UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Application of the dynamic one-dimensional model to pulp and paper mill wastewater clarifiers Lo, Ing-Wei 2004

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2004-0260.pdf [ 11.15MB ]
Metadata
JSON: 831-1.0063472.json
JSON-LD: 831-1.0063472-ld.json
RDF/XML (Pretty): 831-1.0063472-rdf.xml
RDF/JSON: 831-1.0063472-rdf.json
Turtle: 831-1.0063472-turtle.txt
N-Triples: 831-1.0063472-rdf-ntriples.txt
Original Record: 831-1.0063472-source.json
Full Text
831-1.0063472-fulltext.txt
Citation
831-1.0063472.ris

Full Text

Application of the Dynamic One-dimensional Model to Pulp and Paper Mill Wastewater Clarifiers by ING-WEI LO B.Sc, ChungShing National University, Taiwan, 1989 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Civil Engineering; University of British Columbia; Pollution Control and Waste Management We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA Apr 2004 © Ing-Wei Lo, 2004 Library Authorization In presenting this thesis in partial fulfillment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that cdpying or publication of this thesis for financial gain shall not be allowed without my written permission. Name of Authof^frwease print) Date (dd/mm/yyyy) Title of Thesis: fà^ciïfyn, 4Â?  M>ll [/{Jajfcvdè^ CÂir&GKs -Degree: M^^k^u *j AffÀept Sc^c^ear. Q^fijk Department of Cl £/)/ The University of British Columbia Vancouver, BC Canada This thesis reports the results of research to calibrate the parameters of the Takacs model for pulp and paper sludge and to identify how the Takacs model works on pulp and paper wastewater clarifiers, as compared with municipal wastewater clarifiers. This research work involved zone settling tests for calibration of the model parameters, pilot scale tests and full scale clarifier simulations. The estimated parameter values for the pulp and paper (P&P) and municipal sludge are summarized in Table 1. Both the pulp and paper primary and secondary sludge can be simulated by the Takacs model under normal operating conditions. Table 1 Summary of the Parameters of the Takacs Model for P & P and Municipal Sludge Item r„(xl0-4) r p(xl0' 3) Vo Vo' y Pulp and paper sludge Primary sludge 2.728 4.864 133 85 41 Secondary sludge 3.073 4.696 226 141 7 Municipal sludge Primary sludge 1.617 2.286 107 65 22 Secondary sludge 8.148 7.888 210 97 4 Four clarifiers at the pulp and paper and municipal wastewater treatment plants were simulated. The results showed that all of the circular clarifiers could be simulated by this model, but the rectangular settling tank of this study required 2 or 3-dimensional modeling for a proper simulation. The waste sludge from both the primary and secondary clarifiers of the pulp and paper mill was about 1% higher in concentration than in the simulated result due to the fact that the sludge hoppers provide a long detention time for sludge thickening. Application of the calibrated Takacs model, however, can be of assistance in the design of new clarifiers. It can also be beneficial in operation assistance and risk management assessment for wastewater treatment plants. However, since the sludge at different treatment plants is specific and unique to each plant, the sludge in different situations may have different characteristics and therefore should be calibrated separately. Sections Page Abstract ii Table of Contents iv List of Tables vii List of Figures x Acknowledgement xiv 1 Introduction 1 2 Literature Review 3 2.1 Settling Models Devolvement 3 2.1.1 Discrete sedimentation 3 2-1.2 Solids Flux Model 3 2.1.3 One-Dimensional Models 8 2.1.4 Two-and Tree Dimensional Model 10 2.2 Model Selection 11 2.3 The Takacs Model 12 2.4 Sensitivities of Parameters of the Takacs Model 15 2.5 Conclusions of application of the Takacs model from Literatures 18 2.6 Characteristics of Sludges 20 2.7 Summary of Literature Review 21 2.7.1 Models Devolvement and Selection 21 2.7.2 Characteristics of pulp and paper sludge 22 3 Research Objectives and Approach 23 4 Materials and Methods 25 4.1 Introduction: The Effluent Treatment Plant of NorskeCanada Inc. 25 Port Alberni Pulp and Paper Mill 4.1.1 Manufacture Process 25 4.1.2 Effluent Treatment Process 27 4.1.3 Primary Clarifier 30 4.1.4 Secondary Clarifiers 32 Sections Page 4.2 Introduction: The Comox Valley Municipal Wastewater Treatment 35 Plant 4.2.1 Collection System 35 4.2.2 Wastewater Treatment Process 35 4.2.3 Primary Clarifiers 38 4.2.4 Secondary Clarifiers 40 4.3 Experimental Procedures 41 4.3.1 Batch Settling Tests and Calibration of Parameters 42 4.3.2 Steady State Tests (Pilot Scale) 44 4.3.3 Field Sampling Tests (Full Scale) 46 4.4 Analysis Methods 46 3.4.1 Suspended Solids 46 3.4.2 Consistencies 46 4.5 Mathematics Procedures 47 4.6 Quality Assurance and Quality Control 47 5 Results and Discussion 49 5.1 Batch Settling Test and Parameters Calibration 49 5.1.1 Primary Sludge of Pulp and Paper Wastewater 49 5.1.2 Secondary Sludge of Pulp and Paper Wastewater 59 5.1.3 Primary Sludge of Municipal Wastewater 68 5.1.4 Secondary Sludge of Municipal Wastewater 77 5.1.5 Discussion 86 5.2 Steady State Tests (Pilot Scale) 92 5.2.1 Primary Sludge of Pulp and Paper Wastewater 92 5.2.2 Secondary Sludge of Pulp and Paper Wastewater 94 5.2.3 Primary Sludge of Municipal Wastewater 95 5.2.4 Secondary Sludge of Municipal Wastewater 99 5.2.5 Discussion 101 5.3 Field Sampling Tests (Full Scale) 103 5.3.1 Primary Clarifier of Pulp and Paper Wastewater 103 5.3.2 Secondary Clarifier of Pulp and Paper Wastewater 105 Sections Page 5.3.3 Primary Clarifier of Municipal Wastewater 106 5.3.4 Secondary Clarifier of Municipal Wastewater 110 5.3.5 Discussion 113 5.5 Precision of Solids Analysis 123 6 Conclusions and Recommendations 124 6.1 Conclusions 124 6.2 Recommendations for Further Work 126 7 Engineering Significance 128 References 129 Appendices A MatLab Program for Dynamic simulation of the Takacs Model Appendices B Data of Settling Test Tables Page Table 1 Summary of the Parameters of the Takacs Model for P&P and Municipal ii Sludge Table 2.1 Summary of the Takacs Model Sensitivities Analysis - Settling Velocity 17 Table 2.2 Summary of the Takacs Model Sensitivities Analysis - Concentration Profile 18 Table 2.3 Summary of Calibrated Parameters of the Takacs Model from the Literature 20 Table 4.1 Influent Composition of the P&P AST Plant 29 Table 4.2 Summary of Primary and Secondary Clarifiers at the Port Alberni P&P Mill 31 Table 4.3 Summary of Primary and Secondary Clarifiers at the Comox WWTP 41 Table 5.1 Operating Conditions of the P&P Primary Clarifier During the Settling Tests 49 Table 5.2 Settling Velocities of P&P Primary Sludge - Test # 1 52 Table 5.3 Settling Velocities of P&P Primary Sludge - Test # 2 54 Table 5.4 Settling Velocities of P&P Primary Sludge - Test # 3 56 Table 5.5 Summary of the Parameters of the Takacs Model for P&P Primary Sludge 58 Table 5.6 The Operation condition of P&P Secondary Clarifier During the Settling 59 Tests Table 5.7 Settling Velocities of P&P Secondary Sludge - Test # 1 61 Table 5.8 Settling Velocities of P&P Secondary Sludge - Test # 2 63 Table 5.9 Settling Velocities of P&P Secondary Sludge - Test # 3 65 Table 5.10 Summary of the Parameters of the Takacs Model for P&P Secondary Sludge 67 Table 5.11 Operating Conditions of the Municipal Primary Clarifier During the Settling 68 Tests Table 5.12 Settling Velocities of Municipal Primary Sludge - Test # 1 70 Table 5.13 Settling Velocities of Municipal Primary Sludge - Test #2 72 vii Tables Page Table 5.14 Settling Velocities of Municipal Primary Sludge - Test #3 74 Table 5.15 Summary of the Parameters of the Takacs Model for Municipal Primary 76 Sludge Table 5.16 The Operating Conditions of the Municipal Secondary Clarifier During the 77 Settling Tests Table 5.17 Settling Velocities of Municipal Secondary Sludge - Test #1 79 Table 5.18 Settling Velocities of Municipal Secondary Sludge - Test # 2 81 Table 5.19 Settling Velocities of Municipal Secondary Sludge - Test #3 83 Table 5.20 Summary of the Parameters of the Takacs Model for Municipal Secondary 85 Sludge Table 5.21 Summary of the Parameters of the Takacs Model for Pulp and Paper and 87 Municipal Sludge Table 5.22 Pilot Scale Simulation for the Settling of P&P Primary Sludge 93 Table 5.23 Pilot Scale Simulation of the Settling of P&P Secondary Sludge 96 Table 5.24 Pilot Scale Simulation of the Settling of Municipal Primary Sludge 98 Table 5.25 Pilot Scale Simulation of the Settling of Municipal Secondary Sludge 100 Table 5.26 Full Scale Simulation of the P&P Primary Clarifier 104 Table 5.27 Full Scale Simulation of the P&P Secondary Clarifier 107 Table 5.28 Full Scale Simulation of the Municipal Primary Clarifier 109 Table 5.29 Full Scale Simulation of the Municipal Secondary Clarifier 112 Table 5.30 The Evaluation of Mass Balance for the Takacs Model 115 Table 5.31 Sensitivity of the Simulation of the P&P Primary Clarifier 117 SSin vs SSeff-and Waste Sludge Table 5.32 Sensitivity of the Simulation of the P&P Secondary Clarifier - SS;n vs SSeff, 118 RAS and Waste Sludge Table 5.33 Sensitivity of the Simulation of theP&P Primary Clarifier - Q i n vs SSeffand 118 Waste Sludge Table 5.34 Sensitivity of the Simulation of the P&P Secondary Clarifier - Q i n vs SSeff, 119 RAS and Waste Sludge Tables Page Table 5.35 Sensitivity of the Simulation of P&P Clarifiers - Qw vs SS e f f and Bottom 119 Sludge Table 5.36 Sensitivity of Simulation of P&P Clarifiers - Depth vs SSeff and Bottom 120 Sludge Table 5.37 Concentrations of P&P Primary Sludge Thickening with Time 122 Table 5.38 Precision of Solids Analysis 123 Table 6.1 Summary of the Parameters of the Takacs Model for P&P and Municipal 124 Sludge \ Fig 2.1 Paragenesis diagram for particle settling behavior 5 Fig 2.2 Chronological Process of a Batch Settling Test in Different Stages. 6 Fig 2.3 Settling of Sludge Interface in Different Stages. 7 Fig 2.4 Settling Velocity Curve of the Takacs Model 14 Fig 2.5 Soils Balance Across Settler Layers 16 Fig 4.1 The Manufacturing Process of the Port Albemi P&P Mil l 26 Fig 4.2 Flow Chart of the Effluent Treatment Plant at the Port Alberni P&P Mill 28 Fig 4.3 Schematic diagram of the Primary Clarifier at the Alberni Specialties 33 Fig 4.4 Schematic diagram of Secondary Clarifier #1 at the Alberni Specialties 34 Fig 4.5 Flow Chart of the Comox Valley WWTP 36 Fig 4.6 Influent Flow Pattern of Comox Valley WWTP 37 Fig 4.7 Schematic Diagram of the Primary Clarifiers at the Comox Valley WWTP 39 Fig 4.8 Schematic Diagram of the Peripheral Feeding Secondary Clarifier 40 Fig 4.9 Settling Cylinder for Zone Settling Tests 43 Fig 4.10 Schematic of Pilot Scale Test Apparatus 45 Fig 5.1 Hindered Settling Test # 1 of P&P Primary Sludge 52 Fig 5.2 Calibration of rh for P&P Primary Sludge - Test #1 53 Fig 5.3 The Curve of the Takacs Model for P&P Primary Sludge - Test #1 53 Fig 5.4 Hindered Settling Test # 2 of P&P Primary Sludge 54 Fig 5.5 Calibration of rh for P&P Primary Sludge - Test #2 55 Fig 5.6 The Curve of the Takacs Model for P&P Primary Sludge - Test #2 55 Fig 5.7 Hindered Settling Test # 3 of P&P Primary Sludge 56 Fig 5.8 Calibration of rh for P&P Primary Sludge - Test #3 57 Fig 5.9 The Curve of the Takacs Model for P&P Primary Sludge - Test #3 57 Fig 5.10 The Curve of the Calibrated Takacs Model for P&P Primary Sludge 5 8 Fig 5.11 Hindered Settling Test # 1 of P&P Secondary Sludge 61 Fig 5.12 Calibration of rh for P&P Secondary Sludge - Test #1 62 Fig 5.13 The Curve of the Takacs Model for P&P Secondary Sludge - Test # 1 62 Fig 5.14 Hindered Settling Test # 2 of P&P Secondary Sludge 63 Fig 5.15 Calibration of rh for P&P Secondary Sludge - Test #2 64 Fig 5.16 The Curve of the Takacs Model for P&P Secondary Sludge - Test #2 64 Fig 5.17 Hindered Settling Test # 3 of P&P Secondary Sludge 65 Fig 5.18 Calibration of rh for P&P Secondary Sludge - Test #3 66 Fig 5.19 The Curve of the Takacs Model for P&P Secondary Sludge - Test #3 66 Fig 5.20 The Curve of the Calibrated Takacs Model for P&P Secondary Sludge 67 Fig 5.21 Hindered Settling Test # 1 of Municipal Primary Sludge 70 Fig 5.22 Calibration of rh for Municipal Primary Sludge - Test #1 71 Fig 5.23 The Curve of the Takacs Model for Municipal Primary Sludge - Test #1 71 Fig 5.24 Hindered Settling Test # 2 of Municipal Primary Sludge 72 Fig 5.25 Calibration of rh for Municipal Primary Sludge - Test #2 73 Fig 5.26 The Curve of the Takacs Model for Municipal Primary Sludge - Test #2 73 Fig 5.27 Hindered Settling Test # 3 of Municipal Primary Sludge 74 Fig 5.28 Calibration of rh for Municipal Primary Sludge - Test #3 75 Fig 5.29 The Curve of the Takacs Model for Municipal Primary Sludge - Test #3 75 Fig 5.30 The Curve of the Calibrated Takacs Model for Municipal Primary Sludge 76 Fig 5.31 Hindered Settling Test # 1 of Municipal Secondary Sludge 79 Fig 5.32 Calibration of rh for Municipal Secondary Sludge - Test #1 80 Fig 5.33 The Curve of the Takacs Model for Municipal Secondary Sludge - Test #1 80 Fig 5.34 Hindered Settling Test # 2 of Municipal Secondary Sludge 81 Fig 5.35 Calibration of rh for Municipal Secondary Sludge - Test #2 82 Fig 5.36 The Curve of the Takacs Model for Municipal Secondary Sludge - Test #2 82 Fig 5.37 Hindered Settling Test # 3 of Municipal Secondary Sludge 83 Fig 5.38 Calibration of rh for Municipal Secondary Sludge - Test #3 84 Fig 5.39 The Curve of the Takacs Model for Municipal Secondary Sludge - Test #3 84 Fig 5.40 The Curve of the Calibrated Takacs Model for Municipal Secondary 85 Sludge Fig 5.41 Piolt Scale Simulation for the Settling of P&P Primary Sludge (Ln scale 94 inSS) Fig 5.42 Piolt Scale Simulation for the Settling of P&P Secondary Sludge (Ln 97 scale in SS) Fig 5.43 Piolt Scale Simulation for the Settling of Municipal Primary Sludge (Ln 99 scale in SS) Fig 5.44 Piolt Scale Simulation for the Settling of Municipal Secondary Sludge 101 (Ln scale in SS) Fig 5.45 Full Scale Simulation of P&P Primary Clarifier (Ln scale in SS) 105 Fig 5.46 Full Scale Simulation of P&P Secondary Clarifier (Ln scale in SS) 108 Fig 5.47 Full Scale Simulation of Municipal Primary Clarifier (Ln scale in SS) 110 Fig 5.48 Full Scale Simulation of Municipal Secondary Clarifier (Ln scale in SS) 113 Fig 5.49 The Concentrations of Underflow Sludge Increase with Time 122 Acknowledgement I would like to express my gratitude to the following individuals and organization: o Dr. Eric Hall, my thesis supervisor, for his assistance throughout on the creation of this thesis, o Dr. Troy Vassos, my co-supervisor, for his help with measurement equipment and assistance on this thesis, o Larry Cross for his assistance with data collection and experimentation at the Port Alberni P&P mill and Jim Elliott for his assistance at the Comox Valley wastewater treatment plant, o Pablo Baranao and Joan Liu, for helping with mathematical programming, and o NSERC (Natural Sciences and Engineering Research Council of Canada), for their financial support. The assistance given by all of the above was very much appreciated. 1. Introduction Sedimentation is an important method of removing suspended solids from wastewater. It is also a vital component of biological wastewater treatment systems. In the conventional activated sludge system, settleable organic and inorganic matter is removed by gravity sedimentation in primary clarifiers, which reduces the load of both suspended solids (SS) and biochemical oxidation demand (BOD) for biological treatment. In a secondary clarifier, SS and BOD contained in suspended solids are removed from the wastewater to meet an effluent standard. The secondary clarifier also allows part of the settled biological sludge to be returned to the bio-reactor to maintain the substrate to biomass ratio (F/M ratio) and mean sludge age (6C). The excess biological sludge is also typically wasted from the secondary settler. The efficiencies of both the primary and secondary clarifiers directly affect the biological treatment stage and the quality of the treated effluent. A number of models, including one-dimensional, two-dimensional and three-dimensional models, have been proposed to describe the behavior of sedimentation systems. In one-dimensional models, the settlers are divided into several fixed layers, within which the suspended solids concentration is assumed to be constant and the settling velocity is assumed to be dependent on the suspended solids concentration. In 1991, Takacs et al. (1991) proposed a model derived from traditional flux theory, which assumed that at high SS concentration, the settling velocity decreases with the concentration of sludge, but below a certain concentration (threshold sludge concentration), the settling velocity increases with the sludge concentration. Takacs' dynamic one-dimensional model was identified as the most effective among the one-dimensional models (Grijspeerdt et al., 1995). Almost all the research conducted concerning the Takacs model has been carried out for simulating municipal wastewater treatment facilities. At University of British Columbia (UBC), Sreckovic (2001) used the Takacs model to simulate the primary and secondary clarifiers of the Port Alberni pulp and paper mill. Satisfactory results were obtained for underflow SS, but the model was poor for simulation of overflow SS. Sreckovic used the default parameter values for municipal wastewater, without calibrating them specifically. Also, the limited historical data available from plant records were not sufficient to describe and predict the behavior of the clarifier. In his report, Sreckovic recommended future work to apply the model with data originating from a full scale treatment plant treating pulp and paper mill effluent. The main objective of the present research was to calibrate the parameters of the Takacs model for pulp and paper sludge and to identify how well the Takacs model works on pulp and paper wastewater clarifiers compared to those for municipal wastewater treatment. From an engineering perspective, not only was the effluent S S simulated, but the waste sludge was also modeled. The approach taken for both pulp and paper clarifiers and municipal clarifiers included: o calibration of model parameters, „ pilot scale tests, „ full scale analysis, and „ short term simulation. 2. Literature Review 2.1 Settling Models Devolvement 2.1.1 Discrete sedimentation The foundation of sedimentation theory was developed by Hazen (1904). He proposed a theory for continuous sedimentation of discrete particles having a common settling velocity according to the size and shape of the particle. In his model, the fractional removal was a function of the relative overflow rate (settling velocity/ hydraulic loading rate) with a situation where the settling velocity equals the hydraulic loading rate. Camp (1936, 1946) improved Hazen's settling theory by including discrete particles that had a distribution of settling velocities. Camp assumed an ideal basin with homogeneous horizontal flow, uniform inlet distribution, free settling and solids removal at the bottom. The effluent solids concentration was only dependant on the overflow rate and particle settling velocity, and independent of the depth of the settler and the hydraulic retention time. A major criticism of Hazen and Camp's theory was that the model focused only on the removal of solids from a suspension without considering the high concentration settling phenomena occurring within the settler with high solids concentration (Dick, 1970). 2.1.2 Solids Flux Model Settling Classification of Sludge Eckenfelder et al. (1957) and Ekama et al. (1997) described the settling behavior of suspended solids that can be in one of four markedly different regimes, which are governed principally by the dilution of the solids (concentration) and the relative tendency of the particles to flocculate. The relationships between concentration and flocculation and the four regimes are shown on Figure 2.1. Class I settling (discrete non-flocculent settling): the particles represented on the top left corner of the figure are completely dispersed, with no tendency to flocculate. Each particle could settle at its own characteristic terminal velocity. Class II settling (discrete flocculent settling): the particles represented on the top right corner are still discrete, but with a strong tendency towards flocculation. With time, the particles collide and form small floes, which have an increasing tendency to settle individually. Class III settling (zone settling): in the regime of the middle area of the figure, floes are close enough to form a uniform matrix. The forces among floes are sufficiently strong to drag particles along in the same direction at the same velocity. Thus a clear interface forms between the supernatant liquor above and the subsiding solids. Class IV settling (compression settling): the bottom area of the figure represents the idea that when the concentration increases, particles are not solely supported by hydraulic forces but come into closer mechanical contact with one another. In the compression settling zone, the compressive force set up between particles squeezes the particle layers together and forces the water upward. In biological wastewater treatment, the discrete solid particles of primary sludge tend to be represented by the left side portion of Figure 2.1. However, since secondary sludge is more homogeneous and easily forms floes in the liquor, it tends to fall on right side portion of the figure. Low Cone. High Cone. Class I discrete non-flocculent Class II discrete flocculent Class IV compression settling Particulate Flocculent Figure 2.1 Paragenesis diagram for particle settling behavior (redrawn from Ekama et al, 1997 ) Settling Behavior of Sludge Ekama et al (1984) considered a batch settling test column filled with sludge of a certain concentration such that zone settling occurs instantly (see Figure 2.2 (a)). Initially, the concentration is uniform and region B occupies the whole depth. Immediately after settling occurs, a solid-liquid interface develops in region A, and a clarified liquor is formed (see Figure 2.2 (b)). In region B, solids still settle at a constant velocity, which is a function of the concentration at the start of the test. At the same time, the compression of sludge in region D and the transition stage from zone settling to compression in region C occurs. Region C is still a zone settling region but the concentration increases down the zone. The depth of the transition zone is generally considered to remain constant until the solid-liquid interface reaches it. The order of appearance at the solid-liquid interface of the four regimes of settling are: lag stage, hindered or zone settling stage, transition stage and compression stage (Figure 2.3). In the lag stage, the relatively high turbulence which results from filling the measuring column, dissipates slowly, allowing the movement of sludge caused by turbulence to decrease to zero. After the lag stage, zone settling occurs as described above. The solid-liquid interface attains a constant settling velocity, which is called hindered or zone settling. In the transition stage, denser and denser concentrations appear at the interface as the upper, less concentrated sludge sinks into the lower, more concentrated layers. The settling rate of the interface decreases gradually. The transition stage ceases when the last layer sinks into the compression layer. The compression stage follows the transition stage. As the interface continues to subside, the velocity is governed by the compression behavior of the sludge. With time, when the concentration of solids is higher, the settling rate decreases. V 1,000 6 s "p &'5O0 V 1 - i . . - \ 2 - l-„ \ 1 1 1 - 1 1 I " 1: Lag phase " ": ' , 2: Zone settling phase 3: Transition phase 4: Compression phase 7 SVao = \ t _ 230 ml l • l ~ • tSVI-= S V 3 o / X ( m l / g ) ' - - - « , - -• -i • i i i • t - r--1 , i A 40 - 80 Time (min)" ' 120 Figure 2.3 Settling of Sludge Interface in Different Stages ( from Ekama et al, 1997 ) Flux Theory Vesilind (1968) and Dick and Young (1972) proposed the following mathematical expression linking the settling velocity of the solid-liquid interface (Vs) and the suspended solids concentration (X), respectively: Vs = V0 Exp ( - n X) (2.1) Vs = Vo (X) - (2.2) where n is the constant of the flux theory and V0 is the theoretical maximum settling velocity. In an ideal clarifier, the movements of sludge and water are in the vertical direction only. The mixed liquor flows into the clarifier and then separates at the point of entry into the underflow, moving downwards, and the overflow, which moves upwards. The flux theory is considered to be a one-dimensional model which states that solids are carried into the bottom of the clarifiers by the gravity settling flux (Js) and the bulk flux (JB). The bulk flux results from water moving downwards due to recycling or wasting pumps, which withdraw liquor from the bottom of the clarifier (Qù). The total solids flux (Jr) is the sum of the gravity settling flux and the bulk flux. JT = Js+JB=Voexp(-nX)+Xqu (2.3) where qv is surface underflow rate, which equals the underflow rate (Qu) divided by the cross-sectional area of the clarifier. Based on a mass balance of the solids flow into and out of the clarifier, for some selected surface underflow rates smaller than the critical rate (qv < qu.cRir), a specific value of Xm Equation 2.3 makes JT a minimum. This X concentration and the associated flux are called the limiting concentration (XL) and flux (JL), respectively. This limiting flux constrains the rate at which solids can reach the bottom of the clarifier, and therefore the applied solids loading rate or flux (JQF) must not exceed this limiting flux for safe operation of the clarifier (JQF JL). 2.1.3 One-Dimensional Models In the one-dimensional approach to clarifier modeling, solids and liquid movement in the vertical direction are assumed to be dominant and horizontal movement is ignored. The settling tank is divided into a number of layers in the vertical direction and a numerical technique is used to solve the mass balance equations in the vertical direction. The solution to the mass balance equations provides the solids concentration profile in the settling tank, and the solids concentration in the effluent and underflow. When considering a one-dimensional model of a settling tank, three separate zones must be considered (see Section 2.4): 0 the zone above the feed layer, „ the feed layer, and 0 the zone below the feed layer. Above the feed layer, the bulk fluid movement (i.e. surface overflow rate) is upwards; therefore, any solids transport associated with the bulk fluid movement is upwards as well. Below the feed layer, the bulk fluid movement is downwards, so solids transport associated with the bulk fluid movement also is downwards. At the feed layer, the solids mass loading of the feed must be considered, and there is both upwards and downwards bulk fluid movement. Also, the top layer in the first zone and the bottom layer in the third zone require special consideration. The top layer in the first zone is unique as there is no solids flux into it from above. The bottom layer in the third zone is also unique as there is only bulk flux out of it. Another requirement of the one-dimensional settling models is a quantitative relationship between solids concentration and settling velocity. Busby (1973), Stenstrom (1976) and Hill (1985) simulated the thickening process using the solids flux theory in a one-dimensional, layered settler. They used a fixed number of layers of constant thickness to describe the thickening function of the settlers. Vitasovic's study (1986, 1989) considered the upward bulk movement of the liquid in the layers above the feeding point and thus described the process more accurately. However, the Vitasovic model did not include a clarification component, so the suspended solids in effluent were not predicted during normal operating conditions. Five different groups of layers were present in the model, depending on the position relative to the feeding layer: Toy layer (1) d z ( AX / d t ) = J u p ( 2 - „ - J i ( 1 - 2 ) - ( Q e / A ) X, (2.4) Layers above feeding layer (2 to m-1) d Z ( AX / d t ) = Jup(n-l-n) " Jup(n-n+l) + Jj(n-l-n) - Js(n-D+1) (2-5) Feeding Layer (m) dz( A X / d t ) = ( Q j / A ) Xj„ - Jup(m^m-l) - J_n(m->m-+l) + Js(m-l^m) " Js(m-m+l) (2-6) Layers below feeding layer Cm+1 to b-1): d z ( AX / d t ) = Jdn(n-l^n) " Jdn(n-n+l) + Js(n-l~n) - Js(n^n-l) (2-7) Bottom layer (b): d z ( AX / d t ) = Jd.Kb.i-b, + J s ( b . , ~ b ) - ( Q u / A ) X b (2.8) where dz is the thickness of the layers, J_„ is the solids flux caused by the bulk flow to the underflow, J, is the flux due to gravity settling, Qe is the upflow rate, Qu is the underflow rate, Xt is the solids concentration of the top layer, Xb is the solids concentration of the bottom layer, and A is the surface area of the clarifier. Some researchers have developed different models using settling velocity functions, including: the Vesilind model (1968), the Takacs model (1991), the Hartel Q function (1992), the Otterpohl model (1992), the Cho model (1993) and the modified Vesilind model by EnviroSim Associates Ltd., Flamborough, Ontario. Anderson (1981), Hamilton et al. (1992) and Ozinsky et al. (1994) also introduced an additional eddy diffusion term to modify one-dimensional models. 2.1.4 Two- and Tree Dimensional Model In a real clarifier, there are many complex interrelationships between performance and boundary or flow conditions that cannot be reflected in a 1-dimensional model. Three categories of such influences can be identified (Ekama etal., 1997): 0 geometry, for example, shape of tank, inlet and outlet arrangement, 0 flow, for example, density effects causing uneven velocity profiles, short-circuits from the inlet to the outlet and an adverse residence time distribution, resuspension of settled floes from the surface of the sludge blanket, and turbulence effects, and „ the sludge removal process, which is a cause of various unsteady effects and where most of the sludge at the bottom of the tank is diluted. Instead of the upward and downward (vertical) movement of solids and water of a 1-dimensional model, 2-D models also consider the movement in a horizontal direction. Hie relationship between hydrodynamics and sedimentation was first discussed by Anderson (1945), and Larsen (1977) who did the pioneering work on rectangular clarifiers. Schamber, et al. (1981), Imam et al. (1983), Abdel-Gawad (1984), Celike et al. (1987), Lyn and Radi (1989), Lyn et al. (1992), Adams et al. (1990), Stamou et al. (1989), Stamou (1991), Szalai et al. (1994) and Krebs (1995) made further progress on the 2-D simulation of sedimentation. There are at least three types of equations that are used to describe the hydraulics of 2-D settling tanks: the continuity equation, the momentum equation and the turbulence modeling equation. Circular clarifiers generally have axial-metric flow and can be simulated by 2-D models. Rectangular and square tanks tend to exhibit more complex flow and might require 3-D models (Sreckovic, 2001). In 1995, a 3-D model was successfully applied to the rectangular clarifiers at Passaic Valley wastewater treatment plant (WWTP) in Newark, N.J., USA (Zhou et al. 1996). However, there are more factors that need to be considered in applying 3-D models such as inlet kinetic energy, potential energy, gravitational work, energy dissipated in friction etc, which make application of these models difficult (Ekama, 1997). 2.2 Model Selection The two and three-dimensional models that describe the complex hydraulics occurring in clarifiers are the most comprehensive ones. However, their mathematical complexity makes these models difficult to apply. In addition, these models are complex to define and calibrate. Thus, the one-dimensional model was chosen as a simulator for the present study. Grijspeerdt et al. (1995) evaluated the six one-dimensional models proposed by Laikari (1989), Takacs et al. (1991), Otterpohl and Dahl, (1995), Dupont and Henze (1992), Hamilton et al. (1992), as well as a combination model of Takacs et al. (1991) and Otterpohrs (1992). The study concluded that the Takacs model provided the most realistic results when compared with the others, for both steady state and dynamic simulations. The Takacs model has also been used in the upgraded version of BioWin32 (EnviroSim Associates Ltd., Flamborough, Ontario, Canada), which is a Microsoft Windows-based simulator used world-wide in the analysis and design of wastewater treatment plants. Sreckovic (2001) also applied this model to his research on pulp and paper wastewater treatment. Sreckovic chose the Takacs model because it is relatively simple, but provides reasonably good results and had already been studied by some researchers. Therefore, the Takacs model was also selected for the research reported in this thesis. 2.3 The Takacs Model Takacs et al. (1991) presented a double exponential settling function to describe the change in settling velocity Vs with concentration Xover the entire concentration range: Vs = V0 expf - rh * (X-Xmin )] - VQ expf -rp*(X- Xmin )] (2.9) Where : Vs sludge settling velocity (m/d) X concentration of solids (mg/L or g/m3) V0 maximum theoretical settling velocity (m/d) rh settling parameter associated with the hindered settling component oftheSSCmVg) rp settling parameter associated with lower concentrations and slower settling velocity component of the SS (m3/g) Xmin non-settleable solids concentration (mg/L or g/m3) Four concentration regions that are treated separately are defined, Region I to Region IV. (see ;ure 2.4) o Region I : The non-settleable solids concentration Xmi„ is defined as a fraction of influent X.When the concentration is lower than Xmim the settling velocity is zero. The non-settleable concentration is subtracted from the total solids X, such that the settling velocity function is related to (X-Xmin ) rather than to X. o Region II : Parry and Takacs (1992) observed that, in this concentration range, the sludge 13 is flocculent and settling velocity is a function of the local concentration. Therefore, the mean particle diameter, and hence the settling velocity, increase with concentration. Because rp is typically one order of magnitude higher than rh, the secondary term on the right hand side of Equation 2.9 rapidly diminishes with increasing concentration and the settling velocity is influenced only when X is low. o Region III : As the solids concentration increases, a region exists where the average settling velocity reaches a maximum value V0'. This region is assumed to be the transition zone between the low solids concentration region (Region II) and the hindered settling condition (Region TV). 0 Region IV: This region is dominated by the zone settling behavior of the sludge. At relatively high concentrations, the first term on the right hand side of Equation 2.9 is dominant and the function is very like the classic flux theory (Equation 2.1). Settling Vo Velocity \ Vo' i 1 m j IV Solids concentration Figure 2.4 Settling Velocity Curve of the Takacs Model The Takacs et al (1991) modeling process is based on the work of Vitasovic (1986) and considers the low concentration settling of the clarification layers. They also added the threshold concentration X, to the model for each layer, which is designed to limit the downward flux of solids that can be handled by the layer below. The modified formulations are summarized as Figure 2.5. To apply the model, one has to know that the dynamic process is based on the following assumptions: „ the gravity settling velocity is related to the solids concentration, » the solids concentration of each layer is independent of the other layers, 0 the solids concentration is uniform in each layer, and » the overflow and underflow velocities are uniform in each layer. 2.4 Sensitivities of the Takacs Model Kennedy (1994) attempted to reach a thorough understanding of the Takacs model through sensitivity analysis. The parameter values were varied individually to determine the corresponding impact on the settling velocity model and the predicted concentrations at each layer. Kennedy concluded that the parameter V0 is critical for concentrations less than 10,000 mg/L, rAis highly sensitive for concentrations greater than 2,000 to 30,000 mg/L and rp is influential in the low concentration ranges below 500 mg/L. The summaries of the sensitivities analysis are shown as Table 2.1 and Table 2.2, respectively. In this study, the manner in which the variation of influent SS, the influent flow rate and the underflow rate affect the simulation results will be discussed later on in Section 5.3.5. BULK M O V E M E N T T O P L A Y E R Q *X ->* 1 GRAVITY SETTLING Layer 1 J up, 2 - Q e X 2 L A Y E R S A B O V E F E E D L A Y E R J up, 3 = Q e X3 J up, m — Q e X m F E E D L A Y E R -Q i n * X i n / A J dn, m+1 = Q e Xm+i L A Y E R S B E L O W F E E D L A Y E R J dn, m+1 — Q u X m +1 J dn, b-1 = Q u Xb_i B O T T O M L A Y E R Layer 2 t + Layer m J S i 1 = min ( V S 1 X L V S 2 X 2 ) or J s , i = V s 1 X 1 , i f X 2 5 ï X t J s 2 = min (V S 2 X 2 ,Vs3X 3 ) or Js,2 = VS2X2 , if X 3 ^ X f J s , m - i = min ( V S m - i X m . i , V S m X m ) or «Js,m-1 - V s i X ^ . i f X ^ X , 1 / + + > J Layer m+1 s , m - rnin ( V s m X m , V s m + i X m + i ) ^ Js , m+1 = min (Vsm+lX m +i ,Vsm+2X m + 2) J S . M = min ( V s b - i X b - i , V b X b ) v Q under X b / A Figure 2.5 Solids Balance Across Settler Table 2.1 Summary of the Takacs Model Sensitivity Analysis - Settling Velocity (from Kennedy, 1994) Parameter Parameter Impact on Settling Velocity Vo „ highly sensitive for concentrations less than 10,000 mg/L. o predicted settling velocity increases with increasing V0. Vo' o very sensitive for the maximum practical settling velocity concentrations, generally less than 3,000 mg/L; settling velocities increase with increasing Vo'-0 no sensitivity for concentrations less than 50 to 100 mg/L and greater than 4,000 mg/L. n „ moderately sensitive in low concentration range (below 100 mg/L). „ highly sensitive for concentrations greater than 2,000 to 30,000 mg/L. o increasing rh has the effect of decreasing model settling velocity values. rP „ highly sensitive in the low concentration ranges (less than 500 mg/L), » slightly sensitive in the 3,000 to 5,000 mg/L range. o increasing rp has the effect of increasing model settling velocity values. Y o slightly sensitive in the low concentration ranges (less than 100 mg/L). o virtually no sensitivity in concentrations greater than 100 mg/L. „ increasing Xmin has the effect of increasing model settling velocity values. Table 2.2 Summary of the Takacs Model Sensitivity Analysis - Concentration Profile (from Kennedy, 1994) Parameter Parameter Impact on Concentration Profile Vo o highly sensitive in upper layers (above the feeding layer). 0 decreased sensitivity in the layers near the bottom layer. o increasing V0, decreases layer concentration above the feeding layer. Vo' o high sensitivity in all layers. „ increasing Vo ' decreases concentration expected below the sludge blanket. n „ moderately sensitive in layers above the feeding layer. „ low sensitivity in diluted blanket layers, including the feeding layer. 0 high sensitivity to location of sludge blanket height. 0 increasing rh decreases concentration above the feeding layer. rP „ highly sensitivity in layers above the feeding layer. „ low sensitivity in layers below the feeding layer. 0 increasing rp decreases layer concentration above the feeding layer. Xmin 0 no sensitivity expected in the top layer. o increasing Xmin has the effect of increasing the concentration of each layer. 2.5 Conclusions of Application of the Takacs Model from Literature Few studies of using the Takacs Model on municipal wastewater have been published. Takacs used the data from a secondary clarifier at an activated sludge process wastewater treatment plant under low, medium and high loading rates to calibrate the parameters of the model, and estimated V0 = 172-370 m/d, V0' = 112-150 m/d, rh = 2.93~3.78xl0-4, rp = 2.68~5.71xl0"3 and/-, = 1.23-2.59xl0"3. Takacs also verified the parameters for storm conditions as: V0 = 712 m/d, V0' = 340 m/d, rh = 4.2ÔX10"4, rp = 5.0xl0"3 and/M = 1.0xl0"4(/„s = XminIX). Both the steady state and dynamic simulations produced very good results for the clarification and thickening processes. In Kennedy's (1994) pilot scale experiment, the estimated parameters for secondary sludge were close to those used in the simulation of storm conditions in Takacs' study of 1991. They were: V0 = 634 m/d, V0 ' = 280 m/d, rh = 4.89X10 - 4, rp = 2.94xl03 andXmin = 2 mg/L. Jeppsson and Diehl (1996) proposed a model called the Diehl Model. They compared the simulation results of their model with the Takacs Model. The parameters used in their study were: V0 = 145 m/d, V0' = 100 m/d, rh = 4.20xl0"4, rp = 5.0xl0"3 andXmin = 10 mg/L. Gernaey's (1999) study focused on the application of the Takacs model to primary sludge in order to combine the simulated results with the Activated Sludge Model No.l (ASM1). The parameters were calibrated as: Vo - 96 m/d, V0' = 80 m/d, rh = 1.90X10"4, rp = 7.0xl0"4and/ s = 2.40xl0"3. Sreckovic's (2001) research was the first application of the Takacs model to pulp and paper wastewater treatment. In his study, the parameters were estimated as: V0 = 784 m/d, V0' = 433 m/d, rh = 7.10xl0"4, rp = 2.90xl0"3and f„s = 1.36x103 for both primary and secondary sludge. Those parameters were determined by empirical judgment instead of experimental calibration. The mill chosen for that research was the same as that chosen for the present study, the effluent treatment plant of NorskeCanada Inc at Port Alberni, B.C., Canada. Table 2.3 Summary of Calibrated Parameters of the Takacs Model from the Literature Researcher Vo(m/d) *V(m/d) rA(m3/g) r„(m 3 /g) fns Note Municipal Activated sludge 214.2 150.2 3.64X10-4 5.71xl0"3 1.23xl0"3 Low load Takacs 370.0 142.9 3.78xl0"4 2.68xl0-3 2.28xl0"3 Medium load (1991) 172.8 112.1 2.93xl0"4 2.70xl0"3 2.59xl0"3 High load 712.0 340.0 4.26 xlO"4 5.0 xlO"3 1.0 xlO"4 Storm condition Kennedy (1994) 634.0 279.5 4.89 xlO - 4 2.94 xlO"3 2mg/L* Jeppsson (1996) 145.0 100.0 4.20 xlO"4 5.0 xlO"3 10mg/L* * V • - 'Minn Municipal Primary sludge Gernaey(1999) 96 80 1.90 xlO - 4 7.0 xlO' 4 2.40x10"3 Pulp and Paper Sludge (primary and activated sludge) Sreckovic (2001) 784 433 7.10 xlO"4 2.90x10° 1.36xl0'3 Empirical judgment 2.6 Characteristics of Sludges Only a few literature reports deal with the characteristics of sludge from pulp and paper mills. Sreckovic (2001) investigated the solids in the raw effluent of the Port Alberni pulp and paper mill including bark particles, sand, grit, coating and filter particles - which are used as additives in paper making - lime mud, green liquor dregs, lime and other chemically induced floes from the water treatment process. Calcium lignin complexes are formed during treatment, as well as fiber from wood and microbial cells from the pulp secondary treatment operations. The latter two are the most abundant and most important. The solids in municipal primary sludge may be due to sand, clay, silt, excrement, toilet paper, food waste, microorganisms and other organic and inorganic matter (Metcalf & Eddy, Inc., 2003). Toilets are known to generate 30 to 40 % of the total municipal wastewater and is assumed to be a major source of solids entering the sewer system (Friedler, 1996). North Americans produce an average of 100 -130 g (wet mass) / capita / day of faeces (Feachem, 1993). A survey in United Kingdom showed an average person uses 19.4 g of toilet paper, based on manufacturers' data. Approximately 1 g / capita / day solid waste material is screened out, while the rest flows into wastewater treatment plants (Haddon, 1995). Activated sludge biosolids (both pulp and paper wastewater and municipal wastewater) tend to flocculate after going through the aerating biological tanks. In an ideal activated sludge system, floc-forming organisms and filamentous organisms grow in balance and form strong and large floes, which enhance the settling ability of the solids in water (Ekama, 1997). Particle size distributions and flocculation behavior of activated sludge have been measured by Parker et al. (1997). They found the particle size distribution with a small particle size group in the 0.5 ~ 5 jim range and a larger particle size group in the 25-2,500 fi m range. There were few particles in the 5 ~ 25 fi m range. 2.7 Summary of Literature Review 2.7.1 Models Devolvement and Selection Three modeling approaches have been developed after decades of research: solids flux theory, one-dimensional models and multi-dimensional models. Solids flux models are empirical and steady state models. These models simplify or ignore the complex hydraulics in clarifiers. One-dimensional models are the extension of solids flux models and consider some hydrodynamic phenomena. They are comparatively simple, but provide reasonably good results. Multi-dimensional models that describe the complex hydraulics occurring in clarifiers are the most comprehensive models available. However, their mathematical complexity makes these models difficult to apply. Thus, the one-dimensional model was chosen as a simulator for the present study. Grijspeerdt et al. (1995) evaluated the six one-dimensional models. The study concluded that the Takacs model provided the most realistic results when compared with the others, for both steady state and dynamic simulations. The Takacs model is relatively simple, but provides reasonably good results and has already been studied by some researchers. 2.7.2 Characteristics of Pulp and Paper Sludge Fiber from wood and microbial cells from the pulp mill secondary treatment operations are the two most abundant and most important sources of suspended solids in raw pulp and paper wastewater. The solids in municipal primary sludge may be due to sand, clay, silt, excrement, toilet paper, food waste, microorganisms and other organic and inorganic matter. Faeces and toilet paper are the main source of suspended solids in the sewer. The activated sludge biomass formed during both pulp and paper wastewater and municipal wastewater treatment tends to flocculate after going through the aerating biological tanks. The particle size distribution seems to be dominated by a small particle size group in the range of 0.5 ~ 5 [i m and a larger particle size group in the 25-2,500 n m range. 3. Research Objectives and Approach The major research aim was to develop a comprehensive model for predicting the settling behavior of suspended solids associated with pulp and paper wastewater treatment using the activated sludge system. The literature review (Chapter 2) indicated that one-dimensional models are comparatively simple, but provide reasonably good results for realistic simulation. The one-dimensional Takacs model has been reported to be very suitable simulating sludge settling. Although, the parameters of the Takacs model for municipal sludges have been estimated by several researchers, due to differences in the characteristics of municipal and pulp and paper sludge, the parameters may need to be recalibrated for simulating pulp and paper sludge settling. A second objective was to compare settling characteristics of the pulp and paper sludges with those of the municipal sludges. The wastewater treatment plants at the Port Alberni Paper Division of NorskeCanada Inc. and the municipal treatment plant in Comox, B.C. were chosen for the present research. The approach taken for both pulp and paper clarifiers and municipal clarifiers included: „ calibration of model parameters, „ pilot scale tests, 0 full scale analysis, and „ short term simulation. For calibrating model parameters, sludge zone settling tests were conducted within testing cylinders (see Section 4.3.1). Matlab was used to solve the Takacs model equations and to simulate the settling behavior of sludge. From an engineering perspective, not only the effluent SS was simulated, but the waste sludge was also modeled. 4.1 Introduction: The Effluent Treatment Plant of NorskeCanada Inc. Port Alberni Pulp and Paper Mill The Port Alberni Paper Division of NorskeCanada Inc. is located at the head of the picturesque Alberni inlet on the west coast of Vancouver Island in the Province of British Columbia. It is one of the largest producers of telephone directory paper and lightweight coated paper in North America. The annual output of 432,000 tonnes of groundwood paper is used to print a wide spectrum of products, including telephone books, magazines, catalogues, flyers and airline schedules. The mill first operated as a Kraft mill in 1947. In 1956, it gained the paper machines to begin manufacturing paper, which it produces presently. Due to stricter environmental requirements, which were implemented at the beginning of 1991, the Kraft pulping process was replaced by the stone groundwood process and chemical-thermal-mechanical pulp (CTMP) process. 4.1.1 Manufacturing Process (see Figure 4.1, information provided by NorskeCanada Inc.,) Woodroom The products of the Port Alberni pulp and paper mill are manufactured from Douglas fir, western cedar and hemlock. Logs and wood chips are transported to the mill via water and trucks. In the woodroom, after the logs are debarked, they are either sawed into 4 foot lengths for the stone groundwood process, or chopped into chips by a chipper for the CTMP process. The Stone groundwood mill The logs are loaded onto the grindstone surface, which has thousands of grooves. These grooves fill with fiber as the stone passes over the wood. The grinder gets very hot from the friction created between itself and the wood; therefore, grinder showers are used to cool it down and clean off the stone. This also helps to suspend the fiber in the slurry. The slurry goes through a primary and a secondary screen and then enters a gravity decker to be dewatered to approximately 4%. Sodium hydrosulfite is added to brighten the pulp just prior to being pumped into the storage tank. The bleached pulp is mixed with refiner CTMP at the storage tank before being proportioned to the paper machines' specifications. The stone groundwood mill produces 630 dry tonnes of pulp per day. CTMP The CTMP plant produces 350 tonnes dry weight of CTMP per day. Chips are heated and stored for 2 hours and then are washed with hot water to remove sand and heavy materials. Next, the chips are discharged into a bin for the first stage of steaming, which takes 2 hours. The softened chips are conveyed onto a vessel and impregnated with the sulphite, breaking them into consistent sizes. The treated chips are refined by primary and secondary refiners and are screened as pulp for paper making. Pulp is pumped to the CTMP blend chest, which supplies pulp to paper machines, and is brightened by sodium hydrosulfite in line. Pulp is also pumped to and brightened at the peroxide bleach plant. Peroxide Bleach Plant The peroxide bleach plant was designed to bleach 350 tonnes of CTMP or stone groundwood pulp up to 72 brightness level. The bleach solution is made of hydrogen peroxide, sodium silicate, Epsom salts and caustic. A wash press washes the bleached pulp and discharges 2,600 m3/d of wastewater per day. Fig 4.1 The Manufacturing Process of the Port Alberni P&P Mill ( provided by NorskeCanada Inc.) -26-Paper Making Three paper machines manufacture 1,180 tons of directory paper and lightweight coated paper per day. On-site CTMP and stone groundwood pulp facilities meet most of the pulp requirements, with the remainder supplied by other providers, including Kraft pulp and de-inked pulp made from old newsprint and magazines. About 55,000 m3/d of wastewater is generated from paper machines. 4.1.2 Effluent Treatment Process The Port Alberni pulp and paper mill is located in the Alberni Valley (the community claims it is "the Salmon Capital of the world"), which abounds with natural and recreational resources. In 1991, the Province passed laws that imposed new regulations on all B.C. mills with "special regulations" for the Port Alberni pulp and paper mill, which was 2 to 3 times stricter than other B.C. coastal mills. In 1992, the Federal Government also passed "special regulations" for this mill that were similar to B.C. Provincial "special regulations". A new facility for effluent treatment was constructed to fit those requirements. The mill effluent treatment plant is divided into two separate treatment facilities. Kraft production wastewater is treated in an aerated stabilization basin (ASB), or aerated lagoon, with approximately 10 days hydraulic retention time. Paper machine wastewater is treated in an activated sludge treatment (AST) plant (see Figure 4.2 for flow chart of the treatment plant). Although these systems are operated separately, there are some crossover capabilities to protect the activated sludge system. In this research, the AST plant was chosen to do the tests. Influent The current average influent flow rate at the AST plant is about 77,000 cubic meters per day (m3/d), and is derived from the sewers shown on Table 4.1. These streams were chosen to be treated in the AST system due to either high fiber content or high BOD/toxicity levels. Average characteristics of the influent (after primary sedimentation) are: SS of 450 mg/L, pH oïl2 (after adjustment), temperature of 32°C, andBODs of 250 mg/L and COD of 600 mg/L. Influent Bar Screen Lift Station Primary Clarifier Primary sludge l Activated Sludge Basin #1~#5 Secondary Clarifier #1, #2 Effluent i l Recycling Sludge Waste Sludge Sludge Blend Tank - -• Sludge Screw Dryer - -• Power Boiler Stream Figure 4.2 Flow Chart of the Activated Sludge Treatment Plant at the Port Alberni P&P Mill (information provided by NorskeCanada Inc.) Table 4.1 Influent Composition of the P&P AST Plant (information provided by NorskeCanada Inc.) Sewer Average Flow Rate (m3/d) CTMP 9,500 Groundwood 5,000 Paper Machines 55,000 Woodroom 5,000 Bleach Plant 2,500 Total Influent 77,000 Pretreatment and Primary Sedimentation The combined paper machine effluent becomes treatment plant influent, which first enters a lift station after the pH is adjusted to near neutral, and which then passes through a bar screen. The wastewater is then pumped by variable speed pumps into the influent well in the primary clarifier. Fiber and other settleable solids settle to the bottom of the clarifier and are withdrawn by sludge pumps. Primary sludge (PS) is pumped to a blend tank, to be mixed with secondary sludge for dewatering. The primary clarifier overflow then enters the bioreactor at a feed box. Biological Treatment The paper machine effluent is biologically treated in a series of four, connected, completely mixed aeration tanks with a residual dissolved oxygen concentration of 2.0 mg/L. The system provides a nominal hydraulic retention time (HRT) of about 6 hours. The aeration tank outflow enters two central-feeding circular secondary clarifiers. Settled sludge is returned as return activated sludge (RAS) through variable speed RAS pumps. The waste activated sludge is withdrawn from the sludge hopper in the clarifier and is pumped into a sludge blend tank. The mean sludge age range is 5 ~ 7 days, with the mixed liquor suspended solids (MLSS) maintained at 2,600-3,400 mg/L, and food microorganism ratio (F/M) at 0.27 ~ 0.35. Effluent The treated effluent is combined with aerated stabilization basin (ASB) effluent before discharge into the Somass River. The design criterion for the SS of the effluent from the AST is equal to or less than 45 mg/L; the SS is below 12 mg/L. The target effluent BOD5 is 18 mg/L or less, and the actual average BOD5 is approximately 7 mg/L. Sludge Primary sludge and waste activated sludge (WAS) are well mixed in a sludge blend tank prior to rotary screen thickeners and then dewatered on two screw presses. Dewatered sludge at 35 ~ 40 % solids consistency is transferred on conveyors to the steam plant at the mill to used as fuel to generate steam for manufacturing. The press filtrate returns to the feed box of the bioreactor for treatment. 4.1.3 Primary Clarifier (see Figure 4.3 and Table 4.2) A 61-meter diameter primary clarifier with a surface area of 2,920 m 2 and a side water depth of 4.5 meters is operated to remove larger and heavier suspended solids from the wastewater. At an average flow rate of 77,000 m3/d, Jhe hydraulic retention time is 5.3 hours. The average concentration of the influent SS is 450 mg/L. With 90% removal efficiency, the effluent SS of this unit is about 45 mg/L. The central feed pipe conducts influent to an 18 m diameter and 3 m depth flocculating well, which is installed in the center. A scraper sweeps settled sludge on the 1:12 sloped floor into a sludge pit at the center. The pit is 8.5 m in diameter and 1.5 m deep. Primary sludge with a consistency of about 5% ~ 7% is discharged by variable speed primary sludge pumps at flow rates between 3 and 10 L/s. Table 4.2 Summary of Primary and Secondary Clarifier Characteristics at the Port Alberni P&P Mill Item Unit Primary Secondary Type Circular Central feeding Circular Central feeding Diameter m 61 61 Side water depth m 4.5 4.5 Bottom slope 1:12 1:12 Diameter of flocculation well m 18 18 Depth of flocculation well m 3.0 2.85 Overflow rate m/d Normal flow rate 29 15 Maximum flow rate 41 21 Solids loading kg/m2 • d Normal flow rate — 57 Maximum flow rate — 119 Hydraulic retention time hr 3.8-5.4 5.0-7.8 Influent SS mg/L ' 450 2,600-3,400 Effluent SS mg/L 45 <25 S S removal % 90 99 SS of RAS mg/L — 10,000 RAS flow rate L/s — 125-250 WAS consistency % 5.0-7.0 2.5-3.5 WAS flow rate L/s -5 -3 Mean sludge age day 5-7 4.1.4 Secondary Clarifiers (see Figure 4.4 and Table 4.2) Two parallel secondary clarifiers (# 1 and # 2) are operated to remove secondary sludge from the final effluent. In the present study, Clarifier # 1 was chosen for study and simulation. The design of secondary Clarifier # 1 is similar to that of the primary clarifier: 61 m diameter, 4.5 m side water depth, 1:12 floor slope and a sludge hopper with an 8.5 m diameter and 1.5 m deep. There are two concentric wells in each clarifier. The inner well, which is for influent distribution, has a diameter of 6.7 m and a depth of 1.9 m. The outer well, which functions as a flocculation well, is 18 m in diameter and 2.85 m. Settled secondary sludge is removed in two ways: as RAS, which is pumped out from the sludge blanket through suction tubes; and as WAS, which is removed through the center hopper. Most of the solids (RAS) are siphoned from the bottom of the clarifier by eight 20 cm suction pipes attached to each scraper arm. Solids are drawn up to a sludge sightbox. The siphoning action is created by maintaining the level of liquid in the sightbox lower than the water level of the clarifier. The liquid level in the sightbox is controlled by RAS pumps. RAS then flows into the sightbox and from the sightbox back into the center pier, after which it is sent to RAS pumps. The solids concentration of RAS is about 10,000 mg/L at a normal recycling rate of 200 L/s for Clarifier # 1. Settled solids that are not removed as RAS are drafted towards the center pit by a scraper, and are discharged by WAS pumps to the sludge blend tank. Since the WAS has longer retention time in the clarifier, the SS concentration of waste sludge is thickened to about 35,000 mg/L, and the wastage rate is about 3.2 L/s. Figure 4.3 Schematic diagram of the Primary Clarifier at the Alberni P&P Mill (from operation manual) Figure 4.4 Schematic diagram of Secondary Clarifier #1 at the Alberni P&P Mill (from operation manual) 4.2 Introduction: The Comox Valley Municipal Wastewater Treatment Plant The Comox Valley Water Pollution Control Centre (CVWPCC) is a secondary wastewater treatment facility located in Comox, B.C. and is operated by the Regional District of Comox-Strathcona for the communities of Courtenay, Comox, and CFB Comox on Vancouver Island. The CVWPCC was first proposed in 1978 as a replacement for aging municipal and military wastewater treatment facilities. From an initial concept of comminution and discharge to a long marine outfall, the design of the CVWPCC included primary and secondary treatment with aerated static pile composting of dewatered sludge. The CVWPCC has been in operation since 1984. 4.2.1 Collection System The sewer collection system is separate from the storm drainage system. Five wastewater pumping stations discharge to either a 930 mm diameter, 8 km long force main or a 400 mm diameter, 2 km force main, which join together at the plant headworks and discharge an average daily flow of 14,000 m3/d of wastewater. Four of the stations feature submersible pumps ranging in size from 20 to 75 HP while the fifth has three variable speed 175 HP pumps in a dry pit configuration. 4.2.2 Wastewater Treatment Process (see Figure 4.5) Influent The minimum flow rate in dry weather occurs at around 2:00 am to 4:00 am and the peak flow happens at around 9:00 am. The flow rates on a dry day (Apr 16, 2003) and a rainy day (Apr 8, 2003) are shown in Figure 4.6. The average BODs and SS from January 2003 to April 2003 were 138 mg/L and 217 mg/L, respectively. The average temperature in summer is 25°C, and is 12°C in winter. The pH is near neutral. Lift Station Influent Landfill Bar Screen Aeration Grit Removal Primary Clarifier P[LmAry_sJudc]e_ Activated Sludge Reactors Secondary Clarifiers > Effluent _R_eÇYÇiLnJ3 .Sludge, _Wa_s_te Sludge. _ Dissolved Air Flotation i Gravity Thickener L-H Sludge Belt Press Composition Figure 4.5 Flow Chart of the Comox Valley WWTP Flow Rate (CMM) 30 - - - -RainyDay(4/8/03) 25 % * » ' ~ * % Dry Day (4/16/03) * * * 20 ** * 15 10 5 1 1 1 1 1 1 1 1 1 1 I 0 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (Hr) Figure 4.6 Influent Flow Pattern of Comox WWTP (information provided by Comox WWTP) Pretreatment and Primary Sedimentation (Headworks) At the headworks, wastewater is screened and then degritted in aerated grit chambers before entering the rectangular primary clarifiers. Primary sludge and grit are withdrawn continuously through degritting cyclones and a grit classifier. Grit is landfilled and degritted primary sludge is thickened in a pair of gravity thickeners. Biological Treatment The biological treatment used at the CVWPCC is a conventional activated sludge system. Primary effluent passes into two aeration basins that can be operated in either a plug flow or step feed mode. Aeration rates are adjusted by means of in-tank dissolved oxygen (DO) probes which feed back to motorized inlet valves on the aeration blowers. The mixed liquor in the aeration tank flows to a pair of peripheral feed, flat-bottomed secondary clarifiers. Settled sludge is returned as return activated sludge (RAS) through a pair of variable speed RAS pumps. WAS is withdrawn downstream of the RAS pumps, then thickened by dissolved air flotation. The mixed liquid suspended solid (MLSS) is maintained at a target level of 2,340 mg/L and the mean sludge is 6 days, but at the time of study, the plant was operated with a mean sludge age of 1 ~ 4 days,. Effluent Final effluent is discharged through a re-aeration cascade to a gravity outfall which terminates in the Strait of Georgia at a distance of 3 km offshore in 80 meters of water. There is no disinfection of final effluent. The average SS value from January 2003 to April 2003 was 9 mg/L. The average BOD5 from January 2003 to April 2003, was 17 mg/L. Flow rates during this period ranged from 10,600 to 29,600 m3/d. Sludge Thickened primary sludge and WAS are stored at a sludge holding tank prior to dewatering on a 2.0 m belt filter press. Dewatered sludge of approximately 20% solids is transferred in screw conveyors and loaded into a dump truck for transporting to an offsite interim composting facility. Thickener supemate and belt press filtrate are returned to the headworks for treatment. 4.2.3 Primary Clarifiers ( see Figure 4.7 and Table 4.3) Three rectangular primary clarifiers are operated in parallel. Each of them is 32.65 m long, 6.1 m wide (the surface area is 199 m 2 for each tank) and 2.8 m deep at the side of the outlet end. The bottom floor is 1:12. The influent flows through three ports (1.2 x 0.6 m) near the water surface of each tank. The flight scraper pulls the sludge from the outlet side to a 1.4 m deep sludge hopper at the inlet side of the tank. The collected sludge is drawn off by primary sludge pumps. Primary sludge pumps are operated alternatively and constantly discharge the sludge at about 10 L/s for each clarifier. Scum is collected at the end section of each clarifier in scum pits and then the effluent flows into the launder. The primary clarifiers were designed to remove 60% of SS and 35% of BOD5. Presently, the influent SS is about 215 mg/L and the SS of the effluent from the primary clarifier is about 65 mg/L. The overflow rate is 23 m3/m2 • d with a hydraulic retention time of 3.0 hours. The concentration of primary sludge discharged by sludge pumps is about 800 mg/L. The information of the primary clarifier and the secondary clarifier introduced in next section was provided by Comox WWTP Scum draw-off Figure 4.7 Schematic Diagram of the Primary Clarifiers at the Comox WWTP (adapted from Ekama et al., 1997) 4.2.4 Secondary Clarifiers (see Figure 4.8 and Table 4.3) There are two circular secondary clarifiers at CVWPCC with diameters of 23.15 m (surface area is 420 m2), and with a depth of 3.05 m. Instead of central feeding, the influent is fed peripherally to distribute the concentration of influent solids. Beside the feed channel, a skirt baffle is fixed straight down to half of the side water depth to avoid short-circuiting currents and to enhance the settling performance. The bottom sludge is swept into a sludge hopper at the center of the clarifier and withdrawn by recycling activated sludge (RAS) pumps at the pumping rate of about 40 L/s, including waste sludge. The waste activated sludge (WAS) is discharged by the WAS pumps from the RAS recycling pipe. The plant data of effluent SS of secondary clarifiers from January to April 2003 averaged about 9 mg/L. The influent suspended solids concentration is equivalent to the concentration of mixed liquor suspended solids, which ranged from 1,100 to 2,210 mg/L. The surface overflow rate was 17 m3/m2 • d, and the solids loading rate was 34 kg/ m 2 • d. The average nominal hydraulic retention time was 3.3 hours. F e e d < m e I : " C e n t r e l i n e 1 Î -•»- - ... • * , '» 1 1 S k i r t = 4.,." mm S l u d g e b l a n k e t * i Figure 4.8 Schematic Diagram of Peripheral Feeding Secondary Clarifier (from Ekama etal., 1997) Table 4.3 Summary of Primary and Secondary Clarifiers at the Comox WWTP Item Unit Primary Secondary Type Rectangular Circular Peripheral feeding Length m 32.65 — Width m 6.1 — Diameter m — 23.15 Side water depth m 2.8 3.05 Bottom slope 1:12 0 Overflow rate (Qave) m3/m2 • d 23 17 Solids loading (Qave) kg/m2 • d — 34 Hydraulic detention time (Qave) hr 3.0 3.3 Influent SS mg/L 215 2,400 Effluent SS mg/L 65 <24 S S removal % 60 99 SS of RAS mg/L — 6,000 RAS flow rate L/s -40 SS of WAS mg/L 800 6,000 WAS flow rate L/s 10 -2 Mean sludge age day — 6 4.3 Experimental Procedures In this study, a four-step procedure was developed to simulate the settling behavior of pulp and paper mill sludge using the Takacs Model, including: parameter calibration, pilot test, full scale test and dynamic simulation. This procedure was also applied to the Comox WWTP to make a parallel comparison. 4.3.1 Batch Settling Test and Calibration of Parameters Four parameters needed to be calibrated for the modeling work. These were: rh, Vo, and rp andXm i„. a. rh - settling parameter associated with the hindered settling component of the SS and J o -maximum theoretical settling velocity To determine rh and V0, zone (or hindered) settling tests were conducted in transparent plastic cylinders. The settling cylinders were 30 cm in diameter and 65 cm in height. Each column was equipped with a 1 rpm stirrier (Kennedy, 1994, see Fig 4.9). Sludge withdrawn from full scale clarifier sludge sampling valves was diluted to a series of concentrations with clarifier supernatant, expect municipal primary sludge, which was not thick enough for the settling test. The municipal sludge with SS concentration at about 1,200 mg/L was settled for more than 3 hours and decanted the supernatant. The thickened sludge could reach a SS concentration at about 30,000 mg/L for the settling test. More than five concentrations of each sludge including both primary and secondary sludge of P&P and municipal wastewater were tested in the cylinders. At the highest concentration, it was necessary to ensure that settling of sludge blanket occurred. Similarly, for the lowest concentration, it was important that a sharp interface of sludge blanket could be detected. After the sludge was poured into the cylinders and well mixed, the positions of the descending sludge blanket interface were recorded against time. Next, the settling velocity curve for each solids concentration was obtained by plotting the height of the interface against time. Then the liquid was well mixed in the settling cylinders and samples were taken for SS analysis. The zone settling velocity (V2S) could be calculated as the maximum slope of the interface height vs time curve. From the experiment described above, the Vesilind Model parameters were estimated by calculating the linear least-square regression of the m(Vzs) against the sludge concentrations. lrpm stirrer Drainage valve Transparent plastic cylinder ( § 30cm x 65cm depth) Figure 4.9 The Apparatus of Zone Settling Tests. b / i s - non-settleable fraction of the influent solids For the non-settleable fraction of influent solids, fits, the sludge was put in a 4 L beaker for 2 hrs. The supernatant was then carefully taken out for SS analysis. The parameter f„s is the fraction of the supernatant SS (Xmin) divided by SS of original sludge (before settling). c. rp - settling parameter associated with the lower concentration and slower settling velocity component of the S S The Takacs model modified the Vesilind model by adding another exponential function (Equation 2.9) to describe the velocities in lower SS concentrations. The parameter rp is best assessed using a non-linear optimization search technique. In this study, the function "goal search" of Microsoft Excel 2002 was used to estimate the best value of rp that could make the response of the Takacs Model most closely resemble that of the Vesilind Model and with a non-seattleable concentration, Xmj„, of zero. The settling velocities of the sludge for which the concentrations were too dilute to form observable interfaces were also estimated by experiment. The slurry was diluted to several concentrations and was put into a 1 L graduated cylinder to measure the settling velocities of the floes. The movement of the floes 10 cm below the one liter graduation on the cylinder was timed over a settling distance of 5 cm. Then, a 25 mL sample was taken using a pipet from each test for SS analysis (Kennedy, 1994). 4.3.2 Steady State Test (Pilot Scale) The pilot scale tests were conducted by using a pilot setting apparatus to simulate a real clarifier (see Figure 4.10). The apparatus included: a transparent settler, which was 30 cm in diameter and 75 cm in depth with a 5 cm central feeding column that introduced the influent at 50 cm below the water surface, an influent pump, an underflow pump and a sampling pump. Al l pumps were variable speed peristaltic pumps that could control flow rates. The overflow rates were controlled at around 0.015 L/s, as this provided surface overflow rates (SOR) of about 18 m/d, and the underflow rates were in the range of 0.002 L/s to 0.005 L/s for different types of sludge to keep the sludge blanket interfaces at steady positions. Pilot scale tests were conducted while the variations of influents of the full scale WWTPs were relatively small as evidenced by plant records. For the pulp and paper mill, the experiments were conducted while the plant was in a state of normal operation; for the Comox WWTP, the tests were conducted between 10:00 am and 2:00 pm. The settler was first filled with the effluent of the clarifiers. Then, the influent of the clarifiers of wastewater treatment plants was this from was pumped into the central column with the underflow pump operating simultaneously. When the sludge blanket height reached a steady position, the apparatus was kept running for 2 hrs before taking samples. The settler height was divided conceptually into 10 layers, plus 2 layers in the sludge hopper. Each layer was 7.5 cm high. A 25 mL pipet linked to a sampling pump was used to withdraw slurry slowly from each layer for 55 analysis. Care was taken to minimize stirring of the other layers. 0 Influent — " " ^ Sample container ' • Underflow (Waste) Overflow (effluent) Diameter = 30 cm Sidewater depth = 75 cm Bottom slope =1:1 Feeding layer = 50 cm below water surface Sampling tube diameter - 5 mm Drainage Figure 4.10 Schematic of Pilot Scale Test Apparatus 4.3.3 Field Sampling Test Full scale tests were also done when the variations in the influents of the clarifiers were relatively small. In order to know the suspended solids distribution in the clarifiers, three sampling locations were chosen for each clarifier: near the inlet, in the middle, and near the outlet. According to the side water depths, the clarifiers were conceptually divided into around 10 layers. For the primary and secondary clarifiers of the pulp and paper mill treatment plant, each layer was 45 cm deep; for the primary and secondary clarifiers of the Comox WWTP, each layer was 30 cm deep. Samples were taken from each layer using a rotary pump via an 8 mm diameter transparent plastic tube. Care was taken to avoid cross-contamination from different layers and stirring of the sludge in adjacent layers. 4.4 Analytical Methods 4.4.1 Suspended Solids The suspended solids concentrations were measured by following Standard Methods for the Examination of Water and Wastewater, Section 2540-D, and the total suspended solids were dried at 103 ~ 105°C. The filters used for SS analysis were 450 mm diameter, glass fiber filters provided by the manufacture Fisherbrand. 4.4.2 Consistencies For suspended solids concentrations of more than 30,000 mg/L, consistencies were used instead of the SS analysis described by Standard Methods. The procedure was as follows. 1. A sample of thoroughly mixed sludge was placed on a tared aluminum dish which had been dried and desiccated (A, mg). 2. The gross weight of the dish with the sludge was recorded (B, mg). 3. After an overnight drying at 103 ~ 105 °C, the dish was transferred to a desiccator for more than 10 minutes. 4. The dish was re-weighed (C, mg). 5. The consistency = (C-A)/(B-A) x 100%. 4.5 Mathematical Procedures Some model calibration was completed using Microsoft Excel as mentioned in 4.3.1. Then the model formulation was programmed for simulation using Matlab. The time step (dt) used for modeling was 15 seconds, (long term dynamic simulation was 2 min or less). For simulation of the pilot tests, the initial condition was set as the concentration of the effluent which was measured in the clarifier, because the settler was filled with the effluent of the clarifiers in the beginning of the pilot test. For the full scale and dynamic simulations, the initial condition was set as the measured concentration of clarifiers. The results of the simulation were saved as Excel files for the convenience of figure drawing and data management. 4.6 Quality Assurance (QA) and Quality Control (QC) At different ranges of SS, different volumes of slurry were required for filtration. Thus, there were different degrees of error. Three ranges of SS were tested to confirm quality assurance: high concentrations (2,000 to 30,000 mg/L), medium concentrations (100 to 2,000 mg/L) and low concentrations (lower than 100 mg/L). Consistencies of solids concentrations higher than 30,000 mg/L were duplicated for QA. For quality control, more than 5% of the samples were duplicated. If the difference between these two tests was less than 10%, an average of the result was taken. If there was a difference larger than 10%, the analysis was done again. 5. Results and Discussion 5.1 Batch Settling Test and Parameter Calibration The parameters of the Takacs model were calibrated individually for each wastewater treatment plant. Three sets of data from each test for both pulp and paper mill and municipal wastewater treatment plants were collected to calculate the parameters of the model. 5.1.1 Pulp and Paper Wastewater Primary Sludge The pulp and paper primary clarifier settling tests were conducted on January 24, February 25, and April 19, 2003 while the mill was under normal operation. The surface overflow rates were below or equal to the design flow rate of 29 m/d. The flow rates and SS of the influent, SS of the effluent, and flow rates and consistencies of waste sludge at the time sludge was sampled are listed in Table 5.1. The sludge of the primary clarifier at the Alberni pulp and paper mill was taken via the sampling valve of the sludge pump to determine the model parameters. Table 5.1 Operating Conditions of the P&P Primary Clarifier During the Settling Tests Item Unit Test 1 Test 2 Test 3 Jan 24 Feb 25 Apr 19 Flow of influent L/s 700 850 750 SS of influent mg/L 330 348 405 SS of effluent mg/L 53 42 40 Wasting rate L/s 7.0 5.5 7.5 Consistency of waste sludge % 4.6 5.3 5.0 Surface overflow rate m/d 21 29 26 The sludge interface settling curves of each test are shown in Figures 5.1, 5.4, and 5.7. The zone settling velocities for the given solids concentrations were determined from the slope of the first straight line portion of the interface height versus time, and are shown in Tables 5.2, 5.3 and 5.4. According to the Vesilind model, Equation 2.1 can be fitted to the experimental data by regression to yield values for the settling parameters, V0 and n, after taking the natural logarithm of both sides of the equation. Vs = V0 exp ( - n X) (2.1) Ln(Vs)=Ln(Vo)-nX (5.1) The linear regressions ofLn(Vs) vs .A'were plotted as Figures 5.2, 5.5 and 5.8. The theoretical maximum settling velocities (V0), which were equal to the values of the EXP of the first term of the right side of Equation 5.1, for tests #1 to #3 were 130, 159, and 110 m/d with a standard deviation of 25 m/d; the constants of the flux theory n (also denoted as rh in the Takacs model) were 3.054 x 10"4, 3.105 x 10"4 and 2.024 x 10"4 m3/g, respectively, with a standard deviation of 0.610 x 10"4 m3/g. The R square numbers were between 0.95 and 0.98, which means the Vesilind model can be used to describe the settling of high concentrations of pulp and paper primary sludge very well. When the parameters V0 and rh had been estimated, the next step was to determine the non-settleable solids concentrations (Xmin) and the settling parameters of the smaller / lighter suspended solids components (rp) of the Takacs model (Equation 2.9): Vs= V0exp[-rh *(X-Xmin)J-V0exp[-rp*(X-Xmin)] (2.9) For the Xmi„, the supernate taken from the influent which was settled for 60 min was measured and calculated as 41 mg/L on January 24. The function "goal search" of Microsoft Excel 2002 was used to search for the best value of rp and Va ' that could bring the curve of the Takacs model into closer agreement with that of the Vesilind model and end the curve with a settling velocity at zero at the non-seattleable concentration Xmin. The resulting estimated values of rp were 6.435 x 10"3, 4.551 x 10"3and 3.604 x 10"3 m3/g, respectively. The standard deviation was 1.441 x 10"3m3/g. Then, at the point the curve of the Takacs model started to depart from that of the Vesilind model, the curve was cut by a horizontal line to the other side of the curve, as the settling velocity is assumed to remain constant in this region. Thus, the calibrated Takacs models were obtained and shown on Figures 5.3, 5.6 and 5.9, and the associated F0'swere 90, 95 and 74 m/d for tests #1 to # 3, respectively, with a standard deviation of 11 m/d. The mean values of the three parameter sets were taken for further simulation: r„ - 2.728 x 10"4 m3/g, rp= 4.864 x 10"3 m3/g, V0= 133 m/d, V0' = 85 m/d andXmin = 41 mg/L (Table 5.5). The model curve using the values above was redrawn as Figure 5.10. In order to find out how the predictions of the calibrated model compared to the measured data, settling velocities at the lower solids concentration were measured using the sludge set of April 19. The measured data on April 19 (test #3) and simulation results are plotted together in Figure 5.10, illustrate how the model predicated the settling velocities at both lower and higher solids concentrations. The results indicated that the calibrated Takacs model, using the parameters with the mean values from the tree tests, could adequately fit the measured data of test #3, except for the fact that measured settling velocities of SS at about 300 to 2,000 mg/L were higher than the simulation results. In Figures 5.3, 5.6 and 5.9, there were also some data with higher settling velocities than the modeled results. The measured maximum settling velocity of the interface of pulp and paper primary sludge of each test was faster than the V0' of the Takacs model. This inferred that the Region III of the Takacs model (see Section 2.3) which assumes that as the solids concentrations increase, a region exists where the average settling velocity reaches a maximum value Vo ', may not be adequate for pulp and paper primary sludge. If one ignores Region III, the Takacs model becomes a continuous curve, which is a better fit to the measured data, as shown in Figure 5.3, 5.6, 5.9 and 5.10. Table 5.2 Settling Velocities of P&P Primary Sludge - Test # 1 SS (mg/L) Interface Settling Velocity (Vs, m/d) In (Vs) 7,248 14.4 2.67 3,720 38.9 3.66 3,554 46.8 3.85 1,802 64.8 4.17 1,246 100.8 4.61 5.00 4.00 3.00 >3 2.00 1.00 0.00 y = -3.0535E-04x + 4.8647E+00 R 2 = 9.7796E-01 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 SS (mg/L) Figure 5.2 Calibration of rh for P&P Primary Sludge - Test #1 140 r 120 i Takacs Model (rp=6.435E-3, Vo'=90 nVd) \ - - - - Vesilind Model (rh=3.054E-4, Vo=130nVd) 100 i / \ • • Measured Data o _o Continuous Takacs Model > 80 C T3 u 60 A \ (ZI u 60 *_l fl 40 55 20 0 C 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 SS (mg/L) Table 5.3 The Curve of the Takacs Model fer P&P Primary Sludge -Test # 1 Settling Time (rrriri) 0 5 10 15 20 25 30 35 Figure 5.4 Hindered Settling Test #2 P&P Primary Sludge Note: The maximum slope of the curve represents the zone settling velocity of the sludge Table 5.3 Settling Velocities of P&P Primary Sludge - Test # 2 SS (mg/L) Interface Settling Velocity (Vs, m/d) ln(Fs) 7,210 17.8 2.88 6,800 21.6 3.07 5,360 27.8 3.33 3,790 39.4 3.67 2,873 58.6 4.07 2,087 80.2 4.38 1,695 122.4 4.81 1 0 I 1 1 1 1 . 1 1 1 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 SS (mg/L) Fig 5.5 Calibration of r h for P&P Primary Sludge - Test # 2 Takacs Model (rp=4.551E-3, Vo'=95 nVd) - - - Vesilind Model (rh=3.105Er4,Vo=159m/d) • Measured Data Continuous Takacs Model 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 SS (mg/L) Table 5.6 The Curve of the Takacs Model for P&P Primary Sludge -Test # 2 Table 5.4 Settling Velocities of P&P Primary Sludge - Test # 3 SS (mg/L) Interface Settling Velocity (Vs, m/d) In (Vs) 25,975 0.7 -0.36 13,800 5.8 1.76 9,500 11.1 2.41 5,600 31.7 3.46 4,700 49.0 3.89 1,990 100.8 4.61 120 100 5,000 Takacs Model (rp=3.604E-3, Vo'=74 nVd) • - - Vesilind Model (rh=2.024E-4,Vo=110nVd) A Measured Data Continuous Takacs Model 10,000 15,000 SS (mg/L) 20,000 25,000 Figure 5.9 The Curve of the Takacs Model for P&P Primary Sludge - Test # 3 Table 5.5 Summary of the Parameters of the Takacs Model for P&P Primary Sludge Primary sludge rh (xlO"4) >>(xl0-3) Vo Vo' 1 3.054 6.435 130 90 2 3.105 4.551 159 95 3 2.024 3.604 110 74 Mean value 2.728 4.864 133 86(85)* 41 Standard deviation 0.610 1.441 25 11 Coefficient of variation 22.36% 29.63% 18.56% 12.71% *: From the curve using mean values of rh and rp. 140 120 100 80 5 1 60 u on u 60 55 40 20 0 Takacs Model ( rp = 4.864Er3, Vo' = 85 nVd ) - - - VesilindModel(rh = 2.728E-4,Vo = 133nVd) A Measured Data ( test #3 ) •Continuous Takacs Model 20,000 25,000 10,000 15,000 SS (mg/L) Figure 5.10 The Curve of the Calibrated Takacs Model for P&P Primary Sludge 5.1.2 Pulp and Paper Wastewater Secondary Sludge The tests for P&P secondary sludge were done on January 23, February 24, and April 19. The manufacturing conditions on these 3 days were reported as normal by the mill. The hydraulic surface overflow rates and solids loading rates were lower than the design criteria at maximum flow rate, which were 21 m/d and 119 kg/m2 • d, respectively. The other operating conditions of the clarifier are listed in Table 4.6. The tested sludge was taken from the sampling valves of the return activated sludge of secondary Clarifier # 1. Table 5.6 The Operating Conditions of P & P Secondary Clarifier During the Settling Tests Item Unit Test 1 Test 2 Test 3 Jan 23 Fed 24 Apr 19 Flow of influent (including RAS) L/s 513 490 575 SS of influent (MLSS) mg/L 3850 3500 3800 SS of effluent mg/L 11 8 7 Flow of RAS L/s 138 190 195 SS of RAS mg/L 13090 8520 12800 Flow of WAS L/s 12.5 5.0 3.0 SS of WAS % 3.0 3.3 4.2 Mean sludge age day 4 4 8 Hydraulic surface loading rate m/d 18 17 20 Solids loading rate kg/m2 • d 68 59 75 The interface settling curves for each test are shown in Figures 5.11,5.14 and 5.17. The zone settling velocities for given solids concentrations are shown in Tables 5.7, 5.8 and 5.9. The linear regressions of Ln(V$) vs Xare plotted as Figures 5.12, 5.15 and 5.18. The theoretical maximum settling velocities (V0) of tests # 1 to # 3 were 199, 217, 261 m/d with a standard deviation of 13 m/d; The values of the constant, rh ,were 3.604 x 10"4, 2.902 x 10"4 and 2.714 x 10"4 m3/g, respectively, with a standard deviation of 0.469 x 10"4 m3/g. The R square values, at between 0.98 and 0.99, showed that the Vesilind model performed well for modeling higher concentration pulp and paper secondary solids. The sludge of January 23 was measured at 7 mg/L for the Xmin. For rp, the values were 6.099 x 10"3, 5.105 x 10"3 and 2.884 x 10 3 m3/g, respectively, and the standard deviation was 0.703 x 103 m3/g. The calibrated Takacs models were developed as shown in Figures 5.13, 5.16 and 5.19. For tests #1 to #3, the values of V0' were 127, 139 and 144 m/d with a standard deviation of 8 m/d. For simulation, the mean values of the parameters were used: rh = 3.073 xlO"4 m3/g, r„= 4.696 x 10"3 m3/g, V0= 226 m/d, V0' = 141 m/d and Xmin = 7 mg/L. Then, the model curve was redrawn as Figure 5.20. The data of settling velocities of April 19 (test # 2), which included the settling velocities of both higher and lower concentrations, are also shown in Figure 5.20 with the simulation results in order to compare them with the output of the calibrated model. The Takacs model with mean values for the parameters could fit the measured data of test #2, expect for the settling velocities at SS from about 500 to 1,500 mg/L. In Figures 5.13, 5.16 and 5.19, there were also some data with settling velocities higher than the modeled result. As discussed in the last section (Section 5.1.1), the measured maximum settling velocity of the interface of pulp and paper primary sludge was faster than the VQ' of the Takacs model. As mentioned in last section, when ignores the Region III, the Takacs model becomes a continuous curve, which is more fit to the measured data as shown in Figure 5.16, 5.19 and 5.20. Table 5.7 Settling Velocities of P&P Secondary Sludge - Test # 1 SS (mg/L) Interface Settling Velocity (16,m/d) \n(Vs) 9,126 6.3 1.84 7,246 16.6 2.81 5,374 34.2 3.53 2,694 67.0 4.20 2,272 86.4 4.46 1,261 126.0 4.84 T3 o o "aï > 3 250 200 150 100 50 Takacs Model (rp=6.099Er3, Vo'=127m/d) - - - Vesilind Model (rh=3.604E-4, Vo=199nVd) • Measured Data Continuous Takacs Model 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 SS (mg/L) 18,000 20,000 Figure 5.13 The Curve of the Takacs Model for P&P Secondary Sludge - Test # 1 Table 5.8 Settling Velocities of P&P Secondary Sludge - Test # 2 SS (mg/L) Interface Settling Velocity (Vs,m/d) ln(tt) 17,660 1.4 0.34 6,680 25.9 3.25 4,700 51.6 3.94 4,327 58.3 4.07 3,123 88.6 4.48 2,253 118.8 4.78 1,546 165.6 5.11 Takacs Model ( rp = 5.105&3, Vo' = 139 nVd) - - - - Vesilind Model( rh = 2.902E-4, Vo = 217nVd) A Measured Data 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 SS (mg/L) Figure 5.16 The Curve of the Takacs Model for P& P Secondary Sludge - Test #2 Table 5.9 Settling Velocities of P&P Secondary Sludge - Test # 3 SS (mg/L) Interface Settling Velocity (16, m/d) ln(Vs) 21,940 0.9 -0.11 15,040 4.3 1.46 13,480 4.9 1.59 7,360 26.9 3.29 4,620 63.4 4.15 3,050 151.0 5.02 1,480 216.0 5.38 300 250 S 200 u > c VI <u ao -o 3 tri T3 ^ 150 100 50 Takacs Model (rp=2.884Er3, Vo'=l 14 nVd) - - - Vesilind Model (rh=2.714Er4,Vo=261nVd) A Measured Data Continuous Takacs Model 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 22,000 SS (mg/L) Figure 5.19 The Curve of the Takacs Model for P&P Secondary Sludge - Test # 3 r Table 5.10 Summary of the Parameters of the Takacs Model for P&P Secondary Sludge Secondary sludge n (xlO4) />(xl0"3) Vo Vo' 1 3.604 6.099 199 127 2 2.902 5.105 217 139 3 2.714 2.884 261 144 Mean value 3.073 4.696 226 127(141)* 7 Stand deviation 0.469 0.703 13 8 Coefficient of variation 15.26% 14.97% 5.64% 6.70% *: From the curve using mean values of rh and rp. > 60 S u CO u 60 •a 250 200 150 100 50 • Takacs Model (rp=4.696E-3, Vo'=141m/d) - - - Vesilind Model (rh=3.073Er4,Vo=226m/d) A Measured Data Continuous Takacs Model 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 SS (mg/L) Table 5.20 The Curve of the Calibrated Takacs Model for P&P Secondary Sludge 5.7.3 Municipal Wastewater Primary Sludge The calibration procedure for the municipal sludge was the same as that completed with the pulp and paper sludge. The primary sludge of April 9, 10 and 14, 2003 was sampled for the settling tests. The weather conditions on these three days were light showers. The hydraulic surface overflow rates were a little larger than the design criterion of 25 m/d at the average flow rate, but still within the value suggested by Metcalf & Eddy, Inc. (2003), which is 24 ~ 32 m/d. The operating conditions of the clarifier are listed in Table 5.11. The settled primary sludge was taken from the sampling valves of the primary sludge pump. Table 5.11 Operating Conditions of the Municipal Primary Clarifier During the Settling Tests Item Unit Test 1 Test 2 Test3 Apr 9 Apr 10 Apr 14 Flow of influent (including RAS) L/s 68 62 58 SS of influent mg/L 207 232 253 SS of effluent mg/L 38 48 54 Flow of waste sludge L/s 9.7 9.8 9.4 S S of waste sludge mg/L 1,220 1,210 1,280 Hydraulic surface loading rate m/d 29 27 25 The municipal primary sludge used for settling velocity tests was from raw sewer, which was screened, grit removed and pH adjusted. No flocculent or other chemical were added before primary clarifiers, the sludge was "pure" primary sludge. The municipal sludge with SS concentration at about 1,200 mg/L was settled for more than 3 hours and decanted the supernatant. The thickened sludge could reach a SS concentration at about 30,000 mg/L for the settling test. The settling curves of the sludge interface of each test are shown in Figures 5.21, 5.24, and 5.27, and the zone settling velocities for given solids concentrations are shown in Tables 5.12, 5.13, and 5.14. The linear regressions of Ln(Vs) vs X are plotted as Figures 5.22, 5.25 and 5.28. The theoretical maximum settling velocities (V0) were 115, 72, 114 m/d with a standard deviation of 24 m/d; the values of the constant rh were 1.269 x 10"4, 1.608 x 10"4 and 1.584 x 10"4 m3/g, for tests #1 to # 3, respectively, with a standard deviation of 0.149 x 10"4 m3/g. The R square numbers at between 0.92 and 0.98, showed that the Vesilind model was adequate for the higher concentration suspensions of municipal primary sludge. TheXmi„ was measured as 22 mg/L using the sludge of April 9. For rp, the values were 1.772 x 10"3, 2.499 x 10'3 and 2.668 x 10"3 m3/g with a standard deviation of 4.758 x lO^mVg. The calibrated Takacs models are shown in Figures 5.23, 5.26 and 5.29. For tests #1 to #3, the values of V0 were 66, 48 and 70 m/d with a standard deviation of 23 m/d. The mean values of the parameters were rh = 1.487 xlO"4 m3/g, rp = 2.313 x 10"3 m3/g, Vo= 100 m/d, Vo' = 63 m/d andXmin = 22 mg/L (Table 5.15). The model curve with parameters above was redrawn as Figure 5.30. The data of the settling velocities of April 14, which included the settling velocities of both higher and lower concentrations, are put in Figure 5.30 to compare them with the output of the model. The curves of the calibrated Takacs model could adequately fit the measured data. Table 5.12 Settling Velocities of Municipal Primary Sludge - Test # 1 SS (mg/L) Interface Settling Velocity (Vs,m/d) ln ( Vs) 18,200 12 2.48 12,520 23.5 3.16 11,700 24.5 3.20 10,080 28.8 3.36 9,460 36.5 3.60 6,720 44.6 3.80 6,120 57.6 4.05 4,360 69.1 4.24 5.00 4.00 3.00 2.00 1.00 0.00 y = -1.2693E-04x + 4.7417E+00 R2 = 9.8268E-01 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 SS (mg/L) Figure 5.22 Calibration of rh for Municipal Primary Sludge - Test # 1 140 120 i? 100 o o •s 1 CO u OJO T3 j s 80 60 40 20 Takacs Model (rp=1.772E-3, Vo'=66 nVd) VesilindModel (rh=1.269E-4, Vo=l 15nVd) A Measured Data Continuous Takacs Model 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 22,000 24,000 SS (mg/L) Figure 5.23 The Curve of the Takacs Model for Municipal Primary Sludge - Test # 1 Table 5.13 Settling Velocities of Municipal Primary Sludge - Test #2 SS (mg/L) Interface Settling Velocity (Vs,m/d) ln(Vs) 14,050 6.6 1.89 11,767 7.2 1.97 11,275 10.1 2.31 7,633 20.5 3.02 4,500 38.9 3.66 2,567 57.3 4.05 4.50 4.00 3.50 3.00 2.50 S 2 0 0 1.50 1.00 0.50 0.00 y = -1.9968Er04x+ 4.5409E+00 R 2 = 9.8257E-01 0 2,000 4,000 6,000 8,000 10,000 12,000 • 14,000 16,000 SS(mg/L) Figure 5.25 Calibration of rh for Municipal Primary Sludge - Test # 2 u > S 1) "—' on u 00 *o 100 90 80 70 60 50 40 30 20 Takacs Model (rp=2.419Er3, Vo'=56 nVd) VesilindModel (rh=1.997&4, Vo=94nVd) • Measured Data Continuous Takacs Model 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 22,000 24,000 SS (mg/L) Figure 5.26 The Curve of the Takacs Model for Municipal Primary Sludge - Test #2 Settling Time (min) 0 10 20 30 40 50 60 70 50 1 ' Figure 5.27 Hindered Settling Test # 3 of Municipal Primary Sludge Note: The maximum slope of the curve is the settling velocity of the sludge Table 5.14 Settling Velocities of Municipal Primary Sludge - Test # 3 SS (mg/L) Interface Settling Velocity (Vs,m/d) m(Vs) 28,142 1.8 0.59 18,933 4.3 1.46 17,267 6.2 1.82 15,733 7.2 1.97 11,720 21.6 3.07 10,667 23.0 3.14 3,640 67.0 4.20 3,120 74.4 4.31 5.00 SS (mg/L) Fig 5.28 Calibration of rh for Municçal Primary Sludge - Test # 3 120 \ Takacs Model (rp=2.668E-3, Vo'=70 nVd) 100 VesilindModel(rh=1.584E-4,Vo=114m/d) \ A Measured Data 8 80 > ki Sludge Settling (m/d) o o o o Sludge Settling (m/d) o o o o Sludge Settling (m/d) o o o o Sludge Settling (m/d) o o o o k i i t i I I I I I I i Sludge Settling (m/d) o o o o m 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 22,000 24,000 SS (mg/L) Figure 5.29 The Curve of the Takacs Model for Municipal Primary Sludge -Test #3 Table 5.15 Summary of the Parameters of the Takacs Model for Municipal Primary Sludge Primary sludge rh (xlO"4) rp(xl0"3) Vo Vo' 1 1.269 1.772 115 66 2 1.997 2.419 94 56 3 1.584 2.668 114 70 Mean value 1.617 2.286 107 64(65)* 22 Stand deviation 0.360 0.462 12 7 Coefficient of variation 22.58% 20.22% 10.84% 11.27% *: From the curve using mean values of rh and rp. 60 is 1 <u — tZl u 60 •a 55 5 , 0 0 0 Takacs Model (rp=2.286E-3, Vo'=65 nVd) - - •VesilmdModel(rh=1.617E-4,Vo=107nVd) A Measured Data Continous Takacs Model é^A 1 0 , 0 0 0 1 5 , 0 0 0 SS (mg/L) 2 0 , 0 0 0 Figure 5.30 The Curve of the Calibrated Takacs Model for Municipal Primary Sludge 2 5 , 0 0 0 5.1.4 Municipal Wastewater Secondary Sludge For municipal secondary sludge, the settling tests were done on April 7th, 11th and 16th, 2003. Again, the weather on April 7 was light showers, while both Apr 11th and 16th were sunny days. The hydraulic surface overflow rates and solids loading rates were higher than the design criteria at average flow rates, 17 m/d and 34 kg/m2 • d, respectively, but still within the range suggested by Metcalf & Eddy, Inc. (2003), which are 16-28 m/d and 96-144 kg/m2/d for the surface overflow rate. The operating conditions of the clarifier are listed in Table 4.16. The tested sludge was taken from the sampling valves of the recycling activated sludge pump of secondary Clarifier # 1. Table 5.16 The Operating Conditions of the Municipal Secondary Clarifier During the Settling Tests (information provided by Comox WWTP) Item Unit Apr 7 Apr 11 Apr 16 Flow of influent (including RAS) L/s 127 119 105 SS of influent (MLSS) mg/L 1867 2067 2175 SS of effluent mg/L 18 2 5 Flow of RAS L/s 40 42 41 SS of RAS/WAS mg/L 6305 6650 6530 Flow of WAS L/s 1.4 4.4 3.8 Mean sludge age day 4 1 1 Hydraulic surface loading rate m/d 26 24 22 Solids loading rate kg/m2 • d 49 51 47 The curves of the sludge interface settling are shown in Figures 5.31, 5.34 and 5.37. The zone settling velocities for given solids concentrations are shown in Tables 5.17, 5.18 and 5.19. The linear regressions of LnfVç) and Xare plotted as Figures 5.32, 5.35 and 5.38. The parameter values for, Vo, from tests # 1 to # 3 were 200, 144, 160 m/d with a standard deviation of 29 m/d; the constant rh values were 8.874 x 10"4, 7.984 x 10"4 and 7.586 x 10"4 m3/g, respectively, with a standard deviation of 0.6595 x 10"4 m3/g. The regressions of the hindered settling test for municipal secondary sludge demonstrated high R square numbers at between 0.96 and 0.99. For the Xmin, the sludge of April 7 was measured and a minimum suspended solids concentration of 4 mg/L was determined. For rp, the values were 7.402 x 10"3, 7.720 x 10"3and 8.543 x 10"3 m3/g. The standard deviation was 0.589 x 10"3 m3/g. The calibrated Takacs Models are shown in Figures 5.33, 5.36 and 5.39. The V0 values were 81, 67, and 96 m/d for tests # 1 to # 3, and the standard deviation was 15 m/d. The mean values of the parameters were rh = 8.148 xlO"4 m3/g, rp = 7.888 x 10"3 m3/g, V0= 168 m/d, Vo -97 m/d and Xmi„ = 4 mg/L (Table 5.20). Using the mean values, the Takacs and Vesilind Models were redrawn in Figure 5.40. In the same figure, the data measured on April 16 (test # 3), including the settling velocities measured at both higher and lower concentrations, were added to compare them with the output of the models. The higher settling velocity of SS at the concentration of the Region HI of the Takacs Model (see Section 2.3) indicates the real maximum settling velocity was larger than the value o V0 ' of the Takacs Model. As mentioned in last section, when ignores the Region III, the Takacs model becomes a continuous curve, which is more fit to the measured data as shown in Figure 5.33, 5.36, 5.39 and 5.40. However, this observed maximum settling velocity existed only in a short time of the beginning of the settling process. It is also discussed in Section 5.1.5. Figure 5.31 Hindered Settling Test # 1 of Municipal Secondary Sludge Note: The maximum slope of the curve represents the zone settling velocity of the sludge Table 5.17 Settling Velocities of Municipal Secondary Sludge - Test # 1 SS (mg/L) Interface Settling Velocity (Vs, m/d) ln(Fs) 7,400 0.5 -0.69 5,400 1.1 0.10 4,380 2.6 0.96 4,280 3.7 1.31 3,300 10.8 2.38 3,100 14.1 2.65 1,820 43.2 3.77 1,027 108.0 4.68 6 0 ^ - N S "° 5 1 U 6 0 T3 En 250 200 150 100 50 •Takacs Model (rp=7.402E-3, Vo'=81 nVd) VesilindModel(rh=8.874E-4,Vo=200nVd) A Measured Data Continuous Takacs Model 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000 SS (mg/L) Figure 5.33 The Curve of the Takacs Model for Municipal Secondary Sludge - Test # 1 Table 5.18 Settling Velocities of Municipal Secondary Sludge - Test # 2 SS (mg/L) Interface Settling Velocity (Vs,mld) \n(Vs) 6,900 0.5 -0.69 6,400 0.9 -0.11 5,267 2.6 0.96 4,620 3.9 1.36 3,140 11.2 2.42 1,940 21.6 3.07 1,460 48.0 3.87 960 79.2 4.37 8,00 SS (mg/L) Figure 5.35 Calibration of n, for Municipal Secondary Sludge -Test #2 160 140 120 5 o > 100 00 ^ s 1 80 a> (Z! u 60 00 T3 J 3 EÔ 40 20 0 -Takacs Model (rp=7.72&3, Vo'=67 m/d) - — VesilindModel (rh=7.984&4, Vo=144nVd) A Measured Data Continuous Takacs Model A A 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 SS (mg/L) Figure 5.36 The Curve of the Takacs Model for Municipal Secondary Sludge - Test # 2 10,000 Table 5.19 Settling Velocities of Municipal Secondary Sludge - Test # 3 SS (mg/L) Interface Settling Velocity (Vs, m/d) In (Vs) 7,900 0.5 -0.69 4,000 6.6 1.89 3,267 9.6 2.26 2,960 17.3 2.85 2,420 20.2 3.01 1,900 37.4 3.62 1,500 60.0 4.09 670 134.0 4.90 s 6 . 0 0 5 . 0 0 4 . 0 0 3 . 0 0 2 . 0 0 1.00 0 . 0 0 - 1 . 0 0 ( - 2 . 0 0 y = -7.5863E-04x+ 5.0751E+00 R2 = 9.8130E-01 2,000 4 ,000 6,000 liJlOO SS (mg/L) Figure 5.38 Calibration of ^ for Municipal Secondary Sludge -Test # 3 180 SS (mg/L) Figure 5.39 The Curve of the Takacs Model for Municipal Secondary Sludge - Test # 3 Table 5.20 Summary of the Parameters of the Takacs Model for Municipal Secondary Sludge Secondary sludge rh (xlO"4) rp (xlO 3) Vo V0' y . 1 8.874 7.402 200 81 2 7.984 7.720 144 67 3 7.586 8.543 160 96 Mean value 8.148 7.888 168 81 (97)* 4 Stand deviation 0.660 0.589 29 15 Coefficient of variation 8.09% 7.47% 17.23% 17.89% *: From the curve using mean values of rh and rp. 180 •Takacs Model (rp=7.888E-3, Vo'=97 nVd) VesilindModel (rh=8.148E-4, Vo=168nVd) A Measured Data Continuous Takacs Model 1000 2000 3000 4000 5000 6000 SS (mg/L) 7000 -A— 8000 9000 10000 Figure 5.40 The Curve of the Calibrated Takacs Model for Municipal Secondary Sludge 5.7.5 Discussion Pulp and Paper Secondary Sludge vs. Municipal Secondary Sludge With regard to settling ability, the current research focuses more on the settleability of activated sludge. The activated sludge tends to flocculate after going through the aerating biological tanks. In ideal activated sludge, floc-forming organisms and filamentous organisms grow in balance and form strong and large floes, which enhance the settling ability of the solids in water. The settling behavior of municipal activated sludge has been thoroughly studied and well described by some models. In the present study, when calibrating the parameters rh and Vo for pulp and paper activated sludge, the average R square of 0.98 showed it followed Region IV (known as the Vesilind model) of the Takacs Model in Figure 2.4 as well as municipal activated sludge did, which also had a 0.98 R square in this study. The summary of the Parameters of the Takacs Model for pulp and paper and municipal sludge were shown in Table 5.21. The value of V0 for pulp and paper secondary sludge (V0- 226 m/d) is larger than the value estimated for municipal secondary sludge (168 m/d). However, the smaller rh of the pulp and paper secondary sludge (3.073 x 10"4m3/g), as compared with the rh of municipal secondary sludge (8.148 x lO^mVg), illustrates that the zone settling velocities of pulp and paper activated sludge are higher than those of municipal activated sludge. During the hindered settling tests, the zone settling began at concentrations of about 1,500 mg/L and 1,000 mg/L for P&P and municipal secondary sludge, respectively. P&P secondary sludge needed a higher solids concentration to form floes. The Xmin = 7 mg/L finding for pulp and paper secondary sludge was also quite similar to that of municipal wastewater, which was 4 mg/L. Region II and Region III are determined using mathematical methods. The V0' of pulp and paper secondary sludge (141 m/d) was larger than the V0' of municipal wastewater (97 m/d). The rp of pulp and paper secondary sludge (4.696 x 10" m3/g) was smaller than that of the municipal secondary sludge (7.888 x 10"3 m3/g) and also showed slower settling velocities of the pulp and paper activated sludge at concentrations below the threshold for zone settling. Table 5.21 Summary of the Parameters of the Takacs Model for Pulp and Paper and Municipal Sludges Item r„ (xlO"4) rp (xlO"3) Vo Vo' y • Pulp and paper sludge Primary sludge 2.728 4.864 133 85 41 Secondary sludge 3.073 4.696 226 141 7 Municipal sludge Primary sludge 1.617 2.286 107 65 22 Secondary sludge 8.148 7.888 168 97 4 To obtain the settling velocities at concentrations below the zone settling threshold, discrete particles were monitored and the results are shown in Figures 5.10, 5.20, 5.30 and 5.40 in order to compare them with the model predictions. These lighter and smaller discrete particles have different dimensional distributions (Parker et al., 1971; Parker, 1983); therefore, the settling velocities also changed with concentrations that contained different sizes of sludge particles. Figures 5.10, 5.20, 5.30, and 5.40 showed that the measured settling velocities at low concentrations were close to the curve of the model, which also supported the Region II of the Takacs model, as Patry and Takacs (1992) have discussed. The settling ability of activated sludge is affected by three major factors: first is the number of filamentous organisms it contains; second is the nitrogen gas generated by denitrification, and third is the bio gas generated by digestion of sludge (Ekama et al., 1997). Dominant filamentous organisms occur when the ratio of substrate to microorganisms (F/M) is low, or the dissolved oxygen (DO) is low, or for other reasons. Denitrification and sludge digestion usually happen when the sludge remains in the bottom of the clarifier too long and forms an anaerobic condition. In the present study, the 30 min sludge volume indices (SVI) of the sludge were very good (< 80 mg/L) and the supernatants were clear. Thus, neither excess filamentous organisms, nor obvious denitrification were observed. Grady et al. (1999) summarized that when the sludge retention time (SRT) is maintained at 1 to 3 days for municipal activated sludge treatment and 3 to 5 days for industrial wastewater, the sludge develops flocculent biomass. In the present research, the SRT of the pulp and paper sludge was 4 to 8 days, and the SRT for municipal sludge was 1 to 4 days. Al l tests of secondary sludge showed that it had well formed floes. Although the increased SRT might cause the loss of biomass due to more endogenous respiration and have a better settling ability, there was no noticeable difference of settleability among the sludges with different S/?rbetween 1 to 8 days. Pulp and Paper Primary Sludge vs. Municipal Primary Sludge The components of primary sludge are more complicated than activated sludge. The solids in municipal sludge may be due to sand, clay, silt, excrement, toilet paper, food waste, microorganisms and other organic and inorganic matter. The fibers released from broken toilet paper by turbulence act as flocculents which drag particles down in the sewage when it subsides, thus, causing floes to form in the sedimentation tank. The observable floe size was about 5 mm, which is larger than activated sludge floes. The solids in the raw effluent of the pulp and paper mill generally include bark particles, sand, grit, coating and filter particles - which are used as additives in paper making - lime mud, green liquor dregs, lime and other chemically induced floes from the water treatment process. Calcium lignin complexes are formed during treatment, as well as fiber from wood and microbial cells from the pulp secondary treatment operations. The latter two are the most abundant and most important (Sreckovic, 2001). The flocculation of the sludge is caused by floc-forming chemicals that are present in the effluent of the pulp and paper mill. The sizes of the floes were similar to the floes of municipal primary sludge, which were larger in size than activated sludge floes of both pulp and paper and municipal wastewater. The average R square of parameters rh and V0 for both the pulp and paper and municipal primary sludge were 0.97 and 0.98, respectively, showing that both of them could be modeled by the Takacs Model's Region IV (see Figure 2.4). The rh and V0 of the pulp and paper primary sludge (2.728 x lO^nvVg and 133 m/d) were close to those of municipal primary sludge (1.617 x 10*4 m3/g and 107 m/d) and this illustrated that the settling velocity of higher concentration sludge showed no evident difference between pulp and paper and municipal sludge. In comparing the pulp and paper primary sludge with pulp and paper secondary sludge, the zone settling of the pulp and paper primary sludge was faster. In fact, zone settling of the pulp and paper and municipal primary sludge began at concentrations of about 1,500 mg/L and 3,000 mg/L, respectively. This suggests that the pulp and paper primary sludge has a greater tendency to form floes. Zone settling of pulp and paper primary sludge happened at an initial concentration of about 1,500 mg/L. The zone setting for secondary sludge happened at an initial concentration of about 1,500 mg/L as well. The Xmin (Region I in Figure 2.4) for pulp and paper primary sludge (41 mg/L) was larger than that for municipal wastewater (22 mg/L) because there are more fine fiber residues in the pulp and paper wastewater. Also the SS of the pulp and paper primary influent was two or three times higher than that of the municipal wastewater. The values of V0' for the pulp and paper primary sludge (85 m/d) and for municipal sludge (65 m/d) were found to be similar. The fact that the rp of the pulp and paper primary sludge (5.507 x 10° m3/g) was larger than that of the municipal primary sludge (2.313 x 10"3m3/g) inferred that faster settling velocities of P&P primary sludge at lower concentrations. The settling velocities at concentrations below the zone settling threshold were measured in test #2 and #3 of pulp and paper primary sludge and test #3 of municipal primary sludge. The results are shown in the Figures 5.6, 5.9 and 5.29, accompanied by model prodictions. The particle sizes of the primary sludge in this region were varied and, as can be seen in the figures, some settling velocities were much higher than those predicted by the models. Stirring Another factor that affects the settling velocity of the sludge interface is stirring. Rachwal et al. (1982) studied the difference between stirred and unstirred fractional settled volumes (fvs) after 30 minutes using extensive sludge settleability tests. In that research, the stirred fvs was observed to be linear with concentration, where as the unstirred fvs was not. At concentrations of more than 5,000 mg/L, the unstirred fvs was essentially 100% (very poor settleability). However, with stirring, an evident clear zone above the sludge and water interface was observed, even in the concentrations as high as 10,000 mg/L. In Rachwal's study, slower settling velocities were found when the tests were administered without stirring and those data are inappropriate for the model. Maximum Settling Velocity The Vesilind model functions well as describing the particle settling behavior during Class Ill-zone settling, which was mentioned in Section 2.1.2. During the settling test, when the solids concentration was increased to form an observable sludge interface, the particles were discrete and flocculent. Not enough solids to support the interface caused the settling velocity at this concentration was higher than that predicted by the Vesilind model. This is also one of the reasons why the measured maximum settling velocity was higher than the maximum settling velocity V0 ' of the Takacs model. However, during the continuous settling process, the solids concentration changed relatively fast with time. The sludge interface caused by discrete floes existed only in the beginning of the settling process and transited to zone settling within first minute. In this study, the simulation work was conducted by using the Takacs model with four regions instead of the continuous function of the model. 5.2 Steady State Tests (Pilot Scale) Using the model parameters obtained from the laboratory scale calibration experiments (Section 5.1), continuous flow pilot tests were conducted by using a settling tank with 30 cm diameter and 75 cm water depth. All of the primary and secondary sludges of pulp and paper wastewater and municipal primary and secondary sludge were simulated by 10-layer and 30-layer Takacs model with 4 regions. 5.2.1 Primary Sludge of Pulp and Paper Wastewater The pilot tests of pulp and paper primary sludge were done on Feb 25, 2003, when the mill was functioning normally. The primary clarifier was operated at an inflow rate of 856 L/s, an hydraulic surface overflow rate of 25 m/d, influent SS of 348 mg/L, effluent SS at 42 mg/L, a underflow sludge solids consistency of 5.3%, and a sludge wasting rate of 7.7 L/s. In order to provide a proper overflow rate to the pilot settling tank (about 15 m/d), the influent of the primary clarifier was pumped at 0.014 L/s into the central feeding column. The bottom sludge was pumped at 0.002 litres per second to maintain the sludge interface at a constant position. After the interface remained at a constant position for about 2 hours, slurry was taken from each of the 10 layers of the cylinder and 2 layers in the sludge hopper below the tank, for SS ' analysis. Two sets of simulated data (10-layer simulation and 30-layer simulation) generated by the Takacs Model were compared with the measured data and are shown in Table 5. 22 and Figure 5.41 for primary and secondary sludge. The parameters used for this pilot simulation were the values list in Table 5.21. (rh = 2.728 x 10"4, rp = 4.864 x 10 - 3, V0 = 133, V0' = 85, Xmin = 41). For pulp and paper primary sludge, the differences between both the 10-layered and the 30-layer simulations at the top layers were -28.1%, and -65.1% and -64.8%) for the waste sludge. The performance of the 10-layer and the 30-layer simulations are discussed in Sections 5.2.5 and 5.3.5. As a result of these tests, it was determined that the 10-layered simulation would be the better one to use for further work in this research. Table 5. 22 Pilot Scale Simulation for the Settling of P&P Primary Sludge Date: 2003/4/19 (10:00 am) Qin = 0.014 L/s Qw = 0.002 L/s Layer 10-layer Simulated Data (mg/L) 30-layer Simulated Data (mg/L) Measured Data (mg/L) 10-layer Difference (%) 30-layer Difference (%) Note Influent 182 1 41 41 57 -28.1 -28.1 Top layer 2 42 41 65 -35.4 -36.9 3 44 41 77 -42.9 -46.8 4 46 41 79 -41.8 -48.1 5 51 42 77 -33.8 -44.2 6 61 46 80 -23.8 -36.3 7 89 89 77 15.6 15.6 Feeding Layer 8 89 89 90 -1.1 -1.1 9 89 89 86 3.5 3.5 10 89 89 124 -28.2 -28.2 11 89 89 136 -34.6 -34.6 Sludge hopper 12 1,020 1,028 2,920 -65.1 -64.8 Sludge hopper Waste sludge 1,028 1,028 932 10.3 10.3 Calculated by mass balance SS (mg/L) 10 100 1,000 1 0 , 0 0 0 E 0 .4 1 \ \ - • 1 — — Measured Data Simulated Data (10 layers) Simulated Data (30 layers) Fig 5.41 Pilot Scale Simulation of the Settling of P&P Primary Sludge (Ln scale inSS) 5.2.2 Secondary Sludge of Pulp and Paper Wastewater The pilot tests of pulp and paper secondary sludge took place on February 24, 2003 when the mill was manufacturing normally. The operating conditions of the secondary clarifier #1 were: an influent flow rate of 470 L/s, a hydraulic surface overflow rate of 14 m/d, an influent SS of 3,535 mg/L, a solids loading rate of 49 kg/m2 • d, an effluent SS of 8 mg/L, an SS of RAS of 9,325 mg/L, with a flow rate of 188 L/s, a WAS consistency of 3.3% with a flow rate of 5 L/s and an SRTof 4.9 days. The parameters used for this pilot simulation were rh = 3.073 x 10"4, rp = 4.696 x 10"3 , V0-226, V0' = 141, Xmin = 7 (Table 5.21). The influent of the primary clarifier was pumped into the central feed column of the settling tank at an overflow rate about 15 m/d, and was measured at a pumping rate of 0.013 L/s , and the bottom sludge was pumped at 0.003 L/s to keep a constant sludge interface position. The simulated results are shown in Table 5.23 and Figure 5.42. There were differences of-43.8% and -50.0% for the 10-layer and 30-layer simulation at the top layer when compared with the measured data for the pulp and paper secondary sludge. The differences in the waste sludge, when compared with the measured data, were -33.4% and -29.0% for the 10 and 30-layer simulations, respectively. As shown in Figure 5.42, the 10-layer simulation would be the better one to choose for further simulations. Both Figures 5.41 and 5.42 show that the measured suspended solids in the testing tank increased with the depth of water and had a smooth curve of SS profile. However, the simulated results had a straight line just below the end of the central feed column, which means the SS kept the same between the feeding layer and bottom layer. The results of the model calculation indicated that in these layers, the suspended solids flux into the layer was equal to the flux out of the layer. Another factor contributing to this difference was the assumption of the Takacs Model that the SS concentration is constant in the same layer. In a real clarifier or settling tank, it is difficult to achieve this ideal situation. 5.2.3 Primary Sludge of Municipal Wastewater The test was done on April 13, 2003, a sunny day. The operating conditions of the full scale clarifier #1 were: an influent flow rate of 45 L/s, a hydraulic surface overflow rate of 20 m/d, an influent SS of 250 mg/L, an effluent SS of 60 mg/L and a waste sludge concentration of 922 mg/L, with a wasting rate of 10 L/s. Tab 5.23 Pilot Scale Simulation of the Settling of P&P Secondary Sludge Date: 2003/2/24 (1:30 pm) Qin = 0.013 L/s Qw = 0.003 L/s Layer 10-layer Simulated Data (mg/L) 30-layer Simulated Data (mg/L) Measured Data (mg/L) 10-layer Difference (%) 30-layer Difference (%) Note Influent 3,464 1 9 8 16 -43.8 -50.0 Top layer 2 11 11 25 -56.0 -56.0 3 14 12 23 -39.1 -47.8 4 18 14 24 -25.0 -41.7 5 28 20 41 -31.7 -51.2 6 64 98 65 -1.5 50.8 7 306 508 122 150.8 316.4 Feeding Layer 8 306 508 184 66.3 176.1 9 306 508 835 -63.4 -39.2 10 306 508 1,640 -81.3 -69.0 11 2,910 508 6,930 -58.0 -92.7 Sludge hopper 12 14,617 14,904 21,933 -33.4 -29.0 Sludge hopper Waste sludge 14,974 14,977 14,957 0.1 0.1 Calculated by mass balance — — Measured Data Simulated Data (10 layers) Fig 5.42 Pilot Scale Simulation of the Settling of P&P Secondary Sludge (Ln scale in SS) To simulate a clarifier in steady state, the clarifier influent was pumped at rate of 0.0185 L/s into the pilot settling tank, and the bottom sludge was pumped at 0.00158 L/s. The parameters used for this pilot simulation were rh = 1.617 x 10"4, rp = 2.286 x 10'3, V0 = 107 V0' = 65, Xmin = 22. The simulated results are shown in Table 5.24 and Figure 5.43. The differences were -32.8% (10-layer simulation) and -43.3% (30-layer simulation) at the top layer. Differences of -64.7% (10-layer simulation) and -61.8% (30-layer simulation) were found for waste sludge. In this simulation, both the 10-layer simulation and 30-layer simulation were close to the measured data. The 10-layer simulation was better than 30-layer of upper layers and the layers near the bottom simulation. Table 5. 24 Pilot Scale Simulation of the Settling of Municipal Primary Sludge Date: 2003/4/13 (8:30 am) Qin = 0.0185 L/s Qw = 0.00158 L/s Layer 10-layer Simulated Data (mg/L) 30-layer Simulated Data (mg/L) Measured Data (mg/L) 10-layer Difference (%) 30-layer Difference (%) Note Influent 152 1 45 38 67 -32.8 -43.3 Top layer 2 56 54 69 -18.8 -21.7 3 65 60 64 1.0 -6.8 4 74 66 64 15.6 3.1 5 84 73 62 35.5 17.7 6 98 85 120 -18.3 -29.2 7 118 122 118 0 3.4 Feeding Layer 8 118 122 126 -6.3 -3.2 9 118 122 123 -4.1 -0.8 10 118 122 110 7.3 10.9 11 118 122 222 -46.8 -45.0 Sludge hopper 12 1,269 1,371 3,590 -64.7 -61.8 Sludge hopper Waste sludge 1,298 1,486 1,062 22.2 39.9 Calculated by mass balance SS (mg/L) 10 100 1,000 10,000 0.2 \ 1 V V 1 > r 1—: i — — Measured Data Simulated Data (10 layers) Fig 5.43 Pilot Scale Simulation of the Settling of Municipal Primary Sludge (Ln scale in SS) 5.2.4 Secondary Sludge of Municipal Wastewater The test for the municipal secondary sludge was conducted on Feb 24, 2003. The weather was cloudy with a few light showers. The secondary clarifier #1 was operated as: influent flow rate 107 L/s, hydraulic surface overflow rate 22 m/d, influent SS 2066 mg/L, solids loading rate 45 kg/m2 • d, effluent SS 8 mg/L, SS of RAS 6,504 mg/L with a recycling rate 41 L/s and a wasting rate 5 L/s. The SRT was 1 day from Comox WWTP monitoring system. In order to provide an overflow rate of about 15 m/d to the settling tank, the clarifier influent to the pilot scale secondary clarifier was set at a rate of 0.0175 L/s, and the bottom sludge wasting rate was set at 0.00556 L/s. The model parameters used for this pilot simulation were rh = 8.148 x 10"4, rp = 7.888 x 10"3, V0 = 168, V0' = 97, Xmin = 4. The simulated results are shown in Table 5. 25 and Figure 5.44. The differences at the top layer were: -12.9% (10-layer simulation) and -33.3% (30-layer simulation). For the waste sludge, the differences were: -55.2% (10-layer simulation) and -36.1% (30-layer simulation) as compared with the measured data. Table 5. 25 Pilot Scale Simulation of the Settling of Municipal Secondary Sludge Date: 2003/4/15 (8:30 am) Qin = 0.0175 L/s Qw = 0.00556 L/s Layer 10-layered Simulated Data (mg/L) 30-layered Simulated Data (mg/L) Measured Data (mg/L) 10-layered Difference (%) 30-layered Difference (%) Note Influent 2,000 1 8 6 9 -12.9 -33.3 Top layer 2 11 8 12 -8.3 -33.3 3 15 9 15 0.0 -40.0 4 20 10 18 11.1 -44.4 5 35 13 32 94.7 -59.4 6 87 22 92 32.4 -25.8 7 416 417 240 73.3 73.8 Feeding Layer 8 416 417 1,300 -87.2 -67.9 9 1,641 417 3,460 -70.2 -87.9 10 3,704 417 4,833 -48.9 -91.4 11 4,834 1,430 7,867 -59.0 -81.8 Sludge hopper 12 5,954 6,262 9,800 -55.2 -36.1 Sludge hopper Waste sludge 6,278 6,083 6,276 0.04 3.1 Calculated by mass balance 0.00 0.10 0.20 0.30 1 0.40 • g D . U 0.50 Q 0.60 0.70 0.80 0.90 10 SS (mg/L) 100 1,000 10,000 — Measured Data — Simulated Data (30 layers) Simulated Data (10 layers) Fig 5.44 Pilot scale Simulation of the Settling of Municipal Secondary Sludge (Ln scale in SS) 5.2.5 Discussion The simulations of these four pilot scale tests did not produce good results due to the fact that the pilot settling tank was not ideal for clarification. First, the outlet for the effluent was a 1/2" pipe instead of an overflow launder. This caused a short circuiting current from the feeding column to the overflow pipe. This resulted in increasing SS in the effluent. Second, the cross-sectional area of the feed column was relatively small. Kennedy (1994) determined that the best conditions for the Takacs model occurred when the cross-sectional area of the feed column was 25% of the surface area of the clarifier. The cross-sectional area of the feed column of the pilot scale tests was only 2.8% of the surface area, so the heavier solids of the influent were injected into the sludge hopper and the lighter particles flowed upwards the moment of the water came out of the feed column. That is, the testing column could not provide constant concentrations near the feeding layer, which does not agree with the assumptions of the one-dimensional models as mentioned in Section 2.4. Moreover, there was no stirring or scraping of the pilot settler. At the beginning of the test, the underflow sludge was pumped out uniformly. However after 2 hours of pumping, the settled solids were "bridged". More water was pumped out of the liquor than was solids material. The solids accumulated in the sludge hopper increased the SS concentrations at the lower layers. For the reasons described above, the concentrations of SS in the effluent and at the bottom layer were much higher than predicted by the model. From the experiments, the 10-layer simulations seemed to perform better than the 30-layer simulations. Due to the phenomena caused by the less-than-ideal settling tank, the results were not sufficient to suggest using either the 10 or the 30-layer simulation. However, both of them were used to model the full scale clarifiers in the next section. 5.3 Field Sampling Tests (Full Scale) In order to understand how the simulations work on full scale clarifiers and to know the SS profiles in different sections of the clarifiers, full scale tests followed the pilot tests for primary and secondary clarifiers of both pulp and paper and municipal wastewater treatment plants. 5.3.1 Primary Clarifier of Pulp and Paper Wastewater Three sections of the clarifier were chosen to do the SS measure: Section #1 was located 30 cm outside the flocculation well, Section #2 was in the middle of the clarifier, and Section #3 was 30 cm adjacent to the overflow launder. Samples were taken every 45 cm from the top to the bottom. Due to the slope at the bottom, there were 13 layers for Section #1, 12 layers for Section #2 and 11 for Section #3. The simulation used the data of April 20, 2003. The operation of the treatment plant was normal. The influent flow rate was 748 L/s with a hydraulic surface overflow rate of 26 m/d; influent SS was 492 mg/L and the sludge wasting rate was 5.4 L/s. The parameters used for this pilot simulation were the values list in Table 5.21. (rh = 2.728 x 10"4, rp = 4.864 x 10'3 , V0 = 133, F 0 ' = 85,Xm,„ = 41). The result was 2.2% (10-layer simulation) and -8.9% (30-layer simulation) differences between simulated and measured data for the SS of the top layer of Section #1, and -8.4% (10-layer simulation) and -1.9% (30-layer simulation) for the SS of the bottom layer (see Table 5.26 and Figure 5.45). If the simulation results for the top layer of Section #3, the SS concentration of which was assumed to be the same as the effluent (35 mg/L) are examined, then the 30-layer simulation (41 mg/L, with a 17.1% difference) was better than the 10-layer simulation (46 mg/L, with a 30.6% difference). Although the simulated results of the medial point layers below the feeding point were far from agreement with the measured data, the model was good for modeling the effluent and waste sludge solids of the primary clarifier of this pulp and paper mill. Table 5. 26 Full Scale Simulation of the P&P Primary Clarifier Date: 2003/4/20 (10:00 am) Qin = 748L/s Qw = 5.4L/s Layer 10 layers Simulated Data 30 layers Simulated Data Measured Data (mg/L) 10 layers Difference % 30 layers Difference % Note (mg/L) (mg/L) Section #1 Section #2 Section #3 with Section #1 with Section #1 influent 492 1 46 41 45 49 35 2.2 -8.9 Top layer 2 52 43 43 40 36 20.9 0.0 3 59 44 43 45 35 37.2 2.3 4 72 47 40 32 30 80.0 17.5 5 99 53 45 33 28 120.0 17.8 6 170 75 58 38 22 193.1 29.3 Feeding layer 7 170 171 276 40 27 -38.4 -38.0 8 170 171 304 37 30 -44.1 -43.8 9 170 171 308 45 35 -44.8 -44.5 10 13,417 7,042 336 64 26 3893.2 1995.8 11 25,646 28,257 28,500 • 25,200 35,900 -10.0 -0.9 12 30,573 30,997 29,900 36,100 2.3 3.7 13 39,472 42,289 43,100 -8.4 -1.9 Bottom layer Waste Sludge 61,305 62,513 63,374 -3.3 -1.4 Calculated by mass balance 10 0.00 .00 2.00 "S 3.00 4.00 5.00 100 SS (mg/L) 1,000 10,000 100,000 • Measured Data (Section #1) - Measured Data (Section #3) • Simulated Data (30-layers) Measured Data (Section #2) — — Simulated Data (10-lay ers) Figure 5.45 Full Scale Simulation of P&P Primary Clarifier (Ln scale in S S) 5.3.2 Secondary Clarifier of Pulp and Paper Wastewater The locations for the sampling of full scale pulp and paper secondary Clarifier #1 were the same as those for the primary clarifier. The test was also done on April 20, 2003. The operating conditions were: an influent flow rate of 636 L/s, a hydraulic surface overflow rate of 19 m/d, an influent SS of about 4,100 mg/L, a solids loading rate of 77 kg/m2 • d, RAS 12,250 mg/L at a 195 L/s flow rate, a WAS flow rate of 3.4 L/s and an SRT of 7.7 days. The parameters used for this pilot simulation were rh = 3.073 x 10-4, rp = 4.696 x 10 - 3, V0 = 226, V0' = 141, Xmin = 7 (Table 5.21). The differences between simulated and measured SS in the top layer of Section #1 were 7.7% (10-layer simulation) and -38.5% (30-layer simulation); the differences at the bottom layer were 13.3 (10-layer simulation) and 15.8% (30-layer simulation), and for RAS, they were 27.3% (10-layer simulation) and -11.1% (30-layer simulation) (see Table 5.27 and Figure 5.46). When comparing the simulation results to the top layer of Section #3(11 mg/L), the 10-layer simulation (12 mg/L) had a 4 mg/L improvement over the 30-layer simulation, which was 8 mg/L. The results indicated that the Takacs model performed well at simulating the pulp and paper secondary clarifier both at the top and bottom layers, but not at the intermediate depths. 5.3.3 Primary Clarifier of Municipal Wastewater The primary clarifier #1 was used to do the simulation. The locations of the three sections of the clarifier were the same as those of the pulp and paper clarifiers: Section #1 was located 30 cm from the inlet ports; Section #2 was in the middle of the clarifier; and Section #3 was 30 cm adjacent to the effluent launder. The clarifier was divided into 30 cm layers, resulting in 11 layers for Section #1,10 layers for Section #2 and 9 layers for Section #3. The operating conditions on the sampling day, April 10, 2003 were: an influent flow rate of 58 L/s with a hydraulic surface overflow rate of 27 m/d, an influent SS of 134 mg/L and a waste sludge rate of 9.5 L/s. The parameters used for this pilot simulation were rh = 1.617 x 10"4, rp = 2.286 x 10"3, V0 = 107 VQ' = 65, Xmin = 22 (Table 5.21). Table 5.27 Full scale Simulation of the P&P Secondary Clarifier Date: 2003/4/20 (1:00 pm) Qin= 636 L/s Qr=195 L/s Qw=3.4 L/s Layer 10 layers Simulated Data 30 layers Simulated Data Measured Data (mg/L) 10 layers Difference % 30 layers Difference % Note (mg/L) (mg/L) Section #1 Section #2 Section #3 with Section #1 with Section #1 influent 3,980 1 12 8 13 13 11 7.7 -38.5 Top layer 2 17 11 15 12 8 13.3 -26.7 3 24 12 14 13 9 71.4 -14.3 4 40 14 14 19 9 185.7 0 5 98 20 18 31 8 444.4 11.1 6 507 98 16 33 10 3,068.8 512.5 Feeding layer 7 507 508 15 33 14 3,280.0 3,286.7 8 507 508 14 60 14 3,521.4 3,528.6 9 507 508 14 60 11 3,521.4 3,528.6 10 507 508 534 152 289 -5.1 -4.9 11 6,192 508 650 12,480 17,780 852.6 -21.8 12 15,589 508 8,776 25,433 226.4 -89.4 13 27,657 28,287 24,420 13.3 15.8 Bottom layer RAS 15,589* 10,887** 12,250 27.3 -11.1 * RAS of the 10-layer simulation was equal to the solids concentration at the 12 layer. * * RAS was withdrawn at layer 36 t h and 38th. For the 30-layer simulation, it was equal to the average concentration of the 36th and the 38th layers, which were 3,431 and 18,342 mg/L. 0 . 0 0 .00 2 . 0 0 t 3 0 0 a 4 . 0 0 5 .00 SS (mg/L) 1 0 0 1 ,000 1 0 , 0 0 0 1 0 0 , 0 0 0 ' % — • — — 1. L 1 Measured Data (Section #1) •Measured Data (Section #3) •Simulated Data (30-layers) • Measured Data (Section #2) Simulated Data (10-layers) Fig 5.46 Full Scale Simulation of P&P Secondary Clarifier (Ln scale inSS) In this case, the differences between the simulated and measured SS of the top layer of Section #1 were 21.8% (10-layer simulation) and -16.4% (30-layer simulation); for the bottom layer, they were -93.9% (10-layer simulation), and -98.6% (30-layer simulation) (see Table 5.28 and Figure 5.47). When compared with the top layer of Section #3, the differences were 139.9% (10-layer simulation) and 64.3% (30-layer simulation). In this clarifier, the influent comes into the rectangular clarifiers at one side near the water surface of the basin. This may make it inappropriate to apply one-dimensional models to this particular clarifier. Table 5.28 Full Scale Simulation of the Municipal Primary Clarifier Date: 2003/4/10 (8:30 am) Qin= 58 L/s Qw= 9.5 L/s Layer 10 layers Simulated Data 30 layers Simulated Data Measured Data (mg/L) 10 layers Difference % 30 layers Difference % Note (mg/L) (mg/L) Section #1 Section #2 Section #3 (with Section #1) (with Section #1) Influent 134 1 67 46 55 32 28 21.8 -16.4 Top layer 2 98 91 71 38 35 38.0 28.2 3 98 109 82 35 30 19.5 32.9 4 98 109 94 35 46 4.3 16.0 5 98 109 76 51 45 28.9 43.4 6 98 109 76 50 51 28.9 43.4 Feeding layer 7 98 109 81 63 57 21.0 34.6 8 98 109 79 69 420 24.1 38.0 9 98 109 75 1340 520 30.7 45.3 10 98 109 81 2100 21.0 34.6 11 468 571 7533 -93.9 -98.6 Bottom layer Waste Sludge 476 583 537 -12.1 6.6 Calculated by mass balance 10 1 0 0 SS(mg/L) 1 ,000 1 0 , 0 0 0 1 i ' r i • i' • 1 ' rj: S ! — \ l i_ i' i ! i 1 1 i Y \ i Li 1 • ' i i i i i 1 I i i Measured Data (Section #1) Measured Data (Section #2) Measured Data (Section #3) — — Simulated Data (10-layers) Fig 5.47 Full Scale Simulation of the Municipal Primary Clarifier (Ln scale in SS) 5.3.4 Secondary Clarifier of Municipal Wastewater For the simulation of the municipal secondary clarifiers, clarifier #1 was chosen. The locations of the three sections were as follows: Section #1 was located 30 cm from the inlet channel (near the rim of the clarifier), Section #2 was in the medial of the clarifier, and Section #3 was 30 cm adjacent to the center of the clarifier. The clarifier was also conceptually divided into 30 cm deep layers. Since the bottom of the clarifier is flat, there were 11 layers for all sections. On the day of the simulation, April 9, the operating conditions of the secondary clarifier were: an influent flow rate of 76.5 L/s with a hydraulic surface overflow rate of 16 m/d, influent SS at 2,480 mg/L with a solids loading rate of 39 kg/m2 • d, the RAS and WAS were around 5,000 mg/L, at a recycling activated sludge flow rate of 40.5 L/s, and the wasting rate of WAS was 3.0 L/s. The model parameters used for this pilot simulation were rh = 8.148 x 10"4, rp = 7.888 x 10"3, V0 = 168, F 0 ' = 97,Xm,„ = 4(Table5.21). The secondary clarifiers at the Comox WWTP were fed peripherally. The differences are 1 mg/L for both the 10- and 30-layer simulations at the top layer of Section #1 (effluent end), and -62.1% of bottom sludge for both the 10- and 30-layered simulations (see Table 5.29 and Figure 5.48). Though the simulation results were not close to the measured data for the bottom layer, they were close to the real RAS and WAS. Therefore, the model could be applied to simulate this clarifier as well. Table 5.29 Full Scale Simulation of the Municipal Secondary Clarifier Date: 2003/4/9 (10:30 am) Qin= 76.5 L/s Qr = 40.5 L/s Qw=3.0L/s Layer 10 layers Simulated Data 30 layers Simulated Data Measured Data (mg/L) 10 layers Difference % 30 layers Difference % Note (mg/L) (mg/L) Section #1 Section #2 Section #3 with Section #1 with Section #1 influent 2,480 1 6 4 5 4 5 20.0 -20.0 Top layer 2 8 5 6 13 6 33.3 -16.7 3 11 5 6 13 8 83.3 -16.7 4 19 6 7 14 34 171.4 -14.3 5 52 9 10 26 53 420 -10.0 6 369 52 16 80 65 2206.3 225.0 Feeding layer 7 369 369 92 259 115 301.1 301.1 8 369 369 550 910 200 -32.9 -32.9 9 369 369 920 3,800 645 -59.9 -59.9 10 1,827 369 10,880 3,633 4,300 -83.2 -96.6 11 4,260 4,260 11.233 6,600 7,967 -62.1 v -62.1 Bottom layer RAS/WAS 4,356 4,360 4,358 -0.001 +0.001 Calculated by mass balance SS (mg/L) 1 10 '100 1,000 10,000 100,000 — sS -1 1 \ 1 1 1 Is ^ — Il 1 1 X ! J ~~~~ M ~'i i 1 ! " ~ r " i i : i : " " " " — • i 1 i l i 1 — - ^ L i - n i Measured Data (Section #1) Measured Data (Section #2) Measured Data (Section #3) — — Simulated Data (10-layers) Simulated Data (30-layers) Figure 5.48 Full Scale Simulation of the Municipal Secondary Clarifier (Ln scale in SS) 5.3.5 Discussion Mass Balance of the Model One of the disadvantages of the flux theory models is that they do not incorporate the concept of a mass balance, due to the fact that the solids concentration is predicted solely by the net flux, which is determined by a modeled gravity settling velocity. The SS of the waste sludge calculated by mass balances around the clarifier's influent and effluent (Equation 5.1 and 5.2) is shown in Table 5.30, where it is compared with the simulation data. The calculated mass-balance data of column (2) and (5) in Table 5.30 are from Table 5.26-5.29. The data of column (3) are calculated by [(l)-(2)]/(2), and column (3) = [(4)-(5)]/(5). Although the model functioned very well in terms of mass balance for the municipal clarifiers, there was a more than 25% error for the pulp and paper primary sludge and WAS. Table 5.30 also shows the 30-layer simulation resulted in better mass balances than the 10-layered simulation. The mass balances equation for primary clarifiers: Q i n x X i n = Q e x X e + Q u x X b (5.1) The mass balances equation for primary clarifiers: Qin X Xj„ = Q e X X e + Q u X Xb = Q e X X e + Q R A S X X R A S + QwAS X X W A S (5.2) where Q j n = Influent flow rate Influent SS Qe = Effluent flow rate Xe Effluent SS Q u = Under flow rate xb = SS of under flow Q R A S = Flow rate of return activated sludge X R A S •= SS of return activated sludge QwAS = Flow of waste sludge XwAS = SS of waste sludge Watts et al. (1996) did an analysis using 10, 20 and 50-layer simulations of the Takacs model. It was concluded that the simulations with more layers produced poor results in the layers at intermediate depths of the clarifier. Thus, they suggested a dispersion coefficient to modify the model in order to get better results in the intermediate layers. In the same year, Jeppsson and Diehl (1996) compared the dynamic simulation of the Takacs model with the Diehl model and suggested: "at least 30 layers are recommended for reliable results under normal operation conditions." From the full scale analyses of the present study, it is agreed that the 30-layer simulation is superior to the 10-layer simulation for the top and bottom layers. Table 5.30 The Evaluation of Mass Balance for the Takacs Model (1) (2) (3) (4) (5) (6) Item 10-layered Simulated (mg/L) 10-layers Calculated by Mass-Balance (mg/L) 10-layers Difference (%) 30-layered Simulated (mg/L) 30 layers Calculated by Mass-Balance (mg/L) 30-layers Difference (%) P&P Primary underflow sludge 39,825 61,305 -35.0 42,641 62,513 -31.8 P&P RAS 15,866* 15,866 0 10,857** 12,250** -11.4 P&P WAS 27,746 40,247*** -31.1 29,590 40,247*** -26.5 Municipal Primary underflow sludge 416 471 -11.7 564 573 2.5 Municipal Secondary underflow sludge 4,260 4356 -35.0 4,360 4,260 0.2 Note: (3)=[(l)-(2)]/(2); (6)=[(4)-(5)]/(5). Data of column (2) and (4) are from Table 5.26-5.29. *: RAS was equal to the 10th layer. : RAS was withdrawn at layer 36 th and 38 th, the concentration was taken the average of these two layers. *: WAS was as the same value as measured data, 40,247 mg/L. Sensitivity Analysis In the Literature Review (Section 2.5), it was noted that Kennedy (1994) summarized the sensitivities of the Takacs Model to parameter variations. In the following section, the impact of the influent SS, influent flow rate and underflow rate on the effluent and waste solids will be discussed. Only the pulp and paper primary and secondary clarifiers were chosen for analysis. The parameters and other flow conditions used for this sensitivity analysis were the same as those used in the full scale simulation, which were listed in Table 5.21. The SS and flow rate of influent were varied to understand how they affected the SS of effluent, bottom sludge, returning activated sludge and waste sludge. Both the 10 and 30-layer simulations were tested. The results shown in Table 5.31 and Table 5.32 indicate that both the effluent SS and the SS of the bottom layer are not sensitive to the value of the influent SS, and the solids concentrations of the waste sludge at higher influent SS were unreasonably high. However, the waste sludge concentration calculated by mass balance is highly sensitive to an increase or decrease of the influent SS. Table 5.33 and Table 5.34 reveal that the effluent SS and the SS at the bottom layer are sensitive to changes in influent flow, when the influent flow exceeds a certain rate. The waste sludge was highly sensitive to changes in the influent flow rate, but in higher influent flow rates, the solids concentrations of waste sludge were unmoral high. Table 5.35 indicates that waste sludge was sensitive to changes in the underflow rate, but the effluent SS was not. Both 10 and 30-layer simulations are relatively insensitive to variations in influent SS, the influent flow rate and the underflow rate when simulating these clarifiers. The 30-layer simulation is less sensitive than the 10-layer one. Although the 30-layer simulation is more accurate at the top and bottom layers, it is not sensitive to variations in input data values. Also, the 30-layer simulation is not good at the layers right above the very high solids concentration layer. The impact of clarifier depth also examined. The depth was changed to double and one-third the actual values of the pulp and paper clarifiers. The outcomes listed in Table 5.36 point out there was almost no change of SS at the top and bottom layers when the depth was changed. Furthermore, from Table 5.31 and Table 5.33, we can see that when the influent SS is at about 500 mg/L and the flow rate at about 750 L/s in the primary clarifier, the simulation results can fit the real operation situation of the plant that the SS of bottom layer was 43,100 mg/L and the waste sludge was about 63,400 mg/L (see Table 5.26). For the secondary clarifier, from Table 5.31 and Table 5.33, when the influent SS (MLSS) was at about 4,000 mg/L and the flow rate was about 650 L/s, the simulated results were similar to the existing operation conditions that the SS of bottom layer was 24,420 mg/L, the RAS was 12,250 mg/L (see Table 5.27) and the WAS was about 3.3 % (see Section 5.2.3). This also showed that the pulp and paper wastewater treatment plant was well designed and operated. Table 5.31 Sensitivity of the Simulation of the P&P Primary Clarifier SSj„ vs SSefr and Waste Sludge Influent SS SS of top layer SS of bottom layer Waste Sludge* (mg/L) (mg/L) (mg/L) (mg/L) 10-layered simulation 1,000 47 39,842 132,055 750 46 39,834 97,563 500 46 39,825 62,933 100 43 39,759 7,938 30-layered simulation 1,000 41 42,641 143,894 500 41 42,641 62,677 100 41 42,641 8,235 *:The data were calculated by mass balances. Table 5.32 Sensitivity of the Simulation of the P&P Secondary Clarifier SS i n vs SSeff, RAS and Waste Sludge Influent SS SS of top layer SS of bottom layer RAS Waste sludge* (mg/L) (mg/L) (mg/L) (mg/L) (mg/L) 10-layered simulation 6,000 15 28,116 17,661 107,641 4,000 14 27,825 12,369 37,022 2,000 14 27,274 5,905 33,647 500 13 26,331 1,099 28,697 30-layered simulation 6,000 9 28,465 18,671 50,358 4,000 9 28,326 12,486 30,968 2,000 9 28,204 6,008 28,383 500 9 27,014 1,362 14,256 *:The data were calculated by mass balances. Table 5.33 Sensitivity of the Simulation of the P&P Primary Clarifier Qi„ vs SSeff and Waste Sludge Influent Flow (L/s) S S of top layer (mg/L) S S of bottom layer (mg/L) Waste Sludge* (mg/L) 10-layered simulation 1,500 63 39,839 121,452 750 46 39,825 62,933 470 42 39,819 39,905 200 41 39,796 19,745 30-layered simulation 1,500 48 42,641 123,378 750 41 42,641 62,677 470 41 42,641 39,295 200 41 42,641 19,745 *:The data were calculated by mass balances. Table 5.34 Sensitivity of the Simulation of the P&P Secondary Clarifier Qin vs SSeff, RAS and Waste Sludge Influent Flow SS of top layer SS of bottom layer RAS Waste sludge* (L/s) (mg/L) (mg/L) (mg/L) (mg/L) 10-layered simulation 1,200 29 28,116 17,661 213,841 650 14 27,825 10,748 50,829 400 9 27,410 6,572 34,308 200 7 26,706 3,057 30,551 30-layered simulation 1,200 16 28,479 19,239 127,167 650 9 28,290 10,748 24,480 400 8 28,046 6,674 23,517 200 7 27,417 3,383 11,854 *:The data were calculated by mass balances. Table 5.35 Sensitivity of the Simulation of P&P Clarifiers Qw vs SSeff and Bottom Sludge Qw (L/s) S S of top layer (mg/L) SS of bottom layer (mg/L) 10 layers 30 layers 10 layers 30 layers Primary clarifier 50 45 41 20,305 24,361 20 45 41 32,694 35,555 5.4 46 41 39,825 42,641 2 46 41 41,946 43,645 1 46 41 42,629 43,775 Secondary clarifier 30 14 9 22,166 22,696 15 14 9 24,424 25,459 3.4 14 9 27,825 28,290 1.5 9 9 28,568 31,135 0.5 9 9 29,047 32,193 Table 5.36 Sensitivity of Simulation of P&P Clarifiers Depth vs SSeff and Bottom Sludge Depth (m) S S of top layer (mg/L) S S of bottom layer (mg/L) 10 layers 30 layers 10 layers 30 layers Primary clarifier 1.95 46 41 38,226 39,613 5.85 46 41 39,825 42,641 11.7 46 41 39,920 43,208 Secondary clarifier 1.95 14 9 29,847 29,110 5.85 14 9 29,590 27,845 11.7 14 9 28,876 26,685 Sectional Area of the Feeding Well As discussed in Section 5.2.5, the best hydraulic arrangement for the Takacs model is a 25% cross-sectional area of the feeding well (Kennedy, 1994). The cross-sectional area of the total full scale P&P clarifiers is 9%, and, for the municipal secondary clarifier, it is 8%. These percentages cause the solids to not distribute uniformly in the same layer of the medial part of the clarifiers. Thus, the simulated results were not optimal at the medial layers of the clarifiers. The Compression of Sludge with Time The simulated solids concentrations at each layer generated by the Takacs model were calculated by the solids change at each layer, which means it only relates to the depth of the clarifier, but is independent of the clarifier sludge retention time, or hydraulic retention time (HRT). In Section 5.3.1 and 5.3.2, it was noted that the waste sludge in the clarifiers at the Port Alberni P&P mill was much higher than the solids concentrations yielded by the model. This is because the sludge hopper (8.5 m in diameter and 1.5 m deep) at the center of the clarifier offers additional retention time for sludge thickening. In the primary clarifier, the sludge retention time in the hopper was calculated to be 8 hrs, and for secondary clarifier #1, the sludge retention time was 12 hrs. This retention time produces higher solids concentrations at these levels than those of the solids on the clarifier floor. In order to understand why the solids concentration increased with thickening time, some other compression tests were done. Primary and secondary sludges from the pulp and paper mill and the Comox WWTP were poured into the settling test cylinder (Figure 4.9) and stirred. The sludge was sampled every hour for 6 hours from the drainage valve on the bottom of the cylinder to do SS or consistency analysis. The results are listed in Table 5.37. There are natural log relations between suspended solids concentration and thickening time, as shown in Figures 5.49. The R square numbers of the pulp and paper sludge and municipal secondary sludge were from 0.94 to 0.97, only municipal primary sludge was 0.83, which is lower than 0.90. Although these results were not strong enough to express how the solids thicken with time, the results led roughly to the conclusion that solids concentrations have a natural log relation with thickening time. Table 5.37 Concentrations of P&P Primary Sludge Thickening with Time Compression Time (Hr) Concentration (mg/L) P&P Primary P&P Secondary Municipal Primary Municipal Secondary 0 30,700 29,000 11,600 30,700 1 39,300 32,300 28,700 39,300 2 41,500 36,100 34,600 41,500 3 44,700 40,600 35,900 44,700 4 46,800 41,400 36,000 46,800 5 47,900 41,500 36,100 47,900 6 51,100 41,900 36,400 51,100 B c: o •f3 § c o U 55000 50000 45000 40000 35000 30000 25000 20000 15000 10000 5000 y = 6364.2Ln(x) + 38246 R2 = 0.9448 * y = 5692.6Ln(x) -~" "* R2 = 0.936 . - A . - - - - - A " " " " ~ A * *• * y=4060.9Ln(x) + 30183 • g -• • " y = 3606Ln(x) + 8998.6 K = u.ybSi 3 4 Compression Time (Hr) X P&P Primary Sludge A Municipal Primary Sludge Ln(P&P Primary Sludge) - - • Ln(Municipal Primary Sludge) • P&P Secondary Sludge • Municipal Secondary Sludge Ln(P&P Secondary Sludge) Ln(Municipal Secondary Sludge) 32736 Figure 5.49 The Concentrations of Underflow Sludge Increasing with Time 5.4 Precision of Solids Analysis Three sets of low (lower than 100 mg/L), medium (between 100 and 2,000 mg/L) and high concentration (higher than 2,000 to 30,000 mg/L) SS samples and one set of solids consistency (higher than 30,000 mg/L) samples were measured in duplicate in order to determine the precision of the analysis. Each set included three replicates. The sample volumes for filtration and the analysis results are shown in Table 4.37. The coefficients of variation for low, medium, and high SS, and solids consistency were 8.2%, 3.5%, 3.2% and 4.5%. The coefficients of variation of the SS analysis were lower than 5%, except for SS lower than 100 mg/L. This infers that the measured SS in present study had relatively high reliability. For the lower SS concentrations, the lower concentrations produced larger deviations. This might also cause a larger error and difference between the simulation data and laboratory data on the lower concentration water such as effluent. Table 5.38 Precision of Solids Analysis Sample Filtered volume (ml) SS (mg/L) Sample Filtered volume(ml) SS (mgT.) L o w concentration # 1 50 34 High concentration # 1 5 20,500 #2 50 38 #2 5 18,900 #3 50 40 #3 5 19,260 Mean value 37 Mean value 19,553 Stand deviation 3 Stand deviation 631 Coefficient of variation 8.2% Coefficient of variation 3.2% Medium concentration #1 25 316 Consistency # 1 4.29% #2 25 348 #2 4.46% #3 25 336 #3 4.84% Mean value 333 Mean value 4.53% Stand deviation 12 Stand deviation 0.21 Coefficient of variation 3.5% Coefficient of variation 4.5% 6. Conclusions and Recommendations 6.1 Conclusions The major purpose of this research was to calibrate the parameters of the Takacs model for pulp and paper sludge and identify how the Takacs model works on pulp and paper wastewater clarifiers in comparison with municipal wastewater. Therefore the results show that the simulations of both pulp and paper primary and secondary clarifiers can be achieved. Identification of Parameters The calibrated parameters of pulp and paper sludge are summarized in Table 6.1 along with those for municipal sludge. Both the pulp and paper primary and secondary sludge exhibited similar settling characteristics to those of municipal sludge. The model parameters were based on the normal operating conditions of the two treatment plants. Because sludge may have different characteristics in different situations, it needs to be calibrated in each individual situation. Table 6.1 Summary of the Parameters of the Takacs Model for P&P and Municipal Sludge Item A-A(XK)-4) rp (xlO"3) Vo Vo Y • Pulp and paper sludge Primary sludge 2.728±0.610 4.864±1.441 133±25 85±11 41 Secondary sludge 3.073±0.469 4.696±0.703 226±13 141±8 7 Municipal sludge Primary sludge 1.617±0.360 2.286±0.462 107±12 65±7 22 Secondary sludge 8.148±0.660 7.888±0.589 168±29 97±15 4 Note: The values include one standard deviation. The measured maximum settling velocities of the primary and secondary sludge of pulp and paper and municipal wastewater were higher than the maximum settling velocities (V0') of the Takacs model. This contributes to uncertainty of the assumption behind Region of the Takacs Model (see Section 2.3), that as the solids concentrations increase the average settling velocity reaches a maximum value V0 '. Pulp and Paper Secondary Sludge vs. Municipal Secondary Sludge A summary of the parameter values for the Takacs model for pulp and paper and municipal sludge are shown in Table 7.1. The values of the theoretical maximum settling velocity (V0) for both pulp and paper primary and secondary sludge are similar to the values of municipal sludge, respectively. However, the secondary sludge has better settling ability than primary sludge for both pulp and paper wastewater and municipal wastewater in this case. For higher SS concentrations, although the value of r/, and V0 ' of the pulp and paper primary sludge was larger, as compared with the rh and V0' of municipal primary sludge, there is no evidence to illustrate that the zone settling velocities of pulp and paper primary sludge are faster than those of municipal primary sludge; the smaller rh of the pulp and paper secondary sludge comparing with the r/> of municipal secondary sludge, indicate a faster zone settling velocities of pulp and paper activated sludge than those of municipal activated sludge. For lower concentrations, the larger rp represents a better settling ability. The settling ability of pulp and paper primary sludge is better than that of municipal primary sludge, while the municipal secondary sludge has better settling ability. During the hindered settling tests, the zone settling began at concentrations of about 1,500 mg/L and 1,000 mg/L for pulp and paper and municipal secondary sludge, respectively. Pulp and paper secondary sludge needed a higher solids concentration to form floes. The non-settleable SS finding for pulp and paper secondary clarifier effluent was quite similar to that of municipal wastewater. However, the primary clarifier effluent of pulp and paper wastewater was higher than that in municipal secondary clarifier. Although the model can roughly simulate the sludge settling in the clarifiers, the real situations of the clarifiers were more complicated and had more variation of effluent SS and waste sludge than the simulated results. Generally, the amount of simulated effluent SS was higher than in the measured effluent. This may be due to aspects of the specific design of the clarifiers, such as the central flocculation well of the pulp and paper clarifiers and the peripheral feeding channel of the municipal secondary clarifier. The rectangular settling tank of this study required 2 or 3-dimensional modeling to get better simulations. The waste sludge of the pulp and paper primary and secondary clarifiers was about 1% higher in consistency (10,000 mg/L) than in the simulated result because the sludge hoppers provide a long detention time for the sludge to compress to a thicker concentration. Sensitivities of the Model In this study, it was also found that the Takacs model was sensitive to variations in influent flow rates, but was not sensitive to influent SS or underflow rates. The 30-layer model gave better simulation results of effluent SS, underflow SS and mass balance than the 10-layer model. On the other hand, however, the 30-layer simulation was much less sensitive than the 10-layer simulation on effluent and underflow SS. Thus, the 10-layer simulation is recommended for getting a better performance from this model. 6.2 Recommendations for Further Work COD and BOD Conversion in the Clarifiers The Takacs model only considers the suspended solids in the clarifier, but does not consider any biological reactions occurring in the clarifiers. In order to combine the Takacs model with other activated sludge treatment models, the variables in the conversions of COD and BOD have to be studied and added to the model. Bottom Sludge Compression with Time In this study, the bottom solids concentrations were found to have an unspecified natural log relation with thickening time. The specifics of how the settled sludge thickens with time will require more detailed research in order to propose an optimal model for waste sludge treatment. 7. Engineering Significance Application of the calibrated Takacs model, can be of assistance in the design of new clarifiers. It can also be beneficial in operation assistance and risk management assessment for wastewater treatment plants. In a new clarifier design, the model helps to assess the surface area and the depth of the clarifier and the underflow rate of the clarifier, when the flow rate, SS of the clarifier influent and effluent criteria were determined. It also provides the SS concentration and COD - if the compostion of the COD is investigated - of clarifier effluent for bioreactor design and concentration of underflow sludge for sludge handling. When applying the model to an existing clarifier, it predicts the SS of the effluent and underflow sludge, and provides information to operators to control the underflow pumping rate or to bypass the influent to equivalent tanks, or other approaches to reduce the risk of excess discharge of SS. The model also offers diagnosis of the clarifier to evaluate the function of the clarifier. If the measured SS of the effluent is excessively higher than the simulation, it might be caused by the varied characteristics of the influent or the problems of the clarifier, such as short-recruiting currents or dead space, etc. The results of present research could be direct reference of operation to the Port Alberni division of NorskeCanada Inc. and the Comox Valley Water Pollution Control Center. Also it is reference to other pulp and paper mills, municipal wastewater treatment plants and engineering consultants as design, operation and diagnosis. 8. References Abdel-Gawad, S.W. and McCorquodale, J.A. (1984) Strip Integral Method Applied to Settling Tanks. J. Hydraulic Div., ASCE, Vol. 110, No. 1, pp. 1-17. Adams, E.W. and Rodi, W. (1990) Modeling Flow and Mixing in Sedimentation Tanks. J. Hydraulic Engng, ASCE, Vol. 116, No. 7, pp. 895-913. Anderson, H.M. (1981) A Dynamic Simulation Model for Wastewater Renovation System. Ph.D. thesis, Wayne State University, Detroit, Michigan, USA. Anderson, N.E. (1945) Design of Settling Tanks for Activated Sludge. Sewage WksJ., Vol.17, No. 1, pp. 50-63. Busby, J.B. (1973) Dynamic Modeling and Control Strategies for the Activated Sludge Process. Ph.D. Dissertation, Clemson University, Clemson, S.C. Camp, T.R. (1936) A Study of the Rational Design of Settling Tanks. Sewage Works J., Vol. 8, No.5, pp. 742-758. Camp, T.R. (1946) Sedimentation and Design of Settling Tanks. Transactions, ASCE, Vol. I l l , No. 2258, pp.895-958. Celik, I., Rodi, W. and Stamou, A. (1987) Prediction of Hydrodynamic Characteristics of Rectangular Settling Tanks. In Turbulence Measurements and Flow Modeling, Part : Computational Fluid Dynamics, pp. 641-651, Hemisphere Publishing Corp., New York, USA. Cho, S.H., Colin, F., Sardin, M . and Prost, C. (1993) Settling Velocity Model of Activated Sludge. Water Res., Vol. 27, pp. 1237-1242. Dick, R.I. (1970) Role of Activated Sludge in Final Settling Tanks. J. San. San. Div, ASCE, Vol. 32, No. 10, pp. 2951-2958. Dick, R.I. and Young, K.W. (1972) Analysis of Thickener Performance of Final Setting Tanks. Paper presented at 27 Purdue Industrial Waste Conference, Lafayette, Indiana. Dupont, R. and Dahl, C. (1995) A One-dimensional Model for a Secondary Settling Tank Including Density Current and Short-circuiting. Water Sci. Tech., Vol. 31, No. 2, pp. 214-224. Eckenfelder, W.W. and Melbinger, N . (1957) Settling and Compaction Characteristics of Biological Sludge. Sewage and Industrial Wastes, Vol. 29, No. 10, pp. 1114-1122. Ekama, G.A., Pitman, A.R., Smollen, M . and Marais, G.v.R. (1984) Secondary Settling Tanks. Theory, Design and Operation of Nutrient Removal Activated Sludge Process, pp. 8-1-8-14, Pertoria, South Africa. Ekama, G. A., Barnard, J.L., Gunthert,F. W., Krebs, P., McCorquodale, J. A., Parker, D. S. and Wahlberg, E. J. (1997) Determinants of Settling Behavior of Biological Sludge. Secondary Settling Tanks: Theory, Modeling, Design and Operation, pp. 11-15, International Association on Water Quality, Surrey, England. Feachem, R.G., Bradley, D.J., Garelick, H. and Mara, D.D. (1993) Elements and Health Risks of Excreta and Wastewater. In: Sanitation and Disease - Health Aspects of Excreta and Wastewater Management, World Bank Studies in Water Supply Sanitation 3, pp. 3-22. Friedler, E., Brown, M.D. and Butler, D., (1996) A Study of WC Drived Sewer Solids. Water Sci. Tech., Vol. 33, No. 9, pp. 17-24. Gernaey, K., Vanrolleghem, P. A. and Lessard, P. (2001) Modeling of a Reactive Primary Clarifier. Water Sci. Tech. Vol. 43, No. 7, pp. 73-81. Grady, C.P.L., Jr., G.T. Daigger and H.C. Lim. (1999) Biological Wastewater Treatment, 2nd ed., Marcel Dekker, Inc., New York. Grijspeerdt, K., Vanrolleghem, P. and Verstraete, W. (1995) Selection of One-dimensional Sedimentation: Models for On-line Use. Water Sci. Tech. Vol. 1, No. 2, pp. 193-204. Hamilton, J., Jain, R., Antoniou, P., Svoronos, S.A., Koopman, B. and Lyberatos, G. (1992) Modeling and Predenitrification Process. J. Environ. Engng, ASCE, Vol. 118, No.l , pp. 38-55. Hartel, L. and Popel, H.J. (1992) A Dynamic Secondary Clarifier Model Including Processes of Sludge Thickening. Water Sci. Tech. Vol. 25, No. 6, pp.267-284. Hazen, A., 1904, On Sedimentation. Transactions, ASCE, Vol 53, pp. 38-55. Hill, R. D. (1985) Dynamics and Control of Solids-Liquid Separation in the Activated Sludge Process. Ph.D. Dissertation, Rice University, Houston, Tex. Haddon, M . (1995) Out of Your Sight? Out of Your Mind. Wat Bull. No. 642, pp.8-10. Imam, E., McCorquodale, J.A. and Bewtra, J.K. (1983) Numerical Modeling of Sedimentation Tanks. J. Hydraulic Div, ASCE, Vol. 107, No. 12, pp. 1740-1754. Jeppsson, U. and Diehl, S. (1996) On the Modeling of the Dynamic Propagation of Biological Components in the Secondary Clarifier. Water Sci. Tech. Vol. 34, No. 5-6, pp.85-92. Kennedy, A. S. (1994) A 1-Dimensional Quiescent Settler Model. M . Eng. Project Report, McMaster University, Hamilton, Ontario, Canada. Krebs, P., 1995, Success and shortcomings of clarifier modeling. Water Sci. Tech. Vol. 31, No. 2, pp. 181-191. Larsen, P. (1977) On the Hydraulics of rectangular settling basins. Report No. 1001, Department of Water Research Engineering, Lund Institute of Technology, University of Lund, Lund, Sweden. Lyn D.A. and Zhang, Z. (1989) Boundary-fitted numerical modeling of settling tanks. In Proceeding of the 23rd IAHR Conference, pp. A331-338. National Research Council, Ottawa, Canada. Lyn D.A., Stamou, A.I. and Rodi, W. (1992) Density Currents and Shear-induced Flocculation in Sedimentation Tanks. J. Hydraulic Div, ASCE, Vol. 118, No. 6, pp. 849-867. Metcalf & Eddy, Inc. (2003) Wastewater Engineering -Treatment and Reuse. Fourth Edition, McGraw-Hill Inc., New York, USA, PP. 398 and 687. National Bag It and Bin It Campaign (1995) Leedex Public Relations, London. Otterpohl, R. and Freund, M . (1992) Dynamic Models for Clarifiers of Activated Sludge Plants with Dry and Wet Weather Flows. Water Sci. Tech. Vol. 26, No. 5-6, pp. 1391-1400. Ozinsky, A.E., Ekama, G.A. and Reddy, B.D. (1994) Mathematical Simulation of Dynamic Behavior of Secondary Settling Tanks. Research Report W85, Department of Civil Engineering, University of Cape Town, South Africa. Parker, D.S., Kaufman, W.J. and Jenkins, D. (1971) Physical Conditioning of Activated Sludge Floe. J. Wat. Pollut. Control Fed,. Vol. 43, No. 9, pp. 1817-1833. Parker, D.S. (1983) Assessment of Secondary Clarification Design Concepts. /. Wat. Pollut. Control Fed., Vol. 55, No. 4, pp. 349-359. Parry, G. G. and Takacs, I. (1992) Settling of Flocculent Suspensions in Secondary Clarifiers. Water Res. Vol. 26, No. 4, pp. 473-479. Rachwal, A.J., Johnson, D.W.M., Hanbury, M.J. and Chitchard, D.J. (1982) The Application Of Seattleability Tests For Control Of Activated Sludge Plants. In Bulking Of Activated Sludge-Preventative And Remedial Methods ed. pp. 224-242, Ellis Horwood, Chichester. Schamber, D.R. and Larock, B.E. (1981) Numerical Analysis of Flow in Sedimentation Basin. J. Hydraulic Div., ASCE, Vol. 107(HY5), pp. 575-591. Sreckovic, G. (2001) Modeling Activated Sludge Treatment of Pulp and Paper Wastewater. Ph.D. Dissertation, University of British Columbia, Vancouver, B.C., Canada. Stamou, A.I. (1991) On the Prediction of Flow and Mixing in Settling Tanks Using a Curvature-modified k-eModel. Appl. Math. Modeling, Vol. 15 No. 7, pp. 351-358. Stamou, A.I., Adams, E.W. and Rodi, W. (1989) Numerical Modeling of Flow and Settling in Primary Rectangular Clarifiers. J. Hydraulic Res., Vol. 27, No. 5, pp. 665-682. Stenstrom, M . K. (1976) A Dynamic Model and Computer Compatible Control Strategies for Wastewater Treatment Plants. Ph.D. Dissertation, Clemson University, Clemson, S.C. Szalai, L., Krebs, P. and Rodi, W. (1994) Simulation of Flow in Circular Clarifiers With and Without Swirl. J. Hydraulic Engng., ASCE, Vol. 120, No. 1, pp. 4-17. Takacs, L, Patry, G. G. and Nolasco, D. (1991) A dynamic Model of the Clarification - Thickening Process. Water Sic. Tack Vol. 25, No. 10, pp.1263-1271. Vesilind P. A. (1968) Design of Prototype Thickener from Batch Settling Tests. Water Sewage Work. Vol. 115, No. 7, pp. 302-307. Vitasovic, Z. Z. (1986) An Integrated Control Strategy for the Activated Sludge Process. Ph.D. dissertation, Rice University, Houston, Tex. Vitasovic, Z. Z. (1989) Continuous Settler Operation: a Dynamic Model. In Dynamic Modelling and Expert Systems in Wastewater Engineering (Edited by Patry, G. G. and Chapman, D.), pp. 59-81. Lewis Publishers, Chelsea, Mich. Watts, R. W., Svoronos, S. A., and Koopman, B. (1996) One-dimensional modeling of secondary clarifiers using a concentration and feed velocity-dependent dispersion coefficient. Water Res. Vol. 30, No. 9, pp. 2112-2124. Zhou, S.P., Vitascovic, Z. and McCorquodale, J.A. (1996) Improving Performance of Large Rectangular Secondary Clarifiers - 3-D Modeling of PVSC Final Clarifiers. Unpublished report by Reid Crowther Consulting Inc., Seattle, Washington, USA, for Hazen and Sawyer, Consulting Engineers, New York, NY, USA. Appendices A MatLab Program for Dynamic simulation of the Takacs Model % T h i s program uses the Takacs One-dimensuonal Model t o s i m u l a t e the dynamics of 10 l a y e r s of p r i m a r y sludge i n a P&P c l a r i f i e r . % The Takacs Model: Vs=Vo*(exp(-rh)*(X-Xmin)-exp(-rp)*(X-Xmin)) % The out p u t i s a f i g u r e o f s i m u l a t e d SS o f e f f l u e n t . % Programed by: Ing-Wei Lo % Date: Apr 2003 c l e a r ; c l e a r g l o b a l ; cd d : \ c i v i l . u b c \ t h i s e s ; l o a d BPP3 0 . t x t ; format s h o r t ; % D e f i n e Takacs model p a r a r m e t e r s : Vo=145; %m/d (maxium t h e o r e t i c a l s e t t l i n g v e l o c i t y ) rh=0.0002516; %mA3/g ( s e t t l i n g p arameter of h i g h e r c o n c e n t r a t i o n ) rp=0.005507; %m*3/g ( s e t t l i n g parameter of lower c o n c e n t r a t i o n ) Xmin=41; %mg/L ( n o n - s e t t l i n g c o n c e n t r a t i o n ) % D e f i n e model c h a r a c t e r i s t i c s : dz=0.45; % m e t e r ( h e i g h t of l a y e r s ) d t = l / l 0 0 0 ; %day (time i n t e r v a l ) number=14; %the number f o r c a c u l a t i n g (number-1) l a y e r s Xo=[42 51 59 71 98 170 170 170 170 11171 26485 31600 39825]; %g/mA3 ( i n i t i a l c o n c e n t r a t i o n , assumed) A=2920; %mA2 ( s e c t i o n a r e a o f the c l a r i f i e r ) m=6; % f e e d i n g l a y e r time=[0 :dt : 29 . 9] ; he i g h t = [ 0 : d z : ( d z * ( n u m b e r - 1 ) ) ] ; Xin=BPP30(1 : l e n g t h ( t i m e ) , 3 ) ; %g/m*3 (SS of the i n f l u e n t ) Qin=BPP30(1 : l e n g t h ( t i m e ) , 2 ) ; %L/s ( i n f l u e n t f l o w r a t e ) Qw=BPP30(1 : l e n g t h ( t i m e ) , 5 ) ; %L/s (waste sludge f l o w r a t e ) Q r = z e r o s ( l e n g t h ( t i m e ) , 1 ) ; %L/s ( r e t u r n sludge f l o w r a t e ) Q u = z e r o s ( l e n g t h ( t i m e ) , 1 ) ; Q e = z e r o s ( l e n g t h ( t i m e ) , 1 ) ; Qu=Qr+Qw; %L/s (underflow r a t e ) Qe=Qin-Qu; %L/s ( o v e r f l o w r a t e ) %Generate SS a t d i f f e r e n t l a y e r s X = o n e s ( l e n g t h ( t i m e ) , ( n u m b e r - 1 ) ) ; J d n = o n e s ( l e n g t h ( t i m e ) , ( n u m b e r - 1 ) ) ; J u p = o n e s ( l e n g t h ( t i m e ) , ( n u m b e r - 1 ) ) ; J s = o n e s ( l e n g t h ( t i m e ) , ( n u m b e r - 1 ) ) ; X ( l , 1 : (number-1))=Xo; %SS a t l a y e r s a t i n i t i a l time. Jup(1,1:(number-1))=(Qe(1)*(86400/1000)*X(1,1)/A/0.91); Jdn(l , 1 : ( n u m b e r - 1 ) ) = ( Q u ( l ) * ( 8 6 4 0 0 / 1 0 0 0 ) * X ( 1 , 1 ) / A ) ; i f X ( l , l : (number-1))>265 & X ( 1,1 : (number-1))<1695; Js (1,1 : (number-1))=79.9*(X(l,1:(number-1))-Xmin); e l s e Js (1,1: (number-1)) = ( V o * ( e x p ( - r h ) * ( X ( l , 1 ) - X m i n ) - e x p ( - r p ) * ( X ( 1 , 1 ) -Xmin)))*X(1,1 : (number-1)); f o r i = 2: l e n g t h ( t i m e ) %time X ( i , l ) = m a x ( ( X ( ( i - 1 ) , 1 ) + ( d t / d z ) * ( J u p ( ( i - 1 ) , 2 ) - J u p ( ( i - 1 ) , 1 ) - m i n ( J s ( ( i -1 ) , 1 ) , J s ( ( i - 1 ) , 2 ) ) ) ) , X m i n ) ; % l a y e r 1 J u p ( i , 1 ) = Q e ( i ) * ( 8 64 00/1000)*X(i,1)/A/0.91; J d n ( i , 1 ) = Q u ( i ) * ( 8 6 4 00/10 0 0 ) * X ( i , 1 ) / A ; i f X ( i , l ) > 2 6 5 & X(i,1)<1695; J s ( i , 1 ) = 7 9 . 9 * ( X ( i , 1 ) - X m i n ) ; e l s e J s ( i , 1 ) = ( V o * ( e x p ( ( - r h ) * ( X ( i , 1 ) - X m i n ) ) - e x p ( ( - r p ) * ( X ( i , 1 ) - X m i n ) ) ) ) * X (i,1) ; end f o r j=2 : (number-1) i f j<(number-1) i f j<m % l a y e r above f e e d i n g l a y e r X ( i , j ) = m a x ( ( X ( ( i - 1 ) , j ) + ( d t / d z ) * ( m i n ( J s ( ( i - 1 ) , ( j - 1 ) ) , J s ( { i -1) , j ) ) +Jup ( ( i - l ) , ( j + D ) -Jup( (i-1) , j ) -min(Js ( ( i - l ) , j ) , J s ( ( i - l ) , ( j + l ) ) ) ) ) ,Xmin) ; e l s e i f j>m & j<13 % l a y e r lower than than f e e d i n g l a y e r X ( i , j ) = X ( ( i - l ) , j ) + ( d t / d z ) * ( J d n ( ( i - 1 ) , ( j - 1 ) ) + m i n ( J s ( ( i - 1 ) , (j 1)) , J s ( ( i - 1 ) , j ) ) - J d n ( ( i - 1 ) , j ) - m i n ( J s ( ( i - 1 ) , j ) , J s ( ( i - 1 ) , (j+1))) ) ; e l s e % f e e d i n g l a y e r ' X ( i , j ) = X ( ( i -1) ,j) + ( d t / d z ) * (Qin ( i ) * ( 8 6 4 00/1000)*Xin(i)/A+min( J s((i-1) , ( j - 1 ) ) , J s ( ( i - 1 ) , j ) ) -J d n ( ( i - 1 ) , j ) - m i n ( J s ( ( i - 1 ) , j ) , J s ( ( i - 1 ) , ( j + 1 ) ) ) - J u p ( ( i - 1 ) , j ) ) ; end end e l s e %buttom l a y e r X ( i , j ) = X ( ( i - 1 ) , j ) + ( d t / d z ) * ( J d n ( ( i - 1 ) , ( j - 1 ) ) + m i n ( J s ( ( i - 1 ) , ( j - l ) ) , J s ( ( i 1 ) , j ) ) - J d n ( ( i - 1 ) , j ) ) ; end J u p ( i , j ) = Q e ( i ) * ( 8 6 4 0 0 / 1 0 0 0 ) * X ( i , j ) / A / 0 . 9 1 ; J d n ( i , j ) = Q u ( i ) * ( 8 6 4 0 0 / 1 0 0 0 ) * X ( i , j ) / A ; i f X ( i , j ) > 2 6 5 & X ( i , j ) < 1 6 9 5 ; J s ( i , j ) = 7 9 . 9 * ( X ( i , j ) - X m i n ) ; e l s e J s ( i , j ) = ( V o * ( e x p ( ( - r h ) * ( X ( i , j ) - X m i n ) ) - e x p ( ( - r p ) * ( X ( i , j ) -X m i n ) ) ) ) * X ( i , j ) ; end end end % P l o t o f SS vs H i g h t p l o t ( X ( : , 1 ) , ' b : o ' ) %save a=X(:,1) b=X(:,13) save ppdy30day a b - a s c i i - t a bs; %end % T h i s program uses the Takacs One-dimensuonal Model t o s i m u l a t e the dynamics of 10 l a y e r s o f sec o n d a r y sludge i n a P&P c l a r i f i e r . % The Takacs Model: Vs=Vo*(exp(-rh)*(X-Xmin)-exp(-rp)*(X-Xmin)) % The output i s a f i g u r e o f s i m u l a t e d SS o f e f f l u e n t . % Programed by: Ing-Wei Lo % Date: Apr 2003 c l e a r ; c l e a r g l o b a l ; cd d : \ c i v i l . u b c \ t h i s e s ; l o a d BPP3 0 . t x t ; format s h o r t ; % D e f i n e Takacs model p a r a r m e t e r s : Vo=196; rh=0.0002876; rp=0.004369; Xmin=7; %m/d (maxium t h e o r e t i c a l s e t t l i n g v e l o c i t y ) %m^3/g ( s e t t l i n g parameter o f h i g h e r c o n c e n t r a t i o n ) %mA3/g ( s e t t l i n g parameter o f lower c o n c e n t r a t i o n ) %mg/L ( n o n - s e t t l i n g c o n c e n t r a t i o n ) % D e f i n e model c h a r a c t e r i s t i c s : dz=0.45; dt=l/1000; number=14 ; Xo=[12 20 29 %m e t e r ( h e i g h t of l a y e r s ) %day (time i n t e r v a l ) %the number f o r c a c u l a t i n g (number-1) l a y e r s 50 123 633 633 633 633 633 9945 17253 27890] ; %g/m*3 ( i n i t i a l c o n c e n t r a t i o n , assumed) A=2920; %mA2 ( s e c t i o n a r e a o f the c l a r i f i e r ) m=6; % f e e d i n g l a y e r time=[0 :dt: 29.95] ; hei g h t = [ 0 : d z : ( d z * ( n u m b e r - 1 ) ) ] ; Xin=BPP30(1 : l e n g t h ( t i m e ) , 1 0 ) ; %g/m' Q i n = B P P 3 0 ( l : l e n g t h ( t i m e ) , 7 ) ; %L/s Qw=BPP30(1 : l e n g t h ( t i m e ) , 9 ) ; %L/s Qr=BPP3 0(1 : l e n g t h ( t i m e ) , 8 ) ; %L/s Q u = z e r o s ( l e n g t h ( t i m e ) , 1 ) ; Q e = z e r o s ( l e n g t h ( t i m e ) , 1 ) ; Qu=Qr+Qw; %L/s (underflow r a t e ) Qe=Qin-Qu; %L/s (ove r f l o w r a t e ) 3 (SS o f the i n f l u e n t ) ( i n f l u e n t f l o w r a t e ) (waste sludge f l o w r a t e ) ( r e t u r n sludge f l o w r a t e ) ^Generate SS a t d i f f e r e n t l a y e r s X = o n e s ( l e n g t h ( t i m e ) , (number-1) ) ; J d n = o n e s ( l e n g t h ( t i m e ) , (number-1)) ; Jup= o n e s ( l e n g t h ( t i m e ) , ( n u m b e r - 1 ) ) ; J s = o n e s ( l e n g t h ( t i m e ) , ( n u m b e r - 1 ) ) ; X (1,1 : (number-1))=Xo; %SS a t l a y e r s a t i n i t i a l time. Jup(l,1:(number-1))=(Qe(1)*(86400/1000)*X(1,1)/A/0.91); Jdn(l,1:(number-1))=(Qu(1)*(86400/1000)*X(1,1)/A); i f X(1,1 : (number-1))>300 & X(1,1 : (number-1))<1545 ; Js (1,1 : (number-1))=125.4*(X(1,1 : (number-1)-Xmin)); e l s e Js (1,1 : (number-1) ) = (Vo* (exp (-rh) * (X (1, 1) -Xmin) -exp (-rp) * (X (1,1) Xmin)))*X(1,1 : (number-1)) ; f o r i = 2: l e n g t h ( t i m e ) %time X ( i , 1 ) = m a x ( ( X ( ( i - l ) , l ) + ( d t / d z ) * ( J u p ( ( i - 1 ) , 2 ) - J u p ( ( i - 1 ) , 1 ) - m i n ( J s ( ( i -1 ) , 1 ) , J s ( ( i - 1 ) , 2 ) ) ) ) , X m i n ) ; % l a y e r 1 J u p ( i , 1 ) = Q e ( i ) * ( 8 6 4 0 0 / 1 0 0 0 ) * X ( i , l ) / A / 0 . 9 1 ; J d n ( i , 1 ) = Q u ( i ) * ( 8 6 4 0 0 / 1 0 0 0 ) * X ( i , l ) / A ; i f X ( i , l ) > 3 0 0 & X(i,1)<1545; Js ( i , 1) =125 .4* (X ( i , 1)-Xmin) ; e l s e J s ( i , 1 ) = ( V o * ( e x p ( ( - r h ) * ( X ( i , 1 ) - X m i n ) ) - e x p ( ( - r p ) * ( X ( i , 1 ) - X m i n ) ) ) ) * X (i,1) end f o r j=2 : (number-1) i f j <(number-1) i f j<m % l a y e r above f e e d i n g l a y e r X ( i , j ) = m a x ( ( X ( ( i - 1 ) , j ) + ( d t / d z ) * ( m i n ( J s ( ( i - 1 ) , ( j - 1 ) ) , J s ( ( i -1 ) , j ) ) + J u p ( ( i - 1 ) , ( j + 1 ) ) - J u p ( ( i - 1 ) , j ) - m i n ( J s ( ( i - l ) , j ) , J s ( ( i - l ) , ( j + l ) ) ) ) ) , X m i n ) ; e l s e i f j >m & j<12 % l a y e r lower than than f e e d i n g l a y e r X ( i , j ) = X ( ( i - l ) , j ) + ( d t / d z ) * ( J d n ( ( i - 1 ) , ( j - 1 ) ) + m i n ( J s ( ( i - 1 ) , ( j 1) ) , Js ( (i-1) , j ) ) -Jdn( ( i - 1 ) , j ) -min(Js (• ( i - l ) , j ) , Js ( (i-1) , ( j+1) ) ) ) ; e l s e i f j==12 % l a y e r s of RAS withdrawn X ( i , j ) = X ( ( i - 1 ) , j ) + ( d t / d z ) * ( J d n ( ( i - 1 ) , ( j -1 ) ) * ( ( Q r ( i ) / 2 + Q w ( i ) ) / Q u ( i ) ) + m i n ( J s ( ( i - 1 ) , ( j - 1 ) ) , J s ( ( i - 1 ) , j ) ) - J d n ( ( i -1 ) , j ) * ( ( Q r ( i ) / 2 + Q w ( i ) ) / Q u ( i ) ) - m i n ( J s ( ( i - 1 ) , j ) , J s ( ( i - 1 ) , (j+1) ) ) ) ; e l s e % f e e d i n g l a y e r X ( i , j ) = X ( ( i -1) ,j) + ( d t / d z ) * ( Q i n ( i ) * ( 8 64 0 0 / 1 0 0 0 ) * X i n ( i ) / A + J s ( ( i - 1 ) , ( j - 1 ) ) - J d n ( ( i - 1 ) , j ) -m i n ( J s ( ( i - l ) , j ) , J s ( ( i - l ) , ( j + l ) ) ) - J u p ( ( i - 1 ) , j ) ) ; end end e l s e %buttom l a y e r X ( i , j ) = X ( ( i - l ) , j ) + ( d t / d z ) * ( J d n ( ( i - 1 ) , ( j - 1 ) ) * ( Q w ( i ) / Q u ( i ) ) + m i n ( J s ( ( i -1), ( j - 1 ) ) , J s ( ( i - 1 ) , j ) ) - J d n ( ( i - 1 ) , j ) * ( Q w ( i ) / Q u ( i ) ) ) ; end J u p ( i , j ) = Q e ( i ) * ( 8 6 4 0 0 / 1 0 0 0 ) * X ( i , j ) / A / 0 . 9 1 ; J d n ( i , j ) = Q u ( i ) * ( 8 6 4 0 0 / 1 0 0 0 ) * X ( i , j ) / A ; i f X ( i , j ) > 3 0 0 & X ( i , j ) < 1 5 4 5 ; J s ( i , j ) = 1 2 5 . 4 * ( X ( i , j ) - X m i n ) ; e l s e J s ( i , j ) = ( V o * ( e x p ( ( - r h ) * ( X ( i , j ) - X m i n ) ) - e x p ( ( - r p ) * ( X ( i , j ) -X m i n ) ) ) ) * X ( i , j ) ; end end end % P l o t of SS vs H i g h t and save p l o t ( X ( : , D ) a=X(:,l) b=X(:,13) save psdy30day a b - a s c i i - t abs; %end Appendices B Data of Settling Test Blanket level (cm) Column #1 Settling Velocitvfm/d) Column #2 Settling Velocity(m/d) Column «3 Settling Velocitv(m/d) Column #4 Settling Velocitv(m/d) Column #5 Settling VelocitvOn/d) SS lab analysis (mg/L 7248 3720 3554 1802 1246 beginning time 0 0.0 0.0 0.0 0.0 0.0 2 0.4 2.9 0.6 4.3 1.0 7.2 1.5 10.8 5.0 36.0 4 0.8 2.9 1.8 8.6 2.4 10.1 6.0 32.4 14.0 64.8 6 1.2 2.9 6.4 33.1 7.0 33.1 14.5 61.2 28.0 100.8 8 1.6 2.9 12.6 44.6 14.0 50.4 24.0 68.4 34.0 43.2 10 2.0 2.9 15.5 20.9 20.0 43.2 29.0 36.0 37.5 25.2 12 2.5 3.6 17.2 12.2 22.6 18.7 32.0 21.6 39.8 16.6 14 3.0 3.6 18.9 12.2 " 25.0 17.3 34.5 18.0 42.0 15.8 16 3.5 3.6 20.5 11.5 27.0 14.4 36.0 10.8 44.0 14.4 18 4.0 3.6 21.9 10.1 27.5 3.6 36.5 3.6 44.8 5.8 20 5.0 7.2 23.2 9.4 27.9 2.9 36.8 2.2 45.5 5.0 22 6.0 7.2 24.2 7.2 24 7.0 7.2 25.2 7.2 26 8.0 7.2 26.2 7.2 28 9.0 7.2 27.1 6.5 30 10.0 7.2 28.0 6.5 32 110 14.4 28.8 5.8 34 14.0 14.4 29.5 5.0 36 16.0 14.4 29.6 0.7 38 18.0 14.4 29.7 0.7 40 20.0 14.4 29.8 0.7 Max Velocity 14.40 38.9 46.8 64.8 100.8 Blanket level (cm) Column Hi Settling Ve!ociiv(m/dl Column «3 Settling Ve!ocitv(m/dl Column f?4 Settling Vclocitv(m/d> Column fl6 Settling Velocitv(m/di Column K7 Settling Velocitv(m/d) Column &8 Seulini> Velocity! m/dl Column £9 SS lab analysis (nv/L) 72IO 6800 5360 3790 2873 2087 1695 beginning, time 0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1 0.0 0.0 0.0 0.0 0.0 0.0 1.0 14.4 1.0 14.4 1.0 14.4 4.0 •) 0.0 0.0 0.0 0.0 0.0 0.0 2.7 24.5 2.5 21.6 5.0 57.6 12.0 3 0.0 0.0 0.0 0.0 0.5 7.2 4.8 30.2 5.8 47.5 10.5 79.2 21.0 4 0.0 0.0 0.0 0.0 1.0 7.2 7.8 43.2 9.7 56.2 16.2 82.1 27.8 5 0.0 0.0 0.1 1.4 1.5 7.2 10.5 . 38.9 14.0 61.9 21.7 79.2 34.0 6 0.0 0.0 0.5 5.8 3.3 25.9 13.0 36.0 18.0 57.6 26.3 66.2 39.0 7 0.0 0.0 1.0 7.2 5.2 27.4 15.5 36.0 21.7 53.3 30.2 56.2 41.0 S 0.5 7.2 1.4 5.8 7.0 25.9 18.0 36.0 25.4 53.3 34.3 59.0 43.0 9 1.1 8.6 2.8 20.2 9.0 28.8 20.0 28.8 28.7 47.5 37.0 38.9 44.5 10 1.7 8.6 4.0 17.3 11.0 ' 28.8 21.5 21.6 31.2 36.0 39.5 36.0 45.0 11 2.8 15.8 5.5 21.6 12.7 24.5 24.0 36.0 33.2 28.8 40.8 18.7 12 4.0 17.3 7.0 21.6 14.4 ' 24.5 25.5 21.6 35.0 25.9 42.0 17.3 13 5.3 18.7 8.5 21.6 16.0 23.0 27.0 21.6 36.5 21.6 43.0 14.4 14 6.5 17.3 9.9 20.2 17.5 21.6 28.2 17.3 37.7 17.3 43.8 11.5 15 7.6 15.8 11.4 21.6 19.0 21.6 29.2 14.4 38.7 14.4 44.5 10.1 16 8.7 15.8 12.6 17.3 20.4 20.2 30.1 13.0 IS 10.5 13.0 14.9 16.6 23.1 19.4 32.1 14.4 20 12.5 14.4 17.5 18.7 25.2 15.1 33.2 7.9 22 14.2 12.2 19.5 14.4 25 16.5 11.0 22.2 13.0 32 21.5 10.3 vlax Velocity (3min avel 17.8 21.6 27.8 39.4 58.6 80.2 I Test item: P&P primary sludge - lower concentration Sample SS (mp/U Settling Velocity (m/d) 1 208 77.0 2 • 144 75.9 3 136 48.7 4 124 39.7 5 108 40.6 6 78 27.1 Rlankel level (cm) Column «1 Seulini' V'clocitvtrTi/di Column «2 Settling Velocilv(nVd) Column fi3 Setllins Velocitv(m/d) Column «4 Settling Vr!ucitv(m/d) Column tiS Seitlin? Vc]ociiv(m/d) Column *6 Settline Veloci!v(m/d) SS lab analysis (m>>/L) 25975 13800 9500 5600 4700 1W0 oepinnine time 0 0.1 0.0 0.0 0.0 0.0 0.0 1 0.1 0.7 0.2 1 ) 0.0 0.0 0.0 0.0 0.6 8.6 1.5 100.8 •) 0.2 0.7 0.3 1 I 0.0 0.0 0.0 0.0 2.3 24.5 8.5 10O.8 3 0.2 0.7 0.5 2.2 0.1 1.4 0.2 2.9 4.0 24.5 15.5 100.8 4 0.3 0.7 0.6 •> •) 0.2 1.4 0.4 2.9 6.6 37.4 22.5 100.8 5 0.3 0.7 0.8 1 1 0.3 1.4 1.0 8.6 9.4 40.3 29.5 93.6 6 0.4 0.7 0.9 2.2 0.5 2.9 2.8 25.9 13.1 53.3 36.0 57.6 7 0.4 0.7 1.1 2.2 0.7 2.9 5.0 31.7 16.4 47.5 40.0 25.9 8 0.5 0.7 1.2 9 •) 0.9 2.9 7.5 36.0 19.6 46.1 41.8 20.2 9 0.5 0.7 1.4 •> •) 1.2 4.3 9.4 27.4 22.6 43.2 43.2 20.2 10 0.6 0.7 1.5 11 1.5 4.3 11.3 27.4 25.1 36.0 44.3 15.8 11 0.6 0.7 1.8 3.6 2.0 7.2 13.0 24.5 27.0 27.4 45.0 10.1 12 0.7 0.7 2.0 3.6 2.5 • 7.2 14.6 23.0 28.6 23.0 45.5 7.2 13 0.7 0.7 2.3 3.6 3.0 7.2 16.2 23.0 30.2 23.0 46.0 7.2 14 0.8 0.7 2.5 3.6 3.5 7.2 17.8 23.0 31.8 23.0 46.5 7.2 15 0.8 0.7 2.8 3.6 4.2 10.1 19.4 23.0 32.9 15.8 47.0 7.2 16 0.9 0.7 3.0 3.6 5.1 13.0 21.0 23.0 33.9 14.4 47.5 7.2 IS 0.9 0.4 3.5 3.6 6.5 10.1 23.0 14.4 35.9 14.4 48.2 5.0 20 1.0 0.7 4.0 3.6 7.7 8.6 24.5 10.8 37.7 13.0 48.8 4.3 22 1.1 0.7 4.8 5.8 8.9 8.6 25.6 7.9 25 1.3 0.7 6.0 5.8 10.7 8.6 27.0 6.7 30 1.5 0.7 8.0 5.8 13.6 8.4 35 1.8 0.7 10.0 5.8 16.1 8.1 40 2.0 0.7 12.0 5.8 19.2 8.1 45 2.3 0.7 14.0 5.8 22.0 8.1 50 2.5 0.7 16.0 5.S 60 3.0 0.7 1ax Velocity (>3min ave) 0.7 5.8 11.1 31.7 49.0 100.8 I Test item: P&P primary sludge - lower concentration Sample SS (me/D Settling Velocity (m/d) 1 !48 88.3 2 104 58.0 3 92 39.8 4 42 20.3 Blanket level (cm) Column #1 Settling VelocitvOn/d) Column #2 Settling VelocityCm/d) Column #3 Settling Velûcitv(m/d) Column #4 Settling Velocilv(m/d) Column #5 Settling Velocitv(m/d) Column #6 Settling Velocity!m/d.) SS lab analysis (mg/L) 9126 7246 5374 2694 2272 1261 beginning lime 15:30 0 0.0 0.0 0.0 0.0 0.0 0.0 2 0.5 3.6 0.8 5.8 2.0 14.4 4.0 28.8 11.0 79.2 16.0 115.2 4 1.0 3.6 1.9 7.9 3.5 10.8 11.5 54.0 23.8 92.2 34.0 129.6 6 1.8 5.8 3.5 11.5 5.S 16.6 22.0 75.6 35.0 80.6 51.0 122.4 8 2.7 6.5 6.0 18.0 10.2 31.7 31.9 71.3 40.6 40.3 53.7 19.4 10 3.6 6.5 8.1 15.1 15.3 36.7 38.0 43.9 43.2 18.7 55.0 9.4 12 4.3 5.0 10.0 13.7 17.6 16.6 42.1 29.5 45.0 13,0 55.8 5.8 14 5.0 5.0 11.5 10.8 20.5 20.9 45.1 21.6 46.1 7.9 56.8 7.2 16 5.6 4.3 13.0 10.8 • 22.6 15.1 47.3 15.8 47.4 9.4 57.5 5.0 18 6.2 4.3 14.1 7.9 24.3 12.2 20 6.8 4.3 15.3 8.6 25.8 10.8 vlax Velocity (3min ave) 6.30 16.6 34.2 66.96 86.4 126 1 Blanket level (cm) Column «1 Settling Ve)ocilv(m/d Column «2 SelllinE Wlocilvim/d Column «3 Seltlim' Ve!ociiv(m/d' Column w Settlinc Velocitvfm/d Column «5 Settling Velocitvtm/d Column «6 Settling VelocuvlmAJ Column s7 Settling VrioruvfmAJ) S3 lab analvsis (mf/L) 17660 6680 4700 4327 3123 2253 1546 beginning lime 14:00 0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1 0.1 1.4 0.7 10.1 1.8 25.9 2.8 40.3 5.0 72.0 7.0 100.8 10 0 144.0 2 0.2 1.4 1.5 11.5 4.0 31.7 6.0 46.1 10.7 82.1 15.0 115.2 21.0 158.4 3 0.3 1.4 2.0 7.2 6.6 37.4 9.5 50.4 16.8 87.8 23.5 122.4 32.0 158.4 4 0.4 1.4 2.5 7.2 10.5 56.2 13.8 61.9 23.0 89.3 31.5 115.2 44.0 172.S 5 0.5 1.4 3.1 8.6 13.8 47.5 17.6 54.7 28.0 72.0 37.0 79.2 47.5 50.4 6 0.6 1.4 4.5 20.2 17.0 46.1 21.2 51.8 32.5 6J.8 41.0 57.6 49.0 21.6 7 0.7 1.4 6.2 24.5 20.0 43.2 24.9 53.3 35.5 43.2 42.7 24.5 50.5 21.6 8 0.8 1.4 7.5 18.7 22.1 30,2 27.0 30.2 37.5 2S.S 44.0 18.7 51.5 14.4 9 0.9 1.4 9.5 2S.8 24.5 34.6 29.0 28.8 39.2 24.5 45.0 14.4 57.5 14.4 10 1.0 1.4 11.1 23.0 26.5 28.8 30.5 21.6 40.2 14.4 46.5 21.6 53.0 7.2 11 1.1 1.4 12.4 1S.7 27.9 . 20.2 12 1.2 1.4 13.5 15.8 29.1 17.3 13 1.3 1.4 14.S 18.7 30.1 14.4 14 1.4 1.4 16.1 18.7 30.9 11.5 15 1.5 1.4 17.5 20.2 31.5 8.6 16 1.6 1.4 18.8 18.7 32.1 8.6 IS 1.7 0.7 21.3 18.0 32 S 5.0 20 1.8 0.7 22 5.0 ' 33.3 3.6 Max Velocity (2min ave) 1.4 25.9 51.6 58.3 88.6 118.8 165.6 1 Test iiem: PAP secondary sludge - lower concentration Sample SS (mg/L) Settling Velocity (m/d 1 437 153.9 2 136 92.4 3 88 30.5 4 76 33.0 5 60 24.3 6 52 17.0 7 42 19.3 1 i 2 s s S : *7ê o sd I ! a »i S * —. J s i I = S p I s — 1 5 1 s? S S s s s = = i ci 1 s 2 1 s o s S P. S S S î = s s s s s •o. s s S 2 2 O P 2 S 3 2 S s S s s S s 2 S ? 3 3 5 II § S S 8 s S S S ° .o S S S 5 ? S s S ? S s s s 5 3 S P S S 3 2 s S S S => — s S S s S S 2 2 S s 2 S 2 2 2 S s 2 2 2 2 2 2 2 2 2 2 2 2 2 !s s 5 S s S 2 s S s 2 S S s 2 S S I — 11 Ï s S ? s g 1 1 K K s s 3 5 S s •« <> s 5 5 5 s 1 « S 3 s S S I. 1 5 o 3 s g K S II § rs § p s s P p R c s s 2 ? 1 s «, <~; s s 5 3 î i g g = s ? o ?, ? ? S S <-< K S V R R p p K -° 1 5 S II § S 5 5 s S § S r-S S p <^  rs s s S 2 2 5 oc i l g S s 5 3 S 1 § ? 1 g R f s S P S <- -» S •a s H 2 2 2 ? ? â f g -o o 3 p O 2 r-i O •o r-2 S S R s P s S s f-2 2 2 2 2 Ë S i s S K J 1 § £> s s 5 3 g s; — 1 S g s» 1 Z _ z 3 P = = = - = s s P s O i i s s S S T g = â ? i ! il E o o r; p 1 11 1 — i Test item: Municipal primary sludge - 2 Test date: 10-Apr-03 -a Blanket level (cm) Column #2 Settling Velûcitv(m/d) Column #3 Settling Velûcitv(nVd) Column #4 Settling Velûcity(m/d) Column #5 Settling Velùcitv(m/d) Column #6 Settling Velocitv(m/d) Column #7 Settling VeliKiiv(nVd) SS lab analysis (mg/L) 14050 11767 11275 7633 4500 2567 beginning time 0 0.0 0.5 0.0 0.0 0.0 0.0 1 0.0 0.0 0.6 1.4 0.1 1.4 0.1 1.4 1.3 18.7 2.0 23.8 2 0.0 0.0 0.7 1.4 0.2 1.4 0.2 1.4 3.0 24.5 4.0 23.8 3 0.0 0.0 0.8 1.4 0.3 1.4 0.3 1.4 5.0 28.8 6.5 36.0 4 0.1 1.4 0.9 1.4 0.4 1.4 0.4 1.4 7.2 31.7 10.3 54.7 5 0.2 1.4 1.0 1.4 0.5 1.4 0.5 1.4 10.0 40.3 14.1 54.7 6 0.3 1.4 1.1 1.4 0.6 1.4 0.7 ' 2.9 12.8 40.3 18.0 56.2 7 0.4 1.4 1.2 1.4 0.7 1.4 0.9 2.9 15.3 36.0 22.0 57.6 8 0.5 1.4 1.3 1.4 0.8 1.4 1.1 2.9 17.7 34.6 26.2 60.5 9 0.6 1.4 1.4 1.4 1.1 4.3 1.4 4.3 20.0 33.1 28.8 37.4 10 0.7 1.4 1.5 1.4 1.5 5.8 1.7 4.3 22.3 33.1 31.0 31.7 1! 0.9 2.9 1.6 1.4 1.9 5.8 2.2 7.2 24.4 30.2 33.1 30.2 12 1.1 2.9 1.7 1.4 2.5 8.6 2.7 7.2 26.5 30.2 35.2 30.2 13 1.3 2.9 1.8 1.4 3.1 8.6 3.3 8.6 28.0 21.6 37.0 25.9 14 1.5 2.9 1.9 1.4 3.8 10.1 4.0 10.1 29.5 21.6 37.8 - 11.5 15 1.7 2.9 2.1 2.9 4.5 10.1 4.7 10.1 31.0 21.6 33.5 10.1 16 2.0 4.3 2.3 2.9 5.2 10.1 5.4 10.1 32.4 20.2 39.9 10.1 18 2.6 4.3 2.7 2.9 6.6 10.1 7.4 14.' 35.2 20.2 41.3 10.1 20 3.5 6.5 3.6 6.5 8.0 10.1 10.0 18.7 38 20.2 42.6 9.4 22 4.4 6.5 4.6 7.2 9.4 10.1 12.8 20.2 25 5.8 6.7 6.1 7.2 11.5 10.1 17.5 22.6 30 8.1 6.6 8.6 7.2 15.1 10.4 23.5 17.3 35 10.4 6.6 11.1 7.2 1S.6 10.1 27.5 11.5 40 12.7 6.6 13.6 7.2 22.0 9.8 30.7 9.2 45 15.0 6.6 16.1 7.2 25.4 9.8 33.0 6.6 50 17.3 6.6 18.6 7.2 28.8 9.3 35.2 6.3 lax Velocity (>3min ave ) 6.6 7.2 10.1 20.5 38.9 57.3 T > ? s n S 5 9 S g ? s S S 1 s ? •A S s i 1 S i i p § S 2 i 5 s s g ?: S 3 S O R S ? g S s r , 7-, 2 3 S E s s s S s s ? S § S s S 2 s S s P P p p p P p p S S s 5 3 s S s I s S g S S s s 2 s S 2 2 2 2 2 2 2 2 2 2 2 2 2 s 2 2 2 2 2 2 2 § s S ; 5 5 2 p s S 2 S 5 S S 3 5 ï 3 5 3 ' , 3 5 ; 3 3 3 3 3 3 cl'" s S P •t ri s ? w 3 3 „' s S S ? S ? °° oo oo oo oo oo »> OO OO oo oo OO 1 o § 5 2 S s s S S Z Z ™ O S 7, ï S Jj •1 2 := s. ? s f s s ? s S s s s g S 2 S ? ? s; S 2 5 S3 5 ss S ? S S S S 3 3 55 32S3 S S 5 ° S g s s s s 5 2 S S 2 S S S §3 2 S S 5S PS 22PSg ii 333 2 2 S 3 S S S il is 22222S22222822222222222222 gg 55 333333S3222SSSSS s s s s s s s Î Ï S S 1 5 5 5 S ? i a S E: i r-<=> g S ? ? S 5 S o s 1 s i S 5 S S s p § s ; s 3 1 2 S 5 •A ? S S S § s 2 s s I 5 s S ? ? K ? 1-4 r- 5 r; i 2 P g 5 S s 2 s S Z 3 s 2 = 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 § 2 S s r- „ - „ S S S § 2 fi S S S S S S § S 5 s S s ~- *p s s 3 5 2 5 S 2 3 S s S 3 S S 3 3 3 s 3 3 s S 3 3 S 3 S S S S S 3 S S g 5 5 s s 2 5 5 2 2 S 'o S S Z-'J Z- 5 2 a 2 2 2 2 2 2 2 2 2 2 2 2 2 S 2 2 2 2 2 2 2 2 2 s 5 5 S S 3 S S S 2 2 2 2 2 g 2 s S S ê S > i - 2 z 2 c s s C s O C s 5 > i i — 1 1 ? i i 3 •i i s g 3 S s s P g K y ïr. S ! i s S a K = S S s O S l § S s s S s 5 ? a P. i 3 S 1 K K s K s 2 p à S E J 1 g 2 s £ £ ? S 5 f r-S R PI 1 I S Ë s s b s J 1 2 5 S Z!; S 3 s 1 Ê î 3 S ? 5 £ <^  1 *= § 1 1 S S 2 U S S 3 Z 2 "J = ? s S î > - ; S S 3 3 S 2 S 2 o t— o 2 O 2 5 î Z z Z Z Z Z Z z S 2 S 3 ! g 5 3 oo O O o I 5 S s s S s S S S S S S 3 3 S 3 s £ S S 2 3 S S s S i § § 5 5 5 s 3 3 3 s s S s 5 £ 3 S g S 2 s z-i | 1 2 s 2 g Pi s s s 5 S 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0063472/manifest

Comment

Related Items