Vol. 6 3 Trans. Nonferrous Met. Soc. China Sep. 1996 DETERMINING THICKENER UNDERFLOW CONCENTRATION AND UNIT AREA™ Kang Yong, Hu Xiaomin, Deng ( anglie, Luo Q Department of Mineral Engineering, Northeastern University, Shenyang 110006 ABSTRACT has heen discussed, The relationship between the characteristies and the sediment curve in a batch sedimentation and the origination of characteristies has heen analysed by two phases flow wave theory According to the sedimentation theory and experimental data, two mathematical equations for caleulating un~ derflow concentrations and unit are: continuous thickener was developed. and unit area of a continuous thickener. Key words batch sedimentation curve 1 INTRODUCTION Thickeners are used industrially to reduce the amount of liquid in slurries containing such materials as concentrates, tailings and sewage. Traditional design procedures for sizing gravity thickeners have heen based on works by Coe and Clevenger!!!) Kyneh!?!, ‘Talmage and Fitch!*!. The shortcomings of those methods are the part cle settli i is only involved and effects of the ris- ing compression zone upon action of suspension settling are not considered. Recently many schol ler! has dew ars have revised Kynch theory. veloped an equation correlating the tration of the layers that reach the supernatant— lid cocen- suspension interface as a function of the varie tions of its height and the compression disconté nuity height ys time. Font!) developed and dis cussed a mathematical model that may be used to work out the relation between the settling rate and the solids concentration, but this procedure is complicated to calculate the underflow concen tration and the unit area of a continuous thicken The aims of this work are as follows: (1) To analyze origi theoretically nation of characteris ics D_ Received Aug, 21, 1995: accepted Jan, 30, 1996 of a continuous thickener were derived. thickener unit area And a new procedure to design a The procedure is reasonable and valid to calculate underflow concentration underflow concentration (2) To deduce mathematical models that may he used to work out underflow concentration and unit area of a continuous thickeners (3) continuous thickeners. (4) To and unit area of a continuous thickener by various © develop a new procedure to desisy calculate underflow concentration methods and compare theirs results. 2 ORIGINATION OF CHARACTERISTICS Characteristics are lines, whose points corre spond to layers with the same solids concentrar tion and the same settling rate. Kynch!?! ase sumed that all characteristics originate from ore Piller *! gin of the coordinates( height vs time) . recognized that characteristics rise from a sedé ment surface curve in height vs time plot. Fitch'*! and Font!! deduced that every character ie rises from the sediment tangentially. Fiteh does not explain the reasonable result, And Font only interprets origination of characteristics geo metrically, not theoretically In uncompressible two phases flow, there The are continuous waves and ulated waves. velocity of continuous waves and stimulated waves is respectively:. Nonferrous Met. Soc. China Sep. 1996 = oc_ 2G Vo= oe oe (9) or Vs (2) rer eee 0) (10) and the total flux of a hatch sedimentation (3) (4) (5) In Fig. 1, H (height of supernatant— sus- terface) and L (sediment height) are pension plotted against time. ‘The coordinates( + dt H+ dH) and ( t, 1 ) of two adjacent points in the straight line portion of the curve of HT vs tare related by the follow ing equation: C= C2 (6) ie. OC, ay - Olt* aq aH = 0 mM Consider two layers at H+ dH and H . In time dé, the accumulation of particles between the two layers is the difference between the flux of particles Gin through the upper layer and the flux G2 out through the lower layer, per unit edt, (G2- Gia (8) divided by dH dt, the equation becomes, Concentration di height vs time plot Fig. | ontinuity 17 dc OH Combining eqns. (4) and (10), we derive the follow ing equation: ay (12) we find, (13) Comparing eqns. (7) and (12), y.2- we dr The above equation indicates that the slope the velocity of continuous pa e in batch settling ex- of H vs t curve equi waves that are provided by soli les passing down through liquid ph periments. That is to say, the curve of Hf vs ¢ is the locus of the continuous waves. When a fast continuous wave catches up with a slow one, the two waves overlap and cause a stimulated wave, At the begining of particles settling, = 0, sus pension is uniformly successive( the gradient of When +> 0, the solid particles, which exists near the bottom of the concentration is zero.) the cylinder affected by gravity force, move down to the bottom and then stop there, Suppos ing the velocity of the first continuous wave neat the cylinder hottom is Ve. then lim Vier Ve whent > 0. Vai= 0. In the front of the stimulated wave, the sok (14) id concentration is discontinuous, as shown n Fig. 2. When the velocity of the first stimulated wave reduces to . the following continuous waves overlap each other and hecome stimulated waves again and again, then move down to the hottom of the cylinder (actually the stimulated waves are caused successively). The concentrir tion ontinuity of the front of stimulated waves propagates upwards in opposite direction of the velocity of the stimulated waves. Since the solid particles carried by the stimulated waves cumulate at the bottom, the liquid med) which is compressed by the solid particles, moves upwards, and so the intensity of the stimulatedNa3 Vol. 6 Determining thickener underflow concentration and unit sae ed waves is L= f(t) (17) then v,= 2h. dtu (18) dt dr Combining eqns. (15) and (18), we derive a (19) dt dr The general form of the characteristic equr tion!*! is. L= mt Lo (20) £ wu, i (21) y when Lo= 0, p= 4 (22) Fig. 2 Sediment developing process Above all, the characteristics rise from the waves at the upper interface of the concentration discontinuity, is weakened, and a series of contin= uous stimulated waves were formed. The position of the continuous stimulated waves corresponded in zone © ,as shown in Figure 2. The upper in terface of zone C is adjacent to the front of the continuous waves, and the lower interface is next to stimulated waves. When zone B disappears the last continuous wave overlap with the weak stimulated wave in zone € and moves down until the sedimentation process is over According to the above analysis, the stimu= lated waves in a batch sedimentation cause the concentration discontinuity. The propagation of the concentration discontinuity results from accu= mulation of the solid particles carried by the stimulated waves. es the veloc Eqn. (5) expre ly of the stimulated waves moving down. Since the solid particles at the bottom of the cylinder much dense and uncompressible, the velocity dé rection of the concentration discontinuity is oppor site to that of the stimulated waves. Supposing w stands for the velocity of the concentration dis continuity. then _ GG a wee We- oe (15) which is identical to that in Kyneh theory, us (16) C\- C2 Assuming the locus function of the stimulat= a sediment tangentially. Font’ 7! view is correct. 3. DERIVATION OF EQUATIONS Fig. 3 shows the suspension sedimentation curve of a bateh experiment. We may choose such a point at the first falling rate period! ®! as B(t2, H) and draw a line from this point tan gential to the sediment curve. The tangetial point is A(t), £1) . and the intereept on the ax~ is of ordinate is H) then the line HB is a ch acteristic one. Supposing this characteristic Interface and characteristics in batch sedimentation concentration is C2, Font!”! derivesans. Nonferrous Met Soe, China Sep. 1996 Coll = Jae al (23) In Fig. 3, the curve is drawn according to the experimental data, therefore, if points A and B have been chosen, values of ¢) and 2 are iden tified, and Eqn. (23) may becomes Colt a t OF a LOM el Fig. 3 shows AH BH; © ACBA (25) then —1__ be 1 26) Hp- bi t= Him Hy Substituting Eqn. (26) into Eqn. (24), we obtain Colly to b W.-W a e a! (27) According to Yoshioka construction! *!, the relation of Gm C, and Vis ChmIEcs (28) Fitch! found that when the initial condi tions are identical, the downward velocity V, im= parted to the system by underflow withdrawal of a continuous thickener in a steady state must he negatively equal to the upward propagation veloc ity w of the compression discontinuity with re spect to suspension, and consequently Vi=- u (29) According to Kynch theory. we obt = & us Oe (30) therefore: (31) the sediment has heen dense consequently, Vi= 0, and Eqn. (35) ber comes Vj=-us- (32) v Vo Cos C14 A) = Cx 1- “7) (33) In Fig. 3, Hi= Hz 7 (34) Ho= HH, ue (35) Substituting Eqns. (34) and (35) into Eqn. (33), we derive Has I ay (36) Combinin qns. (28) and (36), we obtain alla __t 4 Ha- Hy tn t n= th (37) +( 29) .(35) and exp[- Combining Eqns. ( 28) (37), we deduce , _ Lol Cn = exp/ n= Pn oy and the unit area of a continuous thickener can be 1 (38) expressed as ea (39) Substituting Eqn. (38) into Eqn. (39), we acquire b= et Ce iCid pentane ee by) From the above derivation, we obtain two equations to calculate the underflow concentrer tion and the unit area of a continuous thickener according to a hatch sedimentation curve. 4 CALCULATION PROCEDURE The new procedure is as follows: (1) Suspension should he mixed, whose int tial proper continuos are identical with those fed to a being designed. After a hatch sedimentation experiment is over, the sed thickener mentation curve be drawn out( Hf and L vst ). shown in Fig. 4. (2)Any point such as B(t2, H 2) at the first falling rate period of the settling curve is chosen, A line tangential to the sediment curve is drawn from point B, the tangential point is A(t, L1) . and the intercept on the axis of ordinates is Hy (3) Substituting values of ¢) 2 sbi dl2 Co and Ho into Eqns. (37) and (40), values of and q can be worked out. (4) Rep and (3). choosing several other points and calculating vak ting processes ( 2)Vol.6 a3 Determining thickener underflow concentration and unit area ome Fig.4 Discontinuity height vs time ues of C,, and q at every point. (5) A plot of C, vs q can be drawn out ac cord nd the r i to values of ( Cy. q)s able values of q is obtained from relative values of C, designed in a new continuous thickener. (6) The total ar the thickener’ s capability is identified. can be worked out when 5 BATCH EXPERIMENTAL RESUI In order to verify the new model, mag 1 d- Co= 8.70%; 5~ Ci (4) ~ unflocculated suspension; Coe 4.08% netite tailings suspension was chosen. Expert flocculated nts were separated into two series and unfloceule Every series in- ial volume concentration 6.29% . 8.70% . 11.2! cluded six v 1.93% » 4.08% . and 14. 1%. The ~ 200 mesh weight percent of the mar is 80.0% . The solid density was 2610kg/ m’, measured with distilled water using a pyc te nometer. The tests were carried out at 25°C or so in a graduated plexiglass cylinder, whose valid height is 600mm and the diameter of cross section is 55mm The curve of discontinuity heights vs time at different initial concentrations are plotted in F 5. 6 COMPARISON OF PROCEDURES TO CALCULATE C, AND ¢ In order to verify the new procedure valid ty, values of C, and q for a continuous thickener calculated by this method were compared with those by Talmage Fitch!*! and Oltmann method which are usually used in designing a continuous thickener. However, much practical experience 1.2%: 6 Cy 14. 1% (b) = floceulated suspensionvue rans. Nonferrous Met. Soc. China Sep. 1996 shows that values of q worked out by Talmage Fitch method are greater than the real results, and smaller by Oltmann method The procedures of Talmage- Fiteh and Olt- The unit area of a continuous thickener is worked out by mann methods are shown in Fig. 6. each of the two methods hased on Eqns. (41) and (42) Collo= CH, (41) L T= Coy (42) 6 Construction of Talmage— Fite! Oltmann methods The underflow concentration of a continur ous thickener was worked out by the new proce: dure based on Eqn. (37) according to experimen tal data of the various initial concentrations: 1.95% 6.29% and 14. 1%, respectively. Unit area of a continuous thi ener is calculated as fok lows: (1) by the new procedure hased on Eqn (40), (2) hy Talmage- Fiteh method and Olt- mann method, hased on Eqns. (41) and (42). Comparison of the results by three methods is shown in Fig. 7. respectiv 7, we may observe that vak ues of unit area of a continuous thickener olr In view of Fig. new method lie between those. almage- Fitch method and Olt= mann method. ‘These phenomena prove that the new procedure is more valid in designing a com tinuous thickener than Talmage~ Fiteh and Olt tained by the worked out by mann methods. We find the higher underflow concentration is, the smaller unit areas is. When the — init concentration identified, there is a maximum value for the underflow Fig. 7 Unit areas ys underflow concentrat for a continuous thickener ( Co = 6.29%) 1 —Talmage- Fitch method: 3—Olimann method (2) ~ unfloceulated suspension: the new method: (b) ~ floceulated suspension concentration of a thickener. When initial and underflow concentrations are identified, the unit area of a continuous thickener processing the flocculated suspension is much smaller than that doing unflocculated suspension. For example. when Co= 6.29%, C,= 53.6%, the unit area of the former is 2.40 x 10°m?/(m'*s) latter is 2.24 105m?/(m*+s) the ability of a continuous thickener for handling, flocculated suspension is 9. 35 times greater than 1 the In other wore that for doing the unflocculated suspension.Vol.6 Na3 Determining thickener underflow concentration and unit area 15 By this new procedure, we can work out the underflow concentration of a continuous thickener, and identify the optimun values of C, and q for a continuous thickener being designed with computer 7 NOTATION V,, —vontinuous wave velocity, m/s: V, stimulated wave velocity, m/s; G volumetric settling flux of solids, m°/ (m G Slower interface flux of the concentra tion discontinuity, m'/(m?+s) 5 G2 —upper interface flux of the concentrax tion discontinuity, m°/ (m H_ height of a descending interface. mm: C\, Cy —volume percent of solids for sedi mentation curve at points 1 and 2, %: u “upward propagation rate of the com pression discontinuity, m/s: H yy —the ordinate of point C in Fig. 3, mm; L, sediment height at ¢;, m; Hy —value of H at tz, m: ht Slime, H, “intercept height of a tangent to the curve Hy = g(t) on H axis, mm; H, —intercept height of a tangent to the curve Ly = f(t) on L axis, mm; Ho initial value of H at t= 0, mm; Gu the maximum flux of solids, m*/(m? V, —lownward velocity of pulp in a com tinuous thickener resulting from underflow with drawal, m/s: mT (mi Co initial feed concentration, % : C, underflow volume fraction of solids, q “unit area of a continuous thickene t, —value of ¢ by Talmage Fitch method and Oltmann method, REFERENCES Coe HS, Clevenger GH. Tran lust Min Eng, 1916, 60: 356, Kynch G J. Tran Faraday Soe, 1952, 48: 166. Talmage W P, Fitch EB. Ind Eng Chem, 1955, 47. 38 Yoshioka et al. Kagaku Kogaku, 1957, 21: 66 Tiller F M. AIChE Journal, 1981, 27: 823. Fitch EB, AIChE Journal, 1983, 29: 940. Font R. AIChE Journal, 1988, 34: 229, Svalovsky L. Solid Liquid Separation( 2nd edt.) London: 1990: 180. ed by Wu Jiaquan)