Wiley Wastewater Quality Monitoring and Treatment_10 pptx

Elements of Modelling and Control of Urban Wastewater Treatment Systems Olivier Potier and Marie-No¨elle Pons 3.1.1 Introduction 3.1.2 Short Description of the Biological Process by Acti

this means that the system is not as fully automated as some might hope and that regular visits to the stations by employees should be foreseen Also, a too complex

‘black-box’ concept of the system leads to a significant loss of data The processing

of the sensor signal to data should be transparent showing what can be done by using PC-based modules for the control of the station A web-based communication enables remote control of the stations and the integration of the data into databases This concept also allows for a full remote control of the station by authorised persons and

a limited accessibility for data consultation by users through the web A better spatial representation can be obtained by embedding the monitoring and the modelling in a GIS system (Vivoni and Richards, 2005)

2.3.4 CONCLUSIONS AND PERSPECTIVES

Monitoring in rural areas needs a different approach than in urban areas The pollu-tion in rural areas cannot be measured at certain points along the water body, but can only be estimated by making evaluations of the water quality together with infor-mation on what and how many polluting substances are applied in the area Models, describing all processes on those substances before entering the water, can provide

a means to evaluate the magnitude of pollution coming from diffuse pollution and

to evaluate scenarios for diffuse pollution reduction Specific data are needed to calibrate and build those models

Therefore, the traditional cycle in water management should be inversed Instead

of starting from the data set to select an appropriate tool and hence use this tool for management, one should first define the problem, select a tool that can support this problem and then design an appropriate monitoring program to feed the tool In that way, money is spent to generate primarily the information that is indeed needed A closer cooperation between monitoring and modelling efforts will make sure that models for diffuse pollution can be used with sufficient reliability

Automated monitoring can help to catch the high variability or short rain-driven events Such tools can only provide reliable data provided that the monitoring system

is transparent and follows quality control procedures with regard to maintenance and calibration While a high level of automation may support such procedures, it still requires considerable manpower that should be foreseen in any monitoring budget

REFERENCES

Arnold, J.G., Williams, J.R., Srinivasan, R and King, K.W (1996) SWAT Manual USDA,

Agri-cultural Research Service and Blackland Research Center, Texas.

Barthelemy, P.A and Vidal, C (1999) A dynamic European agricultural and agri-foodstuff

sec-tor In: Agriculture, Environment, Rural Development, Facts and Figures – A Challenge for Agriculture European Commission Report, Belgium.

Beck, M.B (1987) Water Resour Res., 23(8), 1393.

Bervoets, L., Schneiders, A and Verheyen, R.F (1989) Onderzoek naar de verspreiding en de typologie van ecologisch waardevolle waterlopen in het Vlaams gewest Deel 1 - Het Dender-bekken, Universitaire Instelling Antwerpen In Dutch.

Boschma, M., Joaris, A and Vidal, C (1999) Concentrations of livestock production In: Agri-culture, Environment, Rural Development, Facts and Figures – A Challenge for Agriculture.

European Commission Report, Belgium.

Brown, L.C and Barnwell T.O (1987) The Enhanced Stream Water Quality Models QUAL2E and QUAL2E-UNCAS: Documentation and User Model EPA/600/3-87/007, USA.

Janssen, P.H.M., Heuberger, P.S.C and Sanders, S (1992) Manual Uncsam 1.1, a Software Package for Sensitivity and Uncertainty Analysis Bilthoven, The Netherlands.

Krysanova, V and Haberlandt, U (2001) Ecol Modelling, 150, 255–275.

McKay, M.D (1988) Sensitivity and uncertainty analysis using a statistical sample of input values.

In: Uncertainty Analysis, Y Ronen, ed CRC Press, Inc., Boca Raton, FL, pp 145–186 Montarella, L (1999) Soil at the interface between agriculture and environment In: Agriculture, Environment, Rural Development, Facts and Figures – A Challenge for Agriculture European

Commission Report, Belgium.

Pau Val, M and Vidal, C (1999) Nitrogen in agriculture In: Agriculture, Environment, Rural Development, Facts and Figures – A Challenge for Agriculture European Commission Report,

Belgium.

Poirot, M (1999) Crop trends and environmental impacts In: Agriculture, Environment, Rural Development, Facts and Figures – A Challenge for Agriculture European Commission Report,

Belgium.

Sevruk, B (1986) Proceedings of the ETH, IAHS International Workshop on the Correction of Precipitation Measurements, 1–3 April 1985 ETH Z¨urich, Z¨uricher Geographische Schriften,

Z¨urich, p 23.

Smets, S (1999) Modelling of nutrient losses in the Dender catchment using SWAT Masters dissertation Katholieke Universiteit Leuven –Vrije Universiteit, Brussels, Belgium.

Vandenberghe, V., van Griensven, A and Bauwens, W (2005) Water Sci.Technol., 51(3-4), 347–

354.

Vandenberghe, V., Goethals, P., van Griensven, A., Meirlaen, J., De Pauw, N., Vanrolleghem, P.A.

and Bauwens, W (2004) Environ Monitor Assess., 108, 85–98.

van Griensven, A and Bauwens, W (2001) Water Sci Technol., 43(7), 321–328.

van Griensven, A and Bauwens, W (2003) Water Resour Res., 39(10), 1348.

van Griensven, A., Vandenberghe, V and Bauwens, W (2002) Proceedings of the International IWA Conference on Automation in Water Quality Monitoring, 21–22, May 2002 Vienna,

Austria.

Vivoni, E.R and Richards, K.T (2005) J Hydroinform., 7(4), 235–250.

Elements of Modelling and

Control of Urban Wastewater

Treatment Systems

Olivier Potier and Marie-No¨elle Pons

3.1.1 Introduction

3.1.2 Short Description of the Biological Process by Activated Sludge

3.1.3 Process Parameters

3.1.3.1 Biokinetics 3.1.3.2 Oxygen Transfer 3.1.3.3 Hydrodynamics 3.1.3.4 Wastewater Variability 3.1.3.5 Mass Balance 3.1.4 Sensors

3.1.4.1 In-line Sensors 3.1.4.2 On-line Sensors 3.1.5 Introduction to the Control Methods of a Wastewater Treatment Plant

by Activated Sludge 3.1.6 Conclusion and Perspectives

Acknowledgement

References

Wastewater Quality Monitoring and Treatment Edited by P Quevauviller, O Thomas and A van der Beken

 2006 John Wiley & Sons, Ltd ISBN: 0-471-49929-3

3.1.1 INTRODUCTION

A wastewater treatment plant (WWTP) is an intricate system made of unit operations based on physical, biological and physico-chemical principles Its aim is principally the removal of organic, nitrogen and phosphorus pollution The basic processes are complex and the various arrangements of the unit operations which can be proposed lead to many possible configurations of WWTPs It is difficult to describe in detail all of the processes here and only the basics of biological treatment by activated sludge will be examined It is the most widespread for WWTPs of medium and

large size The interested reader will find more details in Henze et al (Henze et al.,

2000) We focus our attention on the most important parameters for optimization and process control of pollution removal in large plants, where spatial distribution of substrate and nutrient in the reacting system plays a large role In smaller plants, time scheduling can replace spatial gradients as in sequencing batch reactors for example Whatever the case and in spite of the perturbations in terms of flow, composition and concentration experienced at the inlet of any WWTP, specifications on the discharged water should be kept within strict limits to avoid taxes and penalties Different tools for monitoring and process control are also presented

3.1.2 SHORT DESCRIPTION OF THE BIOLOGICAL

PROCESS BY ACTIVATED SLUDGE

The biological step (often called secondary treatment) is an essential part of the WWTP At the inlet of the plant, the water is usually pretreated to remove gross debris (grit removal) and can be further treated in a primary settler, which will elim-inate a large part (usually 40–50 %) of the particulate pollution In doing so, part of the biodegradable pollution is indeed removed, which might not always be a good idea: denitrification, one of the steps involved in nitrogen pollution removal, requires

a certain balance between carbon and nitrogen and an external carbon source is often added in that step This could be avoided (or at least limited) by direct injection into the biological reactor of unsettled wastewater The principle of activated sludge is the intensification in a reactor of the principle of self-purification, which is naturally oc-curring in the environment, in presence of a much higher bacterial concentration than

in rivers or lakes The task of the secondary clarifier (Figure 3.1.1) is to separate the

biological reactor by activated sludge return sludge

purified water clarifier

Figure 3.1.1 Schematic representation of an activated sludge system

aerobic zone anoxic

zone

Post-denitrification

aerobic zone anoxic

zone

Pre-denitrification C

Figure 3.1.2 Schemes of different activated sludge reactors with anoxic zone

flocculated bacteria (sludge flocs) from the treated water The sludge is returned to the inlet of the reactor and the purified water is polished in a tertiary stage (post-treatment of phosphorus, filtration, disinfection, etc.) and/or discharged

In the presence of oxygen, carbon and a small amount of nitrogen (from ammonia and hydrolysed organic nitrogen) are metabolized by heterotrophic biomass and most

of the nitrogen by autotrophic bacteria The latter produced nitrates can be reduced

by heterotrophs under anoxic conditions As indicated previously, organic matter

is needed for this reaction and therefore an addition of carbon (such as methanol)

is often necessary In the case of a pre-denitrification system, mixed liquor from the outlet of the reactor is recycled to the anoxic zone Some of the most classical schemes are presented in Figure 3.1.2

In order to ensure the best process efficiency, different parameters must be known and controlled: the main reactions of pollution removal and their kinetics; the spatial distribution of the substrates with respect to the micro-organisms and therefore the reactor hydrodynamics; the aeration capacity and therefore the oxygen transfer; and the variability of the wastewater, in terms of composition, concentration and flow rate

3.1.3 PROCESS PARAMETERS

3.1.3.1 Biokinetics

Many different compounds and micro-organisms are found in a biological wastewater system In addition, the ecosystem is never at steady state Therefore, an exact and complete kinetic model is out of reach For many years the scientific community has tried to provide models of reasonable complexity, able to describe the main steps

of activated sludge behaviour The basic model is ASM1 (Activated Sludge Model

n◦1), devoted to carbon and nitrogen removal (Henze et al., 1987) Improved versions

have been proposed, such as ASM2, which takes into account phosphorus removal, and ASM3 (IWA, 2000)

ASM1 is a good compromise between the description of the complex reality of biological reactions and the simplicity of a model The identification of any model parameter should be possible theoretically (structural identifiability) and experimen-tally through experiments which can be run in the laboratory as well as on full-scale systems

S s

S o

S NH

X B,H

X B,A

X ND

S ND

S NO

r3

r4

r3

r3

r6

r8

r5

r5

r5

r2

r2

r2

r7

r1

r1

r1

S I : Soluble inert organic matter

S s : readily biodegradable substrate

X s : Slowly biodegradable substrate

X I : Particulate inert organic matter

X B,H : Active heterotrophic biomass

X B,A : Active autotrophic biomass

X p : Particulate products arising from biomass decay

S o : Oxygen

S NO : Nitrate and nitrite nitrogen

S NH : NH 4 + and NH 3 nitrogen

S ND : Soluble biodegradable organic nitrogen

X ND : Particulate biodegradable organic nitrogen

Figure 3.1.3 Schematic representation of the ASM1 kinetic pathways

As ASM1 is more particularly used, it will be described in some detail In ASM1 (Figure 3.1.3), wastewater compounds are divided into different categories: inert (i.e nonbiodegradable) versus biodegradable matter, particulate versus soluble Partic-ulate biodegradable matter should be hydrolysed to become readily biodegradable The biomass is divided into two parts: heterotrophic and autotrophic

Note that toxic events could trigger strong inhibition of bacteria Inhibition terms can be added to the basic ASM1 model for specific purpose (industrial wastewater mainly) Autotrophs are deemed to be more sensitive to toxics than heterotrophs

3.1.3.2 Oxygen Transfer

Influence of oxygen on pollution removal

Bacteria use oxygen for their respiration In the ASM1 model, the oxygen concen-tration is considered to be a substrate:

rFor the aerobic growth of heterotrophs, where readily biodegradable substrate is

consumed:

ρ1 = μ H



S S

K S + S S

 

S O

K O,H + S O



X B,H

whereρ1is the aerobic growth rate of heterotrophs, S S the biodegradable soluble

substrate concentration, S O the oxygen concentration, X B,H the concentration of

heterotrophs, K S the heterotrophic half-saturation coefficient for S S , K O,Hthe het-erotrophic half-saturation/inhibition coefficient for oxygen andμ H the maximum growth rate of heterotrophs

rFor the aerobic growth of autotrophs, where NH4+and NH3 nitrogen are trans-formed into nitrates:

ρ3= μ A



S N H

K N H + S N H

 

S O

K O,A + S O



X B,A

whereρ3 is the aerobic growth rate of autotrophs, S NHthe ammonium

concen-tration, X B,A the concentration of autotrophs, K NHthe autotrophic half-saturation

coefficient for S NH , K O,A the autotrophic half-saturation coefficient for oxygen andμ Athe maximum growth rate of heterotrophs

For the anoxic growth of heterotrophs, a very low concentration of oxygen is required to avoid any inhibition:

ρ2= μ H



S S

K S + S S

 

K O,H

K O,H + S O

 

S N O

K N O + S N O



η g X B,H

whereρ2 is the anoxic growth rate of heterotrophs, SNOthe nitrate concentration,

K NO the heterotrophic half-saturation coefficient for S NOandη gthe anoxic growth rate correction factor for heterotrophs

Thus, the oxygen concentration has a great importance: it should be low in the anoxic stages and nonlimiting in the aerated zones However, excessive oxygen supply should be penalized in terms of cost Oxygen is provided by gas diffusers or surface aerators

Oxygen transfer model

Generally, the gas–liquid transfer is modelled by means of the double film theory

(Roustan et al., 2003), according to which the gas–liquid interface is located between

a gas film and a liquid film For the oxygen–water system, the transfer resistance is found in the liquid film, due to the low solubility of oxygen in water The oxygen flux is a function of the difference between the oxygen concentration at saturation

(S O∗) and the dissolved oxygen concentration in the reactor (S O) and of the global

coefficient of oxygen transfer (k L a) Experimental values of k L a are generally between 2 h−1 and 10 h−1 If it is assumed that the reactor can be modelled as a Continuous Perfectly Mixed Reactor (CPMR) (Figure 3.1.4), with a uniform oxygen concentration, the oxygen mass balance is written as:

Q S O I + k L a(S O∗− S O )V = QS O + r O V + V dS O

dt with S O I the oxygen concentration at the inlet, r O the oxygen consumption rate, V the reactor volume and Q the liquid flow rate.

The coefficient transfer is measured directly in the presence of sludge (k L a), or

in clean water without sludge (k L a) (H´eduit and Racault, 1983a,b; ASCE, 1992;

Figure 3.1.4 An aerated Continuous Perfectly Mixed Reactor

Roustan et al., 2003) In this case, the so-called ‘alpha’ factor ( α) must be taken into

account (Boumansour and Vasel, 1996):

k L a= αk L a Example of oxygen profile in a WWTP bioreactor

To illustrate the open loop behaviour of a biological reactor, with no aeration adjust-ment as a function of the oxygen demand, the oxygen profile was measured during

1 day in a 3300 m3 channel reactor with a large aspect ratio The reactor is 100 m long and 8 m wide and aerated by means of fine bubble diffusers located on its floor The dissolved oxygen concentration was regularly measured in six locations along the reactor with a portable probe (WTW, Weilheim, Germany) (Figure 3.1.5) The

1.8

0.8 0.6 0.4 0.2 0

1.6 1.4 1.2 1

time/h

Figure 3.1.5 Variations of the dissolved oxygen concentration in different locations of an

acti-vated sludge channel reactor during 1 day

air flow rate was constant and equally distributed along the reactor, which was con-tinuously fed by urban presettled wastewater The dissolved oxygen concentration changes with the oxygen consumption, and therefore with biodegradable pollution concentration, which depends on time and space in the reactor

In Figure 3.1.5, it can be seen that dissolved oxygen concentration is higher during the night, when pollution is lower The concentration increases along the reactor as the oxygen consumption decreases due to a decrease in the biodegradable substrate availability During the day, dissolved oxygen concentration remains very low, even near the reactor outlet, which indicates complete pollution removal is not achieved Under such conditions oxygen limitation occurs Better aeration with a larger air flow rate could alleviate such a limitation without increasing the reactor volume

3.1.3.3 Hydrodynamics

In brief, two types of reactor shape are found: a compact, ‘parallelepipedic’ or

‘cylindrical’ design, often fitted with surface turbines for aeration; and an elongated design suitable for gas diffusion devices Elongated reactors are often folded or built as ‘race tracks’, which avoids recirculation pumps (Figure 3.1.6) In this case they are generally called ‘oxidation ditches’ when the aerators are horizontal and

‘carousels’ when they are vertical Many variations have been proposed by various manufacturers, such as sets of several concentric channels as in the OrbalTM sys-tem and OCOTM process, inclusion of anaerobic and anoxic zones equipped with mechanical mixing devices, or combination of spatial gradients along the tanks with alternating mode of operation, such as in the BiodeniphoTM or BiodenitroTM process Capacity, land availability, flow circulation, process type (carbon and/or nutrient removal) are some of the criteria for selection

Hydrodynamics have a great importance in a process, because linked with kinetics, they affect pollution removal efficiency and the bacteria species selectivity Usually the reactor behaviour is compared with one of two ideal types: the Continuous Perfectly Mixed Reactor (or CPMR) and the Plug Flow Reactor

The CPMR is characterized by a uniform concentration of each component in all the volume of the reactor This type of reactor can be found in small WWTPs, where the length is similar to the width

aerobic and anoxic zone

Figure 3.1.6 A ‘race track’ reactor

1 2 3 J

Figure 3.1.7 J CPMRs in series

The Plug Flow Reactor model is very different It is composed of a succession of parallel volumes infinitiely small, perpendicular to the flow, with no transfer between them These volumes move forward from the inlet to the outlet, at a velocity linearly related to the flow There is a progressive change in concentrations However, if the ideal Plug Flow Reactor model could be used for tubular or fixed-bed reactors in the chemical industry, it rarely represents in a satisfactory manner an aerated tank in a WWTP

Models based on CPMRs in series (Figure 3.1.7) offer the best simple alternative to model full-scale plants and generally give a good agreement with experimental data

Theoretically, the number of reactors in series (J ) can vary between 1 and infinity In practice, J is determined by tracing experiments and takes values between 3 and 20 Although a series of J CPMRs is a discrete hydrodynamic model, it can model a

continuous liquid system like a channel reactor

Hydrodynamic characterization

A relatively simple method for the characterization of hydrodynamics is the Resi-dence Time Distribution (RTD) method Each molecule has is own resiResi-dence time

(r t) in the reactor, which depends on the reactor hydrodynamics (Figure 3.1.8) The goal of the RTD method is to measure the different residence times based on statis-tics A pulse of nonreactive tracer is injected at the inlet of the reactor Different chemical substances are used, such as lithium chloride (detection by atomic ab-sorption), rhodamine (detection by fluorescence sensor) and radioactive elements The tracer is dissolved in the mixed liquor in the reactor and behaves as the liquid phase At the reactor outlet, the tracer concentration is measured to calculate the RTD (Villermaux, 1993; Levenspiel, 1999)

Figure 3.1.8 Inert tracing of a reactor