CFD Analysis of Mixing in Large Aerated Lagoons

Engineering Applications of Computational Fluid Mechanics ISSN: 1994-2060 (Print) 1997-003X (Online) Journal homepage: http://www.tandfonline.com/loi/tcfm20 CFD Analysis of Mixing in Large Aerated Lagoons Binxin Wu To cite this article: Binxin Wu (2010) CFD Analysis of Mixing in Large Aerated Lagoons, Engineering Applications of Computational Fluid Mechanics, 4:1, 127-138, DOI: 10.1080/19942060.2010.11015304 To link to this article: https://doi.org/10.1080/19942060.2010.11015304 Copyright 2010 Taylor and Francis Group LLC Published online: 19 Nov 2014. Submit your article to this journal Article views: 381 View related articles Citing articles: 3 View citing articles Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalinformation?journalcode=tcfm20

Engineering Applications of Computational Fluid Mechanics Vol. 4, No. 1, pp. 127 138 (2010) CFD ANALYSIS OF MIXING IN LARGE AERATED LAGOONS Binxin Wu Philadelphia Mixing Solutions Ltd., Palmyra, PA 17078, USA E-Mail: bxwulk@hotmail.com ABSTRACT: A general CFD model is developed to characterize the flow fields in large aerated lagoons. The power number and flow number of a propeller used in aerated lagoons are validated against lab experiments. Numerical tests in a small lagoon with 8 propellers indicate that the pumping action of propeller can be represented by momentum source effectively. CFD simulations of a large lagoon with the existing 33 propellers and with 16 propellers for a proposed redesign are carried out by substituting one momentum source for each propeller. A local RTD model is developed to predict resident time in both designs based on the converged flow fields. Furthermore, a biological oxygen demand (BOD) model is incorporated to simulate BOD removal in the lagoon. Simulated results show that the proposed design should be more efficient than the existing design in terms of BOD removal per unit power consumption. Keywords: aerated lagoon, mixing, computational fluid dynamics, retention time, BOD removal 1. INTRODUCTION Biological treatment of wastewater by lagoons has been widely used throughout the world for more than 40 years, relying on unsophisticated technology and minimal maintenance. Wastewater treatment in lagoons involves a complex combination of physical, chemical, and biological processes that are affected by many parameters such as lagoon types and configurations, system design, and environmental conditions. In terms of the predominant biological environment that exists, lagoons can be classified as facultative, aerated, aerobic, and anaerobic lagoons (NRC CNRC, Canada, 2004). Usually, aerated lagoons serve as an economical means for secondary treatment of wastewater, in which the propellers through mechanical mixing bring the wastewater into contact with microorganisms that decompose the pollutants. Incomplete mixing may result in dead zones in a lagoon where the actual liquid retention time is more than the designed retention time while in other areas the residence time is less than desired. Lagoon hydraulic performance can be experimentally and theoretically investigated. One conventional method is to inject a dye tracer into the lagoon inlet and then measure the tracer concentration at the outlet to obtain the RTD. For large lagoons, however, tracer studies are time-consuming, labor intensive, and expensive (Pougatch et al., 2007). Murphy and Wilson (1974) performed tracer experiments in prototype lagoons to study the mixing behavior of aerated lagoons. It was concluded that a dimensionless dispersion number used to characterize the degree of macro-mixing depends more on the lagoon geometry and flow rate than on aerator horsepower input, and that the mixing patterns in aerated lagoons may differ significantly from either ideal complete mixing or ideal plug flow. Dorego and Leduc (1996) did tracer experiments to obtain the mean retention time in a lagoon system including three facultative aerated lagoons and a maturation pond operating in series, and concluded that the hydraulic flow pattern in the lagoons is intermediate between plug flow and complete mixing. Nameche and Vasel (1998) conducted dye tracer experiments on aerated lagoons and waste stabilization ponds to compare with hydraulic models of dispersion, and developed predictive formulas to estimate the macro-mixing conditions prevailing in lagoons and ponds. The theoretical studies on lagoons and ponds were initially from one dimensional (1-D) or algebraic models. Ferrara and Harleman (1981) presented a hydraulic model used to identify the appropriate hydraulic design of waste stabilization ponds. One dimensional conservation of mass transport equations for both the active and return flow zones were solved, and the model simulations were used to indicate the efficiencies of various pond geometries and inlet and outlet configurations. Arceivala (1983) proposed empirical correlations to obtain the dispersion coefficient for lagoons and ponds with and without baffles by determining the ratios of length to width. Polprasert and Bhattarai (1985) Received: 1 Aug. 2009; Revised: 23 Sep. 2009; Accepted: 27 Oct. 2009 127

developed dispersion prediction formula for ponds relating the pond s dispersion factor to the pond s shape, hydraulic detention time, and the kinematic viscosity of pond water. Agunwamba (1992) compared the differences in the performance of model and prototype waste stabilization ponds resulting from pond geometry and flow as well as environmental conditions, and then put forward empirical models to calculate the dispersion number. With the development of CFD technique, some researchers were involved in developing two dimensional (2-D) and three dimensional (3-D) flow models in lagoons and ponds. Wood et al. (1995) performed CFD modeling on wastewater ponds assumed to be a 2-D rectangle (100 m 50 m). The aerator was simplified as a regional acceleration positioned at the center of the pond with a 45 upward angle. Laminar flow under steady state was simulated to predict fluid velocities for unbaffled, inlet baffled, outlet baffled, and aerated pond. Peterson, Harris and Wadhwa (2000) evaluated the effect of aerators on the sediment quality in an aquaculture pond. CFD simulations of the combined two paddlewheels and four propeller-aspirators in the pond were conducted to present 2-D (220 m 100 m) plots of shear stress contours near the impingement of the jet and velocity contours near the water surface and pond bottom. The results were obtained through 518 iterative calculations using 907 central processing unit (CPU) hours (37.79 days) on a computer of 64 bit precision and 743 MB random access memory. Later, Peterson, Wadhwa and Harris (2001) applied a CFD model to study the arrangement of aerators in a rectangular pond and proposed that conventional aerators should be arranged diagonally or in parallel to minimize the dead zone. Salter et al. (2000) carried out 3-D steady state simulations of a facultative lagoon (430 m 180 m 2.5 m) with and without baffles to predict the RTD and BOD removal at the lagoon outlet. Baleo and Le Cloirec (2000) presented a modeling approach to simulate the spatial distribution of mean residence time in a turbulent axisymmetric flow containing expansions and contractions, and compared the simulations with experimental measurements obtained with a pulse injection method using butane as the trace gas. Subsequently, Baleo, Humeau and Le Cloirec (2001) applied their modeling approach to study RTD in a pilot lagoon without aerators. De Kretser, Matthews and Williams (2002, 2003) examined the flow patterns generated by individual aerators and mixers, and the combinations of three aerators and three mixers in a rectangular block of 70 m 2 with a depth of 4.3 m instead of the whole lagoon. The results showed that aerators produce very good surface mixing but are relatively inefficient at depth, whereas the mixers promote good mixing at depth but are less effective on the surface. Pougatch et al. (2007) developed a 3-D computational flow model for large lagoons including high-speed floating mechanical surface aerators. In this model, the lagoon was divided into two domains with 315.6 m 315 m 2.85 m in each zone. Each aerator was simulated separately and then was applied to the boundary condition in the lagoon model. Stropky et al. (2007) developed a calculation method based on random-walk Lagrangian particle tracking to predict RTD in large mechanically aerated lagoons. To the author s knowledge, the latest CFD model was reported by Pougatch et al. (2007), in which about 1.6 million computational cells were used to generate grids and 10 days of CPU time on a PC with Pentium 3 at 1.2 GHz was taken to obtain one solution. 2. OBJECTIVES The overall goal of this study was to develop a general CFD model requiring less CPU time for large lagoon simulation. The specific objectives of this study were to: 1. Develop a theoretical model integrating water flow, local RTD, and BOD removal in aerated lagoons; 2. Validate the power number and flow number of a propeller used in aerated lagoons against experiments; 3. Quantify the flow patterns in a small lagoon with propellers and momentum sources, and test the feasibility of representing the pumping action of the propellers by momentum sources to save mesh cells and CPU time; and 4. Predict flow fields, residence time, and BOD removal in a large lagoon for the existing design with 33 propellers and the proposed redesign with 16 propellers. 3. MODEL DEVELOPMENT The theoretical model was developed based on the following assumptions: Water flow in the lagoon is 3-D and steady state. 128

Water is an incompressible and isothermal Newtonian fluid. The tracers have the same density as the surrounding fluid and no diffusion occurs. The model is limited to the flow model without considering the heat flow, as the water temperature is constant at 20 C. BOD removal follows first-order kinetics, in which the decay reaction rate is assumed to be 0.23 d -1 at 20 C. The model is single phase without accounting for multiphase interactions. Pumping action of each propeller is represented by one momentum source in the large lagoon simulation. The effect of wind speed on water flow is negligible. 3.1 Flow model In this study, the rotation of the propeller was solved in multiple reference frame (MRF). The general governing equation can be expressed as (Patankar, 1980): ( ρ ) t div ( ρ u ) div ( ) S (1) where t is time, is density of wastewater, Γ is transport coefficient, represents velocity components, turbulent kinetic energy or its dissipation rate, S is source term dependent on, and u is the absolute velocity which can be calculated by: u ( r ) v (2) where is the angular velocity vector, r is the position vector in rotating frame, and v is relative velocity vector in rotating frame. 3.2 Turbulence model Since the Reynolds stress term in the Reynoldsaveraged Navier-Stokes equation for turbulent flow is unknown, many turbulence models have been developed to complete the flow equations. The commonly used turbulence models include standard k-ε, RNG k-ε, Realizable k-ε, and Reynolds stress models, among which the realizable k-ε model was tested to work well for recirculation and rotation flows (Shah, 1979), and thus was used in this study. 3.3 Power number The power number in a stirred vessel is defined as (Paul, Atiemo-Obeng and Kresta, 2004): N P P (3) N 3 D 5 where N is the rotational speed, D is the diameter and P is the power delivered to the fluid through the blades of propeller which can be calculated as: P 2 N F torque (4) 3.4 Flow number The flow number in a stirred vessel is defined as (Paul, Atiemo-Obeng and Kresta, 2004): N Q Q 3 N D (5) where Q is flow rate generated by the propeller. 3.5 Local RTD model RTD model is represented by a scalar transport equation. For an arbitrary scalar k, the general transport equation can be expressed as (ANSYS- Fluent Inc., 2007): ( ρ k ) t x i ( ρ u i k k x k i ) (6) According to the assumption that there is no diffusion between the tracer and the background liquid, if the source term S k is specified as the density of the background liquid ( ), then the resident time of the tracer can be obtained by solving the convection term only under steady state as: x i ( u ) 1 (7) i k where k represents the physical resident time of the tracer. 3.6 BOD removal model The most widely used parameter of organic pollution applied to wastewater treatment is BOD removal. Based on the first-order kinetics, BOD removal in lagoons can be expressed as (Metcalf and Eddy, 2003): e k t C 1 1 / e (8) S k 129

where k is constant rate and t is hydraulic retention time. 4. CFD STRATEGY CFD modeling strategies include grid generation and then numerically solving the governing equations. In meshing a large lagoon with multiple propellers, controlling mesh quantity and ensuring mesh quality are crucial to a successful simulation. Lagoon meshing was implemented by Gambit software (ANSYS-Fluent Inc., 2007). The main meshing procedures included: (1) constructing lagoon geometry with inlets, outlet, and propellers, (2) decomposing the whole domain into a few sub-domains, (3) generating tetrahedral mesh for MRF zones containing propellers and hexahedral mesh for all other regions, and (4) using size functions to control mesh growth. Numerical computing was performed by Fluent software (ANSYS-Fluent Inc., 2007). The key setups under Fluent were: (1) using Green-Gauss node based solver, (2) specifying the influent as velocity-inlet, the effluent as pressure-outlet, the water top surface as symmetry, and walls as no-slip velocity, and (3) starting iterations from the first order upwind and then switching to the second order upwind scheme. Upon obtaining converged flow fields, user defined scalar (UDS) can be active to solve Eq. (7) to predict residence time, and then the custom field function can be used to calculate BOD removal by Eq. (8). camera, a high power laser, and an optical arrangement to convert the laser output light to a light sheet. A fiber optic cable connects the laser to the cylindrical lens setup. The laser acts as a photographic flash for the digital camera, and the particles in the fluid scatter the light that is detected by the camera. Upon obtaining the power input and flow rate of the propeller, the power number and flow number can be predicted by Eqs. (3) and (5). Comparison of the measured and simulated values of power number and flow number indicates that CFD is capable of predicting propeller-driving flow fields (Table 1). (a) Power test 5. RESULTS AND DISCUSSIONS 5.1 Validation of power number and flow number A 20HP Raptor propeller used in this study was tested at Philadelphia Mixing Solutions to obtain its power number and flow number. As illustrated in Fig. 1(a), the experiment was conducted in a rectangular vessel (18 m 15 m 4.5 m) with one propeller aimed 45 down. Power consumption at 1200 rpm of rotational speed (N) was measured and continuously recorded during all test runs. Wattmeter power was multiplied by motor efficiency to determine motor output horsepower, also designated as brake HP. An Esterline Angus Recording Wattmeter was used to measure electrical power to the motor. Flow rate was obtained by measuring the velocity fields through particle image velocimetry (PIV) as shown in Fig. 1(b). PIV apparatus consists of a digital Fig. 1 (b) Flow test Propeller lab test. Table 1 Measured and simulated power number and flow number of a propeller. Measured Simulated Power number 0.41 0.4 Flow number 0.66 0.6 130

(a) Lagoon configuration (b) Arrangement of propellers Fig. 2 A small lagoon with 8 propellers. 5.2 Comparisons of propellers and momentum sources in a small lagoon In order to investigate the feasibility of substituting momentum source for each propeller, a small lagoon (135 m 42 m 4.3 m) with 8 propellers was simulated (Fig. 2). Arrangements of 8 propellers were made in three rows, and all propellers were set aiming 55 down. Propellers 1, 2 and 3 in the first row were pumping towards the outlet, opposite to the pumping direction of propellers 4, 5, and 6 in the third row. To enhance mixing in the central region, propellers 7 and 8 were placed in the second row pumping towards the inlet. All propellers were operated at N=1200 rpm in this case and large lagoon case discussed later. Taking propeller 3 as an example, average velocity magnitude ( v ) that is normal to the exit plane for each MRF zone can be predicted as 4.6 m/s for N=1200 rpm illustrated in Fig. 3(a), and then the momentum source can be specified by fixing two velocity components illustrated in Fig. 3(b). Following this procedure, the velocity components for all other momentum sources can be specified. All the momentum sources have the same velocity magnitudes but the directions are different for different locations. Fig. 4 shows qualitative and quantitative comparisons of velocity contours at some specified planes for propellers and momentum sources. In all displays of velocity contours the velocity magnitudes in the red area are greater than (or equal to) the maximum value specified in the velocity contour bar. From Fig. 4 it can be seen that the overall flow patterns are similar although velocity magnitudes in some areas are different. These differences could be attributed to 131

(a) MRF zone containing propeller (b) Momentum source zone Fig. 3 Velocity vectors in MRF and momentum source zones. Propeller at y = 8.5 m ( v = 0.32 m/s) Momentum source at y = 8.5 m ( v = 0.36 m/s) Propeller at y = 18.3 m ( v = 0.15 m/s) Momentum source at y = 18.3 m ( v = 0.15 m/s) Propeller at y = 33.5 m ( v = 0.32 m/s) Momentum source at y = 33.5 m ( v = 0.35 m/s) (a) z-x planes respectively at y = 8.5, 18.3, and 33.5 m Propeller ( v = 0.35 m/s) Momentum source ( v = 0.38 m/s) (b) x-y plane at z = 0.1 m Fig. 4 Velocity contours at the specified planes in a small lagoon. 132

Table 2 Mesh cells and CPU time for a small lagoon. Propeller Momentum source Mesh cells for each propeller / momentum source 55,237 75 Total mesh cells 1,214,865 601,680 CPU time 8 hours 4 hours (a) Configurations and dimensions (b) Existing design (uniform spacing) (c) Proposed design Fig. 5 A large lagoon respectively with 33 and 16 propellers. the fact that only two velocity components are fixed in each momentum source while propellerdriving has three velocity components. Generally, a momentum source acting as a propeller is acceptable. In this study, lagoon meshing and CFD simulation were performed in an HP workstation xw6200 with 6 dual processors (2 GB memory of each processor), and every solution 133

was checked for grid independence. CFD simulation using momentum sources, evaluated in terms of CPU time and number of mesh cells, shows higher efficiency than that using propellers (Table 2). 5.3 Large lagoon case study As shown in Fig. 5(a), a large lagoon (610 m 274 m 2.4 m) with four inlets and one outlet was used in this study. Fig. 5(b) shows the existing placement of 33 propellers pumping down, and Fig. 5(c) shows the proposed design using 16 propellers aimed 45 down. The propellers used in the existing and proposed designs, represented by the momentum sources, are depicted in Fig. 6. Total mesh cells and CPU time for both designs are shown in Table 3. (a) Existing design Table 3 Total mesh cells and CPU time for a large lagoon. Existing design Proposed design Total mesh cells 1,453,444 1,254,708 CPU time 12 hours 10 hours (b) Proposed design Fig. 6 Propellers represented by momentum sources. (a) z = 0.1 m ( v = 0.06 m/s) (a) z = 0.1 m ( v = 0.15 m/s) (b) z = 1.4 m ( v = 0.04 m/s) (b) z = 1.8 m (v = 0.15 m/s) Fig. 8 Velocity contours at x-y planes in a large lagoon with 33 propellers. Fig. 7 Velocity contours at x-y planes in a large lagoon with 16 propellers. 134

Similar to the small lagoon, all propellers in the large lagoon for the existing design were placed in three rows with uniform spacing. Rows 1, 2, and 3 consisted of propellers 1 to 11, 12 to 22, and 23 to 33, respectively. It can be observed from Fig. 7 that the mixing surrounding each propeller is very intensive but weak along the side walls. The overall flow patterns from top view (xy planes) at different depths are similar. However, the average velocity is 0.06 m/s at z=0.1 m which is greater than 0.04 m/s at z=1.4 m, indicating good mixing at depth. In addition, a few flow channels can be identified from the inlets to some propellers in Row 1 and from some propellers in Row 3 to the outlet. These flow channels might result in hydraulic short-circuiting. In contrast to the existing design, the proposed design reduced the number of propellers and relocated each propeller 45 down. Fig. 8 shows that the global flow direction is clockwise corresponding to the pumping directions of propellers 1 14. Placement of propellers 15 and 16 opposite to the inlet flow direction not only improved mixing in the central regions, but also prevented the short-circuiting from the inlets to the outlet. Furthermore, this arrangement produced two local flow circulations with a large one on the left side and a small one on the right side. The average velocity is 0.15 m/s at both z=0.1 m and z=1.8 m, illustrating uniform mixing throughout the depth in the proposed design. To trace the flow paths, some colored particles were assumed to be injected at the inlets. For steady state flow, each pathline refers to the trace of a single particle in space. Fig. 9 shows the flow paths from the top view in both designs. In the existing design the density of pathlines in the vicinity of each inlet is higher than that in other regions, and few particles flow to the right side of the lagoon. In the proposed design, however, pathlines are relatively uniform over the whole domain and two significant flow circulations and several small ones can be observed. Simulations of local residence times in the lagoon were conducted based on the converged flow fields. Fig. 10(a) shows the contours of residence time at z=0.1 m in the existing design, in which the residence time in the right side is greater than that in the left side. This phenomenon is because of the hydraulic short-circuiting discussed before. The objective of lagoon hydraulic design is to obtain the longest uniform retention time of water. To increase the uniformity of the residence time, one possible way is to maximize the water flow path. In the proposed design, the residence time in the left side of lagoon depicted in Fig. 10(b) has been improved with the increase of flow paths in this region. (a) Existing design (b) Proposed design Fig. 9 Pathlines colored by particles releasing from influent. 135

(a) Existing design (a) Existing design (b) Proposed design (b) Proposed design Fig. 10 Contours of residence time in x-y plane at z = 0.1 m. Fig. 11 Contours of BOD removal in x-y plane at z = 0.1 m Upon obtaining the residence time, the goal of predicting BOD removal (C e ) in lagoons can be achieved. C e was calculated using the custom field function in Fluent software without iterations. From Eq. (8) it can be seen that C e is the function of constant rate (k) and residence time (t). If k is kept constant, then C e increases nonlinearly with the increase of t. As a result, BOD removal shown in Fig. 11 corresponds to the residence time shown in Fig. 10. Table 4 gives quantitative comparisons of BOD removal inside the lagoon and its BOD reduction in the effluent for both designs. Even though BOD removal for the existing design is only a little better than that for the proposed design, the total power consumption for the existing design is much higher than that for the proposed design (Table 5). In the present research stage, the lagoon temperature was assumed to be constant in calculating BOD removal. Actually, the lagoon temperature varies with surrounding soil, weather conditions, and solar radiation, etc. Also the prevailing wind speed should affect the accuracy of the local RTD. To complete this study, there is a need for further research that would integrate the lagoon hydraulic model with heat transfer and wind flow models. Work at Philadelphia Mixing Solutions is continuing towards this goal. Controlling the mesh quantity to reduce CFD computing time is very important for large lagoon simulations. CFD along with local RTD methodology developed in this study provides an effective simulation tool to optimize aerated lagoons through relocating propellers. Prediction of residence time would assist environmental engineers in analyzing the hydraulic performance by evaluating a lagoon s mixing characteristics instead of using the conventional RTD method that treats a lagoon as a black box. In essence this program opens the black box for a more efficient design. Table 4 BOD removal for two designs. Table 5 Power consumption for two designs. Existing design Proposed design Existing design Proposed design Average BOD removal in the whole lagoon BOD reduction in the effluent 31.3% 30.4% 29% 29% Total power consumption Power consumption per unit volume 495 kw 240 kw 1.23 W/m 3 0.6 W/m 3 136

6. CONCLUSIONS From the CFD analysis of mixing in aerated lagoons, the following conclusions can be drawn: 1. The pumping action of the propeller can be represented by a momentum source effectively. 2. A hydrodynamic model that integrates a simple biological model has been successfully developed to evaluate lagoon mixing performance. 3. For a large aerated lagoon, mixing performance in the proposed design with less propellers and relocations has been improved in terms of BOD removal per unit power consumption. NOMENCLATURE BOD biological oxygen demand C e BOD removal, 100% CFD computational fluid dynamics d diameter, m D diameter, m F torque torque, N.m k turbulent kinetic energy, m 2 /s 2 k constant rate, 1/day L length, m MRF multiple reference frame N rotational speed, rpm or rps N P power number, dimensionless N Q flow number, dimensionless P power, W PIV particle image velocimetry Q flow rate, m 3 /s r radius, m Re Reynolds number, dimensionless S source term t time, s or day u absolute velocity, m/s UDS user defined scalar v relative velocity, m/s v average velocity, m/s V flow rate, m 3 /s x, y, z Cartesian coordinate Greek symbols ρ density, kg/m 3 general parameter Γ transport coefficient ω angular velocity, rad/s ε Subscripts dissipation rate of turbulent kinetic energy, m 2 /s 3 i, k tensor form REFERENCES 1. Agunwamba JC (1992). Field pond performance and design evaluation using physical models. Water Research 26(10):1403 1407. 2. ANSYS-Fluent Inc. (2007). Fluent 6.3. Lebanon, N.H. 3. Arceivala SJ (1983). Hydraulic modeling for waste stabilization ponds. Journal of the Environmental Engineering Division 109(1):265 268. 4. Baleo JN, Le Cloirec P (2000). Validating a prediction method of mean residence time spatial distributions. AIChE J 46(4):675 683. 5. Baleo JN, Humeau P, Le Cloirec P (2001). Numerical and experimental hydrodynamic studies of a lagoon pilot. Water Research 35(9):2268 2276. 6. De Kretser D, Matthews B, Williams M (2002). Maryvale mill uses CFD to investigate natural and mechanically induced flows in an aerated lagoon. Paper Technology 43(9):35 38. 7. De Kretser D, Matthews B, Williams M (2003). Use of computational fluid dynamics to investigate natural and mechanically induced flows in an aerated lagoon. Tappi Journal 2(4):12 14. 8. Dorego NC, Leduc R (1996). Characterization of hydraulic flow patterns in facultative aerated lagoons. Water Science Technology 34(11):99 106. 9. Ferrara RA, Harleman DRF (1981). Hydraulic modeling for waste stabilization ponds. Journal of the Environmental Engineering Division 107(4):817 830. 10. Metcalf and Eddy (2003). Wastewater Engineering (fourth edition). McGraw-Hill, New York. 11. Murphy KL, Wilson AW (1974). Characterization of mixing in aerated lagoons. Journal of the Environmental Engineering Division 100(5):1105 1117. 12. Nameche Th, Vasel JL (1998). Hydrodynamic studies and modelization for aerated lagoons and waste stabilization ponds. Water Research 32(10):3039 3045. 137

13. NRC CNRC, Canada (2004). Optimization of lagoon operation. Available at: www.sustainablecommunities.fcm.ca/files/inf raguide/. 14. Patankar SV (1980). Numerical Heat Transfer and Fluid Flow. Hemisphere/ McGraw-Hill, New York. 15. Paul EL, Atiemo-Obeng VA, Kresta SM (2004). Handbook of Industrial Mixing. John Wiley & Sons, Inc, New Jersey. 16. Peterson EL, Harris JA, Wadhwa LC (2000). CFD modeling pond dynamic process. Aquacultural Engineering 23:61 93. 17. Peterson EL, Wadhwa LC, Harris JA (2001). Arrangement of aerators in an intensive shrimp grow out pond having a rectangular shape. Aquacultural Engineering 25:51 65. 18. Polprasert C, Bhattarai KK (1985). Dispersion model for waste stabilization ponds. Journal of the Environmental Engineering Division 111(1):45 49. 19. Pougatch K, Salcudean M, Gartshore I, Pagoria P (2007). Computational modelling of large aerated lagoon hydraulics. Water Research 41:2109 2116. 20. Salter HE, Ta CT, Ouki SK, Williams SC (2000). Three-dimensional computational fluid dynamic modelling of a facultative lagoon. Water Science and Technology 42(10 11):335 342. 21. Shah NG (1979). A new method of computation of radiant heat transfer in combustion chambers. PhD Dissertation. Imperial College of Science and Technology, London, England. 22. Stropky D, Pougatch K, Nowak P, Salcudean M, Pagoria P, Gartshore I, Yuan JW (2007). RTD (residence time distribution) predictions in large mechanically aerated lagoons. Water Science and Technology 55(11):29 36. 23. Wood MG, Greenfield PF, Howes MR, Johns MR, Keller J (1995). Computational fluid dynamics modeling of wastewater ponds to improve design. Water Science Technology 31(12):111 118. 138