deposition and the corre-determined from Eq.(5), which is shown in Fig.4.sponding increase in PAC adhesive forces, pushingFigure 5 shows that the erosion rate increasespore water in PAC outside. With the increase of com-with bed shear stress in a power form, which can bepact timce. the par adhciva. forces reach steady,expressed aswhic中国煤化工d to be stable.0HCN M H Gf0.60 n/s in walerE=At"(14)flume... ...... process of PAC ofdifferent bulk density was measured, as shown iwhere E is the erosion rate, T is the shear stress,Fig.7.4484.0E(t, p)= 077"p2(17)3.0Specifically, Eq.(17) is applicable to the erosionof cohesive scdimcnts with the particle size smaller2 2.0than 62 μm in diameter, and for the scenarios of bulk1.0density varying from 1 550 kg/m' to 1 800 kg/m', andthe lateral cross-sectional averaging shear stress sma-1.50 1.s5 1.60 1.65 1.70 1.75 1.80ller than 1.80 N m 2. Equation (1 7) also shows that thevery sensitive dependence of erosion rates to bulkρI0' kgm"'density is evident.Fig8 Erosion rate of PAC for dfferent bulk densitiesReferencesFigure 8 gives the plot of erosion rate of PAC[1XU Zu-xin, YIN Hai-long. Development of coupledwith compacting time and bulk density. It demonstra-tes that the longer the PAC compacting time is, the1D- 2D mathematical models for tidal rivers[J] Journalof Hydrodynamics. Ser. B, 2004, 16(6): 767-776.less the erosion rate and the corresponding steady sus-[2] YIN Hai-long, XU Zu-xin and WANG Chen et al.pended concentration decrease with increasing PACDevelopment of early warning modeling system forcompacting time which is mainly due to the fact ofsudden chemical spill in plain river network[J]. JournalPAC bulk density increases with compacting time.of Marine Science and Technology, 2011. 19(6): 634-64[3WINTERWERP J. C, Van KESTEREN W. G. M.3. Summary and concluding remarksIntroduction to the physics of cohesive sediment inthe marine environment[M]. Amsterdam, TheBy means of flume experiments described herein,Netherlands: EIsevier, 2004, 576.the effects ofbed shear stress as well as scdiment bulk[4] LICK w. Sediment and contaminant transport indensity on erosion rates of PAC with 230 meshes havesurface waters[M]. Boca Raton, USA: Taylor andbeen examined.Francis, 2009, 45-97.The bed shear stress in flume experiment, it is[5TAN Guang-ming, JIANG Lei and SHU Cai-wen.determined according to the actual water flowingExperimental study of scour rate in consolidated cohe-sive sediment[J]. Journal of Hydrodynamics, 2010,velocity and the bed shear stress in the raw water22(1): 51-57.aqueduct of upstream Huangpu River, Shanghai.[6]GROSS T. F. DADE w. B. Suspended sediment stormExperimental results demonstrate that PAC erosionmodeling[] Marine Geology, 1991, 99(3-4): 343-360.rate exponentially increases with bed shear stress as a[7]ABERLE J, NIKORA V. and MCLEANS. ct al. Afunction of power form, and no critical bed shearstraight benthic flow-throught flume for in sin mea-stress are needed for the PAC erosidn.surement of cohesive sediment dynamics[J]. Journal ofThe effects of sediment bulk density on erosionHydraulic Engineeringerine. ASCE, 203, 129(); 63-67.LICKWMCNrates have also been studied. The bulk density is pri-[8]J. Effects of sediment bulk proper-ies on erosion rates[J]. The Science of the' Totalmarily a function of the water content of the sedimentEnvironment, 2001, 266(1-3); 41-48.and varies as the water content changes due to sedi-[9]MCNEIL J. LICK W. Erosion rates and bulk propertiesment deposition and subsequent compaction. PACof sediments from the Kalamazoo River[J]. Journal ofbulk densities varying with compaction times varyingGreat Lakes Research, 2004, 30(3): 407-418.from 1 d to 15 d have been measured, and erosion[10] BOUDREAU B. P., JORGENSEN B. B. The benthicboundary layer:; Transport processes aand biogeo-rates for different bulk densities determined. For eachcinemistry[M]. Oxford, UK: Oxford University Press,shear stress, the decrease in the erosion rate as thebulk density increases can be clearly seen.[] MADSEN J. D, CHAMBERS P. A. and JAMES W. F.From these experiments, it can be determinedet al. The interaction berween water movement, sedi-that the erosion rate is a function of bulk density andment dynamics and submersed macrophytes[] Hydro-biologia, 2001, 444(1-4): 71-84.bottom shear stress. In general, the data can be appro-De BOER W. F. Seagrass-sediment interactions, posi-ximated by an equation of the formtive feedbacks and critical thresholds for occurence: Areview[J]. Hydrobiologia, 2007, 591: 5-24.E(t,p)= At"p" .(16)[13]ABERLE J, NIKORA V. and WALTERS R. ffcts ofbed material properties on cohesive sediment erosion[].. M larine Geologv. 2004. 207(1-4): 83-93.where E(r.p), A, n and m are constants. The[14]中国煤化工and JAMES C. S. Mea-equation with n=1.13 and m=-9.80 shows ;F and lransport with thegood ftting to the data for all of the experimentalCHC N M H Gand transport fume[J].，Journal o1 Hyaraule Engineering, 2003, 129(11):results, that is862-871. .449[15] WANG Y. H. The interidal erosion rate of cohesive[20] FUGATE D. C., FRIEDRICHS C. T. Determining con-sediment: A case study from Long Island Sound[].centration and fall velocity of estuarine paricle popula-Estuarine, Coast. Shelf. S, 2003, 56(5-6): 891-896.tions using ADV, OBS and LISST[J]. Continental[16] LAU Y. L.. DROPPO I. G. and KRISHNAPPAN B. G.Shelf Research, 2002, 22(11-13): 1867-1886.Sequential erosion/deposition experiments demonstra-[21] CHANSON H, TAKEUCHI M. and TREVETHAN M.ting the effects of depositional history on sediment ero.Using turbidity and acoustic backscatter intensity assion[J]. Water Research, 2001, 35(11): 2767-2773.surrogate measures of suspended sediment concentra-[17] GRABOWSKI R. C., DROPPO 1. G. and WHARTONtion in a small subtropical estuary[]. Journal of Envi-G. Erodibility of cohesive sediment: The importance ofronmental Management, 2008. 88(4): 1406-1416.sediment properties[J]. Earth-Science Review, 201 I,[22] NIKORA V, GORING D. Fluctuations of suspended105(3-4): i01- 120.sediment concentration and turbulent sediment fluxes in[18] DEBNATH K., NIKORA V. and ABERLE J. et al.an open-channel flow[J]. Journal of Hydraulic Engi-Erosion of cohesive sediment: Resuspension, bed load,and erosion patterns from field experiments[J]. Journal[23] VOULGARIS G., MEYERS S. T. Temporal variabityof Hydraulic Engineering, ASCE, 2007, 133(5): 508-of hydrodynamics, sedimncnt concentration and sedi-ment setting velocity in a tidal creek[]. Continental[19] ABERLE J. NIKORA V, and MCLEAN s. et al. Sirai-Shelr Research, 2004, 24(15): 1659- 1683.ght benthic flow-through flume for in situ measurementof cohesive sediment dynamics[J]. Journal of Hydrau-lic Engineering, ASCE, 2003, 129(1): 63-67.中国煤化工MHCNMHG