a = To: 300 Cooke unl eee aa He Burt Wilson C1 For your information oniy 1 Respond to sender by Bob Schroeder Copies: c. p. silverblatt RL. Ennett FILTRATION THEORY 0. =. Albertson Since filters aze Eimco's major business, it certainly Seems that we should have a thorough understanding of filtration theory. Toward that purpose, 1 have derived some expressions for the rate of filter cake formation that avoid the use of the terms V (for volume of filtrate) and ¢ (for the slurry concentration expressed as wt. of solids per unit volune of filtrate). Both of thess terms are used frequently in the literature and are probably > quite convenient when considering pressure filtration, Since the filtrate is usually the valuable product in most pressure filter applications. However, when dering vecuun filtration we are usually i in the solids filtration rate rather than the licuia filtration rate. Therefore, an expression for the rate of cake formation that avoids the V and ¢ terms should be of greater iuturest to us~ rested The choice of an equation from which to start the derivation depends on the reference source that one selects. The same is true of the notation that is used. The starting equation is usually derived from Poiseuilie's equation and is typically in the Zollowing form, which expresses the instantaneous rate of filtrate flow as the ratio of the Giving force to the product of the viscosity ané the sun of the filter cake and filter media resistances: av 2 Bd; = ee mM Pe + aI ‘This can be converted to: (2eege « November 20, 1976 The notation that will be used herein is as follows: Og = cake formation tine V = volume of filtrate u = Liguia viscosity a = filtration area ap = pressure drop across the filter cake % = average specific filtration resistance, (un: length/mass) ¢ = feed slurry solids concentration in terms of mass of solids par unit volune of filtrate R, = resistance of the filter medi, (units of length”) W = mass of ary solids in the Zilter cake per unit area Liguid density we: Eeaction colide in the feed alurcy S. = wt. fraction solids in a formed (but undewatered) Filter cake Integrating equation (2): eee ee Soren Sl] - “ Since w=, ~ “) 5 [e+ | “)Sage 3 November 30, 1976 ww [ie & * ape [s+ *| a It should be clear that the rate of cake formation, expressed as we usually express it (in terns of wt. of Gry cake solids per unit area and per unit time), is equal to W/0g. Rearranging equation (7) yields: 2:8] ® Fle This is a rather simple equation, and it avoids the use of the V term. However,.the c term is one that has an ° vnusual definition, and this definition is probably responsible for nore misunderstandings regarding filtration theory than any other factor. It therefore is beneficia? to replace the c term with encther expression that refers to the feed slurry solids concentration in terms that we can readily understand, i.e., wt.t solids. Actually, the te: 1 will use is 5, the we. fraction solids in the feed slurry, which is alnost the same thing ‘The definition of c is the weight of dry solids in the feed slurry per unit volume of filtrate. The liquid remaining in the formed (but undewatered) cake is, of course, not part of the filtrate, but it certainly was part of the feed slurry and it mist be accounted for. Notice that I refer to the formed but undewatered filter cake. This is because the nechanisn of dewatering oF draining @ filter cake by displacing the entrained liquia with air (as occurs in the cake drying portion of a vacuun filter) has nothing to do with the mechanism of cake formation, which is described by equation (8). Therefore, we must refer to the wt. fraction of solids in a formed but undewatered filter cake, which I will denote with the tern §.- ‘The definition of ¢ could be expressed as: wt. of solids volume of filtrate @Ss November 30, 1976 By dividing c by p,' the density of the Liquid, we can see that: ¢ _ vt of solids 5” WE. of filtrate Qo) Also, we can define $_ as: wt. of solids So" GE. OF formed cake Oy wt. of solide ° WE. OF slurry - we. of Filwate @2) ‘Therefore, we. of filtrate = wt. of slurry ~ We: Of solids (13) and, e wt. of solids 2” Wao of sluzy ~ WE: of Solids” aay PU we. of slurry - ¥& oF Be By multiplying the term containing 8. by (wt. of slurry/wt. gf slurry), I cas) ae) anPage 5 November 30, 1976 Note that we have gone from equation (1) to equation (17) using only mathematics, with no assumptions at all, so that if one accepts the validity of equation (1), then equation (17) is rigorous. his equation should be used by changing only one variable at a tine on the right-hand side in order to see what effect it has on the left-hand side (the "form filtration rate"), For instance: If dp is doubled, the form rate will double. wote that this can only be due to 6, being halved: Since we didn't change W in the right-hand side, therefore W cannot Change in the left-hand side either (since they are the same}. This is filtration at constant cake thickness. On an actual vacuum filter, if the Filer operator doubled ‘4p, in order to get the sane cake thickness as before he would have to increase the filter speed by a factor « of 2 If =~ were doubled, the form rate would double. once pa E again, this is true only if the cake thickness is held constant by changing the filter cycle time. he logical question that follows is: What is the effect of changing a variable if you don't change the filter speed? In other words, filtration at constant cycle tine. In order to evaluate this and at the sane tine avoid the use of a complicated equation, we must make the simplifying assumption that the resistance due to the filter medium, 8, is small compared to the resistance due to the filtel cake, W#/2, This is usually the case in most vacuum filter applications. After setting Ry equal to zero, we obtain: w. 2000 -~ ae tt as Rearzanging: 2ape w = 0, 268 d (is) ee TS 5S tant lentinichiviiahiRORIAE aac in itPage 6 November 30, 1976 and: rr St Ss wf ae (20) Be By taking the logarithm of both sides, ve can see that if W is plotted against 0, on log-log paper, a straight line of 0.5 slope should fesult, (and it generally does). By dividing W by 0, to calculate the form filtration rate, we obtain: BRS ,_S pee an ey By taking the logarithn of both aides, we can see that £2 the torn filesation rate is ploteed sgainae Se on log-log paver, @ straicht line of minus 0.5 slopé should yeaule, (and dt generally Aone) We can see that changing a variable on the right-hand side of equation (21) at constant cycle tine will result in a change in the form fileration rate that ie proportional to the square root of the ratio of the new variable Ragnitude fo the O18 variable magnitude. As an example, if a drum filter attains a filtration rate of 50 1bs/hr-£e2 at 2 vacuum level of 20 inches Hig and suddenly the vacuum drops to 10 inches He, the ney filtration rate will be (50) /10720 = 35.4 ibs/ne-ft?, ‘This assumes that the cycle’ tine didn't change, the submergence didn't change, the feed solids concentration didn't change ,—-nathing but the vacuum level and the cake thiness changed. And it also assumes that the new cake thickness is able to discharge from the filter. ‘This should point out the fact that whenever you are considering the effect of a change in filtration conditions on the filtration rate, you must first decide: Are you interested in the effect at constant cake thickness or at constant cycle time? You can't start to answer the question until, this is decided.Page 7 Novenber 30, 1976 Another point: Equations (17) and (21) show us that the filtzation rate is not proportional to the we.t solids in the slurry, and if's not proportional to the square root of the ratio of the two concentrations in question. The effect of a change in the feed solids concentration is a bit more complicated than that because vou can't analyze this effect using 2 sound theoretical basis until you have determined 5,, the wt, fraction solids in a formed but undewatered°cake. “Ad far as 7 know, this has been done on only three slurries in all of Bines's technical history. The attached table lists factors by which a meaéured filtration rate should be multiplied in order to find tha Eilevation cea te snes seaied Bet jen Bob Schroeder , attachment,TABLE T variable That Constant cake” Constant Filter” Is Changed ‘Thickness Cycle Tine [ey Vacuum level = ay w 2 7 Temperature ae pes i, a Feed Solids Concentration submergence Submergence, /Summergence, Weetecktve) ‘Babnergence, submergence, *yaitiply the filtration rate at condition "1" by this factor to get the rate at condition "2". wns lich lili EE CRG cis GRAS SR RTE