DRINKING WATER - LAB EXPERIMENTS LAB EXPERIMENTS Granular filtration Hydraulics
granular filtration - ydraulics lab experiments Framework Tis module explains te lab experiments on te ydraulics of filtration. Contents Tis module as te following contents: 1. Objectie 2. Experiment set-up 3. Teory 3.1 Hydraulic aspects 3.2 Backwasing 3.3 Caracteristics of te filter material 3.4 Furter application 4. Procedure 4.1 Filter materials 4.2 Head losses and expansion measurements 4.3 Canging te filter bed 4.4 Siee analysis 5. Elaboration Data forms
lab experiments granular filtration - ydraulics 1. Objectie Wen designing rapid filters for drinking water treatment knowledge of te pysical properties of filter materials is required. Parameters suc as grain size and sape, sape factor, bed porosity and specific surface can be found experimentally by measuring te ead losses oer te filter bed. Furtermore te bed expansion for different flow rates can be measured to get information about te quantities of water necessary for backwasing. 2. Experiment set-up Te pilot plant is a perspex column wit a diameter of 67 mm and an area of 35 cm 2 can be used in upward and downward flow direction (see figure 1). Tap water is pumped from a storage tank wit a constant temperature. Te flow elocity for downflow filtration is controlled downstream of te filter bed. Connections for pressure gauges are spirally mounted on te column. Te pressure is measured wit water manometers. 3. Teory 3.1 Hydraulic aspects Te most important part of a rapid filter is te filter material. It as to be free of clay, silt and organics and must be sustainable. Sand is an appropriate filter material. Howeer, te ariation in grains sizes is large. Tis induces stratification during backwasing because te finer particles moe to te top of te bed, wile te bigger particles are moing downward. Te finer particles gie te igest remoal but te filter will clog faster. One as to apply a filter bed as uniform as possible to preent rapid clogging.wit a certain type of filtersand different measurements can be used to determine te specific properties of te sand. According to Carman-Kozeny a linear relationsip exists between te ead loss of te clean bed and te laminar flow elocity: H= 180 ν (1 p) g p d L [1] in wic: ΔH = ead loss (m) ν = kinematic iscosity (m 2 /s) g = graity acceleration (m/s 2 ) p = porosity (-) = elocity (m/s) d = ydraulic diameter of sand grains (m) L = bed eigt (m) Te porosity is calculated from te bed olume and te mass and density of te filter material. Te Carman-Kozeny relationsip can be tested by plotting te ead loss as function of te filtration elocity. Tis sould gie a linear relationsip. Te slope of tis line can be used to calculate te ydraulic diameter of te sand grains. d 180 ν (1 p) = I g p 0 wit H I0 = L [2] Figure 1 - Experiment set-up ydraulics of filtration tis equation is only true if Re 5.
granular filtration - ydraulics lab experiments Re d 1 1 p = ν [3] Te quotient of and I 0 is te permeability coefficient K (m/s) wic is a measure for te ease of flow troug a bed of grains. in wic: p e = porosity of te expanded bed (-) L e = te eigt of te expanded bed (m) Te porosity of te expanded bed can be calculated from te original porosity and te expansion according to: p + E pe = 1 + E [5] E = (L e L)/L = expansion Rewriting gies (1 p) L = (1 p e ) L e [6] At te ery moment were te ead loss equals te weigt of te bed expansion begins, reflected by te following equation: f w H max = (1 p) L r r rw [7] in wic: r w = density of water (kg/m 3 ) r f = density of filter material (kg/m 3 ) Equalizing [4] wit [7], substituting [6] and rewriting [8] gies an equation to find d for tis flow regime: g r r 1.2 f w e 1.8 = d 0.8 0.8 130 ν rw (1 p e ) p [8] Te iscosity is dependent of te temperature according to: Figure 2 - Hydraulics of filtration in te laboratory 3.2 Backwasing During backwasing te flow is in te transitional region between laminar and turbulent flow. Te Carman-Kozeny formule is not alid anymore. For tis type of flow only empirical formulae can be used. A good approac is: ν H= 130 (1 p ) 0.8 1.8 1.2 e 3 1.8 g pe d L e [4] 497 10 ν = (T + 42.5) 6 1.5 [9] 3.3 Caracteristics of te filter material Te filter material used in tis experiment consists of a sieed fraction of particles of ariable size between te lowest and widest mes. A siee analysis gies information about te grain size distribution. 4
lab experiments granular filtration - ydraulics Coefficient of uniformity Te spread in grain sizes sould not be too big to aoid stratification during backwasing. Te spread is caracterized by te coefficient of uniformity U = d 60 /d 10. Preference is gies to a alue of U 1.2. Specific diameter and sape factor Te so-called specific diameter of te grains can be found from a siee analysis wit: W W1 W W Wn = + + +.. d ϕ S S ϕ S S ϕ S S ϕ S S + [10] in wic: W = total weigt of te sand sample (g) W i = weigt on siee i (g) S i = mes of siee i (mm) ϕ i = sape factor for fraction between siee i and siee i+1 (-) Te sape factor is a measure for te deiation of a sperical particle. For an exact calculation of d s te sape factor of te subfractions sould be known. Because tese factors cannot be quantified tey are assumed to be 1. From formula aboe is follows tat: [11] Te ydraulic diameter d as it is determined from te experimental results depends on te bed conditions. Different alues will be found for low porosities, ig porosities and expanded beds. Te ratio between te ydraulic diameter and te specific diameter gies te sape factors for te different ydraulic conditions: ϕ e ϕ ϕ l s 1 1 2 2 2 3 3 3 4 n n n 1 W W d S S + i i = s 1 i i 1 = d /d s for expanded beds = d /d s for ig porosity fixed beds = d /d s for low porosity fixed beds Effectie diameter Te ydraulic diameter for filter materials wit unknown sape factors can be approximated In practice by te so-called effectie diameter de, defined as te d 10 from a siee analysis. From te results is sould be proed tat te ratio between te ydraulic and effectiee diameter is close to 1. Tis means tat te sape factor is determined as ϕ = d e /d s 3.4 Furter application If te specific diameter d s, te sape factor φ and te density are known, tan te flow elocity can be calculated for any expansion using equations [5] and [8], wit te ydraulic diameter d = φ d s. Te porosity tat corresponds wit tat expansion is found wit equation [6]. Te porosity of te fixed bed is found from te mass and density of te filter material and te bed olume. Te ead loss of a clean filter bed is calculated wit equation [4]. On te oter and te ydraulic diameter d for any unknown filter material can be calculated from a single expansion experiment using equations [5] and [8]. Next te flow elocity can be calculated for oter expansions. Te porosity is calculated as explained before. Te experimental results are used for te design of a real-scale treatment plant. It is assumed tat te same filter material is used as in te experiment. 4. Procedure 4.1 Filter materials Te experiment is performed wit two filter materials: - Meuse sand, commercial fraction 1.0 1.6 mm, density 2,650 kg/m 3. - Antracite (an adapted mineral coal), commercial fraction 1.4 2.5 mm, density 1,450 kg/m 3. Te masses of bot filter beds are gien during te experiment. 5
granular filtration - ydraulics lab experiments 4.2 Head losses and expansion measurements to determine te ydraulic diameter Te bed is expanded sortly in order to start wit a ig porosity. Flow rates ae to be conerted to filtration elocities. Te porosity is calculated from te eigt of te filter bed (fixed or expanded) and te density and mass of te filter material. - fill te column alfway wit tapwater. - pour te filter material into te column using a funnel (up to about 1 meter). - backwas te bed wit a ig flow rate for 2 minutes. - switc te pilot plant in te filtration mode. Head loss measurements - measure te water temperature before starting te measurements. - measure te ead losses in te bed for 5 flow rates: 4 8 12 16 20 l/. - tap on te column carefully to decrease te eigt of te filter bed, meanwile reducing te porosity. - measure te water temperature. - measure te ead losses in te bed for te same flow rates: 4 8 12 16 20 l/. Expansion measurements - start wit 30% expansion and reduce te flow rate stepwise. Coose 6 to 10 alues on te upstream flowmeter. Measure te eigt of te expanded bed and te manometer alues. Te alues can be conerted later to backwas elocities and expansion E. - measure te water temperature wen te experiment is finised. - cange (if necessary) te filter bed, according to te procedure gien in 4.3. 4.3 Canging te filter bed Te following steps must be taken to cange te filter bed: - close te water pressure gauges and empty te column completely. - unscrew te bottom of te column. - wen remoing te bottom catc te material in a bucket. - clean te column and replace te filter bottom. - refill te storage tank wit water. - control if te pressure gauges are closed. 4.4 Siee analysis Take 100 gram of bot filter materials and perform a siee analysis wit te sieemacine. Mes dimensions according to DIN. Select eigt siees of wic te biggest is bigger tan te largest expected diameter of te filter material and of wic te smallest is smaller tan te smallest expected material diameter. Weig eery siee before saking. Siee time 5 minutes on amplitude 4 (green button), ten 5 minutes on amplitude 1. Weig eery siee again. 5. Elaboration Te elaboration ereafter is for bot filter materials. Gie your comments were possible. - work out te siee analyses and find te coefficient of uniformity U, te specific diameter ds and te effectie diameter de. - plot in one diagram I 0 as function af te flow elocity for te ead loss measurements; ceck te Carman-Kozeny formula. - determine for bot porosities te permeability coefficient for te laminar flow region. - plot in one diagram for bot filter materials te total ead loss oer te bed and te percentage expansion as function af te flow elocity. - find te ydraulic diameter from te ead loss and expansion measurements and compare bot alues. - next find te tree sape factors of te filter material and explain te differences. - ceck if te ratio between te ydraulic and te effectie diameter is about 1. 6
lab experiments granular filtration - ydraulics Data forms Data form 1 Group number: Date: Head loss measurements sand: 1,0-1,6 mm, density 2,650 kg/m 3. Mass of filter material: kg Hig porosity Flow rate Lowest point Point aboe filter bed ΔH I 0 * 10-2 (m/m) (10-3 m/s) 4 8 12 16 20 Begin temperature ( o C) : Bed eigt : End temperature ( o C) : Pressure gauge bottom : Aerage temp. ( o C) : Resistance oer : Low porosity Flow rate Lowest point Point aboe filter bed ΔH I 0 * 10-2 (m/m) (10-3 m/s) 4 8 12 16 20 Begin temperature ( o C) : Bed eigt : End temperature ( o C) : Pressure gauge bottom : Aerage temp. ( o C) : Resistance oer : 7
granular filtration - ydraulics lab experiments Data form 2 Group number: Date: Head loss measurements antracite: 1,4-2,5 mm, density 1,450 kg/m 3. Mass of filter material: kg Hig porosity Flow rate Lowest point Point aboe filter bed ΔH I 0 * 10-2 (m/m) (10-3 m/s) 4 8 12 16 20 Begin temperature ( o C) : Bed eigt : End temperature ( o C) : Pressure gauge bottom : Aerage temp. ( o C) : Resistance oer : Low porosity Flow rate Lowest point Point aboe filter bed ΔH I 0 * 10-2 (m/m) (10-3 m/s) 4 8 12 16 20 Begin temperature ( o C) : Bed eigt : End temperature ( o C) : Pressure gauge bottom : Aerage temp. ( o C) : Resistance oer : 8
lab experiments granular filtration - ydraulics Hydraulics of Filtration Data form 3 Group number: Date: Expansion measurements sand; temperature: Q (m) (m/) Exp. (%) o C Manometer readings 1 2 3 4 5 6 7 8 Q (m) (m/) Exp. (%) Manometer readings 9 10 11 12 13 14 15 16 Expansion measurements antracite; temperature: Q (m) (m/) Exp. (%) o C Manometer readings 1 2 3 4 5 6 7 8 Q (m) (m/) Exp. (%) Manometer readings 9 10 11 12 13 14 15 16 9
granular filtration - ydraulics lab experiments Data form 4 Group number: Date: Siee analysis of 100 grams of sand Mes (mm) Weigt of empty siee (g) Weigt of siee + sand (g) Weigt of sand (g) Going troug % (cumulatie) 1.8 1.6 1.4 1.25 1.12 1.0 Bottom Siee analysis of 100 grams of antrite Mes (mm) Weigt of empty siee (g) Weigt of siee + antracite (g) Weigt of antracite (g) Going troug % (cumulatie) 2.5 2.24 2.0 1.8 1.6 1.4 Bottom 10