Global J. of Engg. & Appl. Sciences, 2012: 2 (2) Research paper: Pushpendra Kumar Sharma et al., 2012: Pp.178-182 MODELLING OF COD REDUCTION IN A UASB REACTOR Pushpendra Kumar Sharma, Nasim Ahmad Khan* and Sohail Ayub* Dept. of Env. Engg., HCST, Mathura. *Dept. of Civil Engg., AMU, Aligarh ABSTRACT In the present study an attempt has been made to develop a mathematical model for COD reduction in an UASB reactor. The reactor is operated at steady state under a given set of process conditions and the reactor contents are completely mixed. For successful operation of a UASB reactor the operational parameter specific substrate utilization rate kept in the range of 0.30-0.36 kg COD/kg V ss d. which is used in the model development to evaluate effluent substrate concentration. The values of COD removal efficiency predicted by the proposed model are in close agreement with the experimental results available in the literature. INTRODUCTION UASB reactor is getting wide applications during recent years for the treatment of a variety of industrial wastewaters. Over 300 full scale plants have been installed and commissioned in various parts of the world (Chui and Fang, 1994) there is a continuous need for a better understanding of their mechanisms. In this respect mathematical modeling offers a number of potential benefits. A review of literatures reveals scanty informations available regarding the modeling aspects of the process. COD removal is an important parameter to evaluate the performance of the process. Therefore, in the present study an attempt has been made to develop a mathematical model for COD reduction in an UASB reactor. The reactor usually consists of three parts, sludge bed, sludge blanket and the settler. On this basis a COD reduction model can be visualized and presented through Fig. 1. The wastewater enters the reactor at the bottom and flows upwards through a bed of relatively dense sludge and a blanket of sludge particles. The substrate comes in contact with the biomass and eventually gets converted into biogas and cells (Bhatia et al., 1985; Bolle et al., 1986a and Bolle et al., 1986b). The gas bubbles thus produced pass through the bed and cause excellent mixing so that adequate contact is made available between the biomass and the substrate. The fluid flow in the settler can be described as plug flow. The rising gas bubbles, produced in the sludge bed strengthen the bypassing effects but reduce the dead spaces. In the sludge bed, the bacteria are sufficiently supplied, so that enough gas is produced for mixing. In the lower parts of the bed the gas bubbles rise slowly because of the compactness of the bed. At higher places in the sludge blanket, the gas bubbles will rise faster, giving better mixing. The back mix flow from the blanket to the bed is generalized to replenish the volumes of the gas bubbles (plus their wake) that have already left the sludge bed. The flow rate is dependent on the flow rate of gas (Heertjes et al., 1978). For the model shown in Fig 1 the following notations may be considered. Suffix d represents 178 the ideally mixed region in the sludge bed and t represents the same in the blanket and pf represents the plug flow (settler) region. Q = Wastewater flow entering and leaving the reactor. (m3/d). Qbp = The part of flow (Q), bypassing the sludge bed. Qd = The part of flow (Q), entering the sludge bed. Qdt = The liquid flow from bed to blanket. (m3/d) Qtd = Backmix flow of liquid from blanket to bed. (m3/d) So = Influent substrate concentration. (kg/m3) Xo = Influent biomass concentration. ( kg/m3) V d = Volume of sludge bed. (m3) Sd = Substrate concentration in bed. (kg/m3) Xd = Biomass concentration in bed. (kg/m3) rsd = Rate of substrate conversion per unit vol. in bed. (kgCOD/m3.d) rst = Rate of substrate concentration per unit vol. in blanket. (kgCOD/m3.d) V t = Volume of sludge blanket. (m3) St = Substrate concentration in blanket.(kg/m3) Xt = Biomass concentration in blanket. (kg/m3) Xe = Effluent biomass concentration. (kg/m3) Se = Effluent substrate concentration. (kg/m3) D = Specific substrate utilization rate (kg COD/kg.VSS.d) Model Development: Substrate conversion and COD reduction are closely related to each other. To evaluate COD reduction efficiency of the UASB system, mass balance for substrate utilization is carried out around the bed and blanket (Fig.1), resulting in equations (1) and (2) (Bowker, 1983; Hulshoff Pol and Lettings, 1986). Sludge bed: V d. . dSd = Qd.So + Qtd.St – Qdt.Sd – qd.Xd.Vd…(1) dt Sludge blanket V t . dSt = Qbp.So + Qdt.Sd – Qtd.St – Q.St – qt.Xt.Vt…(2) dt Under pseudo – steady state conditions, dSd = 0 and dSt = 0 dt dt ISSN 2249-2631(online): 2249-2623(Print) - Rising Research Journal Publication Global J. of Engg. & Appl. Sciences, 2012: 2 (2) From Eqn.(1), 0 = Qd.So + Qtd.St – Qdt.Sd – qd.Xd.Vd. ……(3) Vd = Qd.So + Qtd.St – Qdt.Sd ………(4) qd.Xd Also from Eqn.(3), St = Qdt.Sd + qd.Xd.Vd – Qd.So……(5) Qtd Now from Eqn.(2) O = Qbp.So + Qdt.Sd – Qtd.St – Q.St – qt.Xt.Vt …(6) Therefore, Vt = Qbp.So + Qdt.Sd – Qtd.St. – Q.St ……(7) qt.Xt Also from Eqn.(6) we get , St = Qbp.So + Qdt.Sd – qt.Xt.Vt …… (8) Qtd + Q Equating Eqn. (5) and (8), Qdt.Sd + qd.Xd.Vd – Qd.So = Qbp.So + Qdt.Sd - qt.Xt.Vt Qtd Qtd + Q or, Qdt.Sd.Qtd + qd.Xd.Vd.Qtd - Qd.So.Qtd + Q.Qdt.Sd + Q.qd.Xd.Vd – Q.Qd.So = Qtd .Qbp.So + Qtd.Qdt.Sd – Qtd.qt.Xt.Vt or, Q.Qdt. Sd = Qtd.Qbp.So – Qtd.qt.Xt.Vt – qd.Xd.Vd.Qtd + Qd.So.Qtd – Q.qd.Xd.Vd + Q.Qd.So = So(Qtd.Qbp + Q.Qd + Qd.Qtd) – qd.Xd.Vd(Q+Qtd) – Qtd.qt.Xt.Vt Hence, Applying mass balance for flow around sludge bed and at point A (Fig.1), Qdt = Qd + Qtd…… (10) Q = Qd + Qbp……… (11) Or, Qd = Q – Qbp … (12) Using Equations (10) and (12) can be further modified as given below , So[Qtd.Qbp + Q(Q-Qbp) + (Q – Qbp) Qtd] Sd = - qd.Xd.Vd(Qd+Qbp+Qtd) – Qtd.qt.Xt.Vt Q.Qdt So (Qtd.Qbp + Q2 – Q.Qbp + Q.Qtd – Qbp.Qtd) Or, Sd = - qd.Xd.Vd(Qd+Qbp+Qtd) – Qtd.qt.Xt.Vt Q.Qdt = Q.So.(Q-Qbp+Qtd) – qd.Xd.Vd(Qbp+Qdt) – Qtd.qt.Xt.Vt Q.Qdt Or, Experimental observations reveal that most of the influent COD is removed in sludge bed which has a very high concentration of biomass as compared to 179 that in the sludge blanket. It implies that the most of the substrate is converted to methane, carbon dioxide and biological solids in the bed itself rather than blanket. Specific substrate utilization rate in the blanket may be considered to be negligible because no appreciable substrate conversion is accomplished in the blanket due to absence of a high concentration of viable biomass. Keeping these in view, following assumptions may be adopted. (i) Xt is very very small as compared to Xd. (ii) Substrate utilization rate (qt) in blanket is very very small as compared to that (qd) in sludge bed. (iii) Sd remains unchanged in blanket and settler i.e (Jewell and Switzenbaum, 1980; Jewell, 1981 and Jones and Hall, 1989). Sd ≈ St ≈ Se Applying these assumptions to equations (13), Sd = Q.So.Qdt – qd.Xd.Vd (Qbp + Qdt) ……… (14) Q.Qdt Assuming that the reactor is operating at steady state under a given set of process conditions and if the reactor contents are completely mixed Qbp may be negligible (i.e. Q = Qd). Thus equation (14) can be further simplified as, Sd = Q.So.Qdt – qd.Xd.V d.Qdt Q.Qdt Sd = Q.So – qd.Xd.Vd ……(15) Q) i.e. Se = Q.So – qd.Xd.vd........ (16) Q Equation (16) represents a simplified mathematical model to evaluate effluent substrate concentration from the reactor. For a reactor under consideration the values of various parameters such as flow rate (Q), influent substrate concentration (So), volume of sludge bed (Vd), biomass concentration in the sludge bed (Xd), hydraulic retention time (HRT), organic loading rate (OLR), sludge loading rate (SLR), etc. affecting the reactor performance are available (Table 1) (Mosey, 1983; Parkin and Owen, 1986). Volume of the reactor = 11.4 l Hydraulic retention time (HRT) = 4 hrs Biomass concentration in sludge bed (Xd) = 70.68 g/l Average values are shown in parentheses These parameters can be used to estimate the value of effluent substrate concentration. Manjunath (1987) suggested that the reactor should be operated at specific substrate utilization rate. (q) values above the limiting values (i.e.q>qlim). The limiting values of substrates utilization rat (qlim) were reported in the range of 0.059 – 0.174 kg COD/kg Vss.d. Keeping this in view various values ISSN 2249-2631(online): 2249-2623(Print) - Rising Research Journal Publication Global J. of Engg. & Appl. Sciences, 2012: 2 (2) of specific substrates utilization rate were selected in the range of 0.05 – 0.38 kg COD/kg Vss.d to evaluate the corresponding values of effluent substrate concentration/treatment efficiency (Table (1) and (2)) Equations (17) – (19) represent the regression analysis of the data presented, through Table (2). So = 1350 mg/1: y = - 3.937x + 1.351; (r = 0.999)…(17) So = 1800 mg/1: y = - 4.692x+1.799; (r = 0.999)……(18) So = 2100 mg/1: y = - 5.529x+ 2.101; (r = 0.999)……(19) Where y = Effluent substrate concentration (kg.m3) x= Specific substrate utilization rate (kg COD/kg Vss d) r= coefficient of correction For successful operation (i.e. COD removal efficiency ≈ 85-95%) of a UASB reactor it has been observed (Table (2) ; Equations (17 –19) that the operational parameters (Table 1), the values of specific substrates utilization rate should be kept in the range of 0.30 – 0.36 kg COD/kg Vss d. The values of specific substrate utilization rate may be adopted as 0.31 kg COD/kg V SS.d (for SO = 1350 mg/1), 0.35 kg COD.kg VSS.d (for SO = 1800 mg/1) and 0.35 kg COD/kg VSS.d (for SO = 2100 mg/1) and Used in the model developed to evaluate effluent substrate concentration (Table 2) (Pol et al., 1983; Van der Meer and Heertjes, 1983 and Lettings et al., 1984). Model Verification Model proposed in the preceding section (Eqn.16) is verified with the help of data available in (Table1). For Influent substrate concentration SO = 1350 mg/1> = 1.35 kg/m3 Organic loading rate (OLR) = 8.1 kg COD.m3d Qd = 0.31 kg COD/kg VSS d Xd = 70.68 kg/m3 Vd = 3.80 x 10-3 m3 (from Table 6.1) Q = V = 11.4 x 10 -3 m3 = 68.4 x 10 3 m3/d T 4/24 d From Eqn(16) Se = 68.4 x 10 -3 x 1.35 – 0.31 x 70.68 x 3.8 x 10-3 68.4 x 10 -3 = [ 92.84 – 83.26] x 10-3] 68.4x 10-3 = 0.13275 Kg/m3 = 132.75 mg/1 COD removal efficiency (%) = (1350 – 132.75) x 100 = 90.0% 1350 For influent substrate concentration SO = 1800 mg/1 = 1.8 kg/m3 Organic loading rate (OLR) = 10.80 kg COD/m3.d Qd = 0.35 kg COD/kg VSS.d Xd = 70.68 kg/m3 180 V d = 4.54 x 10-3 m3 (table 6.1) Q = 68.4 x 10-3 m3/d Se = 68.4x10-3 x 1.8 – 0.35x70.68x4.54x10-3 68.4 x 10-3 = 123.12 x 10-3 – 112.31 x 10-3 68.4 x 10-3 3 = 0.1580 kg/m = 158 mg/1 COD removal efficiency = [1800 – 158] x 100 = 91.0% 1800 And For influent substrate concentration (Table 3) So = 2100mg/1 = 2.1 kg/m3 Organic loading rate (OLR) = 12.60 kg COD/m3.d Qd = 0.35 kg COD/kg VSS.d Xd = 70.68 kg/m3 V d = 5.35 x 10-3 m3 (Table 6.1) Q = 68.4 x 10-3 m3/d Se = 68.4x10-3 x 2.1 – 0.35 x 70.68 x 5.35 x 10-3 68.4 x 10-3 = 143.64x10-3 – 132.35 x 10-3 68.4x 10-3 = 0.1651 Kg/m3 = 165.1 mg/1 COD removal efficiency = [2100-165.1] x 100 = 92.0% 2100 Volume of the reactor = 11.4 1 Hydraulic retention time (HRT) = 4 hrs Biomass concentration in Sludge bed (Xd) = 70.68 kg/m3 Average values are shown in parentheses CONCLUSIONS The present study can be concluded as follows: 1. A mathematical model for COD reduction in an UASB reactor is obtained as : Se = Q.So.Qdt – qd.Xd.Vd(Qbp+Qdt) Q.Qdt When the reactor contents are completely mixed then Qbp may be negligible (i.e. Q = Qd). Thus the model can be further simplified as Se = Q.So – qd.Xd.Vd 2. For successful operation (i.e. COD removal efficiency = 85 – 95%) of a UASB reactor it has been observed that for the operational parameters the values of specific substrate utilization rate should be kept in the range of 0.30-0.36 kg COD/kg V ss d. The value of specific substrate utilization rate may be adopted as 0.31 kg COD/kg V SS.d (for So = 1350 mg/1), 0.35 kg COD/kg VSS.d (for So = 1800 mg/1) and 0.35 kg COD/Kg VSS.d (for So = 2100 mg/1) and used in the model developed to evaluate effluent substrate concentration. 3. Table-3; compares the values of COD removal efficiency obtained by the proposed model with those obtained experimentally. The values of COD ISSN 2249-2631(online): 2249-2623(Print) - Rising Research Journal Publication Global J. of Engg. & Appl. Sciences, 2012: 2 (2) removal efficiency predicted by the proposed model are in close agreement with the experimental results available in the literature. REFERENCES Bhatia, D., Vieth, W.R and K. Vankatasubramaniant. 1985. Steady-state and transient Behavior in Microbial Methanification. Mathematical Modeling and Verification Biotechnology and Bioengineering. XXVII: 1199-1207. Bolle, W.L., Van Breugel, J., Van Eybergen, G.C., Kossen, N.W.F and R.J. Zoe Te Meyer. 1986a. Modelling the liquid Flow in Up-Flow Anaerobic Sludge Blanket reactors. Biotechnology and Bioengineering. XXVIII:1615-1620. Bolle, W.L., Van Breugel, J., Van Eybergen, G.C., Kossen, N.W.F and W. Van Gills. 1986b. An Integral Dynamic model for the UASB Reactor. Biotechnology and Bioengineering. XXVIII: 1621-1636. Bowker, R.P.G. (1983). New wastewater Treatment for Industrial Applications. Environment Progress, 2(4):235-242. Graef, S.P and J.F. Andrews. 1974. Stability and control of Anaerobic Digestion. J.Water Poll. Control. Fed, 46:666-683. Heertjes, P.M. and R.R. Van der Meer. `1978. Dynamics of liquid flow in an Up-flow reactor used for Anaerobic Treatment of wastewater. Biotechnology and Bioengineering. XX: 15771594. Hulshoff Pol, L and G. Lettings. 1986. New Technologies for Anaerobic Wastewater Treatment. Wat. Sci. Tech. 18(12): 41-53. Jewell, W.J and M.S. Switzenbaum. 1980. Anaerobic Attached Film Expand Bed Reactor 181 Treatment. J. Water Pollut. Control, 52(7): 1953-65. Jewell W.J. 1981. Development of the Attached Microbial Film Expanded Bed process for Aerobic and Anaerobic Waste Treatment. Biological Fluidized Bed Treatment of Water and Wastewater. Cooper, P.F. and Atkinson, B. (eds.) Ellis Horwood, Chichester, England. Jones, R.M. and Hall, E.R. (1989). Assessment of Dynamic Models for a High Rate Anaerobic Treatment Process. Env. Tech. Letters, 10(6):551-556. Lettings, G., Pol, L.W.H., Koster, I.W., Homba, W.M., Dezeeuw, W.J., Rinzema, A., Grin, P.C. Roersma, P.E and S.W. Homba. `1984. High Rate Anaerobic wastewater Treatment using the UASB reactor under a wide range of Temperature conditions. Biotechnology and Genetic Engineering. Reviews, II: 392-399. Mosey, F.E. 1983. New Development in the anaerobic treatment of industrial Wastes. Effluent and Waste Treatment Journ. XXIII(5): 85-93. Parkin, G.F and W.F. Owen. 1986. Fundamentals of Anaerobic Digestion of Waste Water Sludge. Journ. Of Env. Engg. Div., ASCE, 112(5):867920. Pol, L.H.W., Dezeeuw, W.J., C.T.M and G. Lettings. 1983. Granulation in UASB Reactors. Wat.Sci.Tech., XV: 291-304. Van der Meer, R.R and P.M. Heertjes. 1983. Mathematical Description of Anaerobic Treatment of wastewater in Up-flow reactors. Biotechnology and Bioengineering, XXV:25312556. ISSN 2249-2631(online): 2249-2623(Print) - Rising Research Journal Publication Global J. of Engg. & Appl. Sciences, 2012: 2 (2) Table-1: Analysis of effluent substrate Influent substrate concentration (mg/l) Organic loading rate (OLR) (kg COD/m3d) COD removal efficiency (%) Effluent substrate concentration (mg/l) Volume of sludge bed (Vd) (l) Sludge loading rate (SLR) kg COD/kg Vss d 1350 1800 2100 8.10 10.80 12.60 90-92 (91) 94-96 (95) 94-97 (95) 135-108 (121.5) 108-72 (90) 126-63 (94.5) 3.80 4.54 5.35 0.339 0.391 0.372 Table-2: Evaluation of the Specific Substrate Utilization Rate Influent Substrate Concentration So = 1350 mg/l Influent Substrate Concentration So = 1800 mg/l Influent Substrate Concentration So = 2100 mg/l Specific Substrate Utilization rate qd (kgCOD/kg Vss d) Effluent substrate concentration Se (kg/m3) COD removal efficiency (%) Specific Substrate Utilization rate qd (kgCOD/kg Vss d) Effluent Substrate Concentration Se (kg/m3) COD removal efficiency (%) Specific Substrate Utilization rate qd (kgCOD/kg Vss d) Effluent Substrate concentration Se (kg/m3 COD removal efficiency (%) 0.05 0.10 0.20 0.30 0.31 0.32 0.33 0.34 1.150 0.960 0.560 0.170 0.132 0.092 0.050 0.014 15 29 56 87 90 93 96 99 0.05 0.10 0.20 0.30 0.31 0.32 0.33 0.34 0.35 1.565 1.330 0.860 0.390 0.346 0.299 0.250 0.200 0.160 13 26 52 78 81 83 86 89 91 0.05 0.10 0.20 0.30 0.31 0.32 0.33 0.34 0.35 1.824 1.548 0.996 0.444 0.390 0.330 0.280 0.200 0.170 13 26 52 79 81 84 87 90 92 0.36 0.37 0.38 0.110 0.060 0.020 94 97 99.9 0.36 0.37 0.38 0.110 0.060 0.002 95 97 100 Table-3: Comparison between COD removal efficiency (Verification) Influent Substrate Concentration So (mg/1) 1350 1800 2100 Organic Loading rate (OLR) (Kg COD/M3d) 8.10 10.80 12.60 COD removal Efficiency (Experimental) (%) 90-92(91) 94-96 (95) 94-97(95) Fig.1 COD reduction model COD removal Efficiency Predicted by the Model (%) 90 91 92 ************ 182 ISSN 2249-2631(online): 2249-2623(Print) - Rising Research Journal Publication