www.mbr-network.

eu

Introduction to Process Modelling Biological Process Modelling (Part I)

Peter A. Vanrolleghem

Summer school on modelling MBR processes July 15-17, 2008, Ghent - Belgium

The model building exercise

purpose prior knowledge

Framework Definition data Model Selection Parameter estimation Validation

Model

www.mbr-network.eu

The model building exercise

Biology Chemistry Physics

prior knowledge

Karplus arc
Parameter estimation

Economy Politics

www.mbr-network.eu

data

The model building exercise real world

system boundary (x, t) observations

observations

input data

model

output data

(Thomas Hug)
www.mbr-network.eu

Use of Models

input

model

?
output

predict

input

?
model

understand

input

output

communicate

www.mbr-network.eu

Modelled processes in WWT

model

www.mbr-network.eu

Modelled processes in WWT

Add a mixing model

model
MP A (R1) 20 Br R1 sim Br A mess 15 Br [g/m-3] Br [g/m-3] 10 15 Br B sim Br B mess MP B (R2)

10

0 -0.1

0.1 0.2 Versuchszeit [d] MP C (R3)

0.3

0.1 0.2 Versuchszeit [d] MP E

0.3

10 8 Br [g/m-3] 6 4 2 0 Br C sim Br C mess

8 Br E sim Br E mess 6 Br [g/m-3] 0 0.1 0.2 Versuchszeit [d] 0.3

www.mbr-network.eu

0.1 0.2 Versuchszeit [d]

0.3

Modelled processes in WWT

model

Add a Add an mixing model aeration model Add an Add an aeration influent model model

www.mbr-network.eu

Modelled processes in WWT

Add a mixing model

model
0 0.05 0.1 0.15 0.2 SB H (m) 0.25 0.3 0.35 0.4 0.45 0.5 0 0.002 0.004 0.006 0.008 0.01 time (d) 0.012 0.014 0.016 0.018 0.02

Add an Add a Add an aeration settler model influent model model

4000 mg/l 1600 mg/l 1200 mg/l 800 mg/l 2800 mg/l

www.mbr-network.eu

Modelled processes in WWT

Add a mixing model

model
0 0.05 0.1 0.15 0.2 SB H (m) 0.25 0.3 0.35 0.4 0.45 0.5 0 0.002 0.004 0.006 0.008 0.01 time (d) 0.012 0.014 0.016 0.018 0.02

Add an Add a Add an aeration settler model influent model Mimick model installed control system
4000 mg/l 1600 mg/l 2800 mg/l 1200 mg/l 800 mg/l

www.mbr-network.eu

Modelled processes in WWT

Add a mixing model

0 0.05 0.1 0.15 0.2 SB H (m) 0.25 0.3 0.35 0.4 0.45 0.5 0 0.002 0.004 0.006 0.008 0.01 time (d) 0.012 0.014 0.016 0.018 0.02

Add sludge Add an Add a Add anline aeration settler model influent model Mimick model installed control system
model
4000 mg/l 1600 mg/l 2800 mg/l 1200 mg/l 800 mg/l

www.mbr-network.eu

Sewage Treatment
Different steps
Primary treatment - Mainly physical-chemical: grit, primary settling, sand trap, fat trap, flocculation Secondary treatment - Mainly biological (see further) - Physical sludge-water separation: settling, membrane filtration Tertiary treatment - Further purification of effluent (polishing) - Active sand filtration, activated carbon, membrane filtration Sludge treatment - Physical: thickening, dewatering - Biological: sludge digestion
www.mbr-network.eu

Sewage Treatment
A conventional activated sludge WWTP
Primary Settling tanks Grit/sand removal Sludge digester

Influent

Final settlers

Aeration tanks

www.mbr-network.eu

Overall modelling principle: Mass balancing

Mass Balance for compound:

dM d (VC ) = = Qin Cin Qout Cout + rV dt dt

transport
with M: C: V: Q: r:

conversion

Mass of compound in system (g) Concentration of compound (g/m3) Volume of system (m3) Flow rate (m3/h) Volumetric conversion rate (g/m3.h)
www.mbr-network.eu

Sewage treatment: a biological process

A culture of collaborating organisms do the job

www.mbr-network.eu

Sewage Treatment - Biological

Treatment Principles
Biological growth and conversion processes Influences on biological processes

Treatment Processes
COD removal Nitrification Denitrification Phosphate Removal (Biological / Chemical) Anaerobic Digestion

www.mbr-network.eu

Biological growth
Growth = multiplication of organisms Requirements for growth:
nutrients (biomass = C5H7O2N, + P, S, ...) favorable environmental conditions (pH, temperature)

Basic reaction : C-source + NH4 + PO4 + H+ ==> Biomass + electron acceptor (O2, NO3) + byproducts + electron donor (C-source) (H2O, CO2, N2, NO3)

www.mbr-network.eu

Biological conversion
Because biomass grows (or at least wants to), a number of compounds are converted, e.g.
Organic pollutants --> CO2 + waste biomass NH4 --> NO3 NO3 --> N2 PO4 --> Poly-P stored in waste biomass Organic pollutants --> biogas (CH4 + CO2)

How much is converted ?

Rate of the conversion reaction

KINETICS Ratio of conversions of the different compounds STOICHIOMETRY

www.mbr-network.eu

Reaction stoichiometry
Suppose following reaction takes place:
C18H19O9N + O2 + H+ --> C5H7O2N + CO2 + H2O for each molecule of pollutants degraded, a proportional amount of other products will be used (left of arrow) or produced (right of arrow) We can therefore write: a C18H19O9N + b O2 + c H+ --> d C5H7O2N + e CO2 + f H2O a,b,c,d,e,f are called yield or stoichiometric coefficients note that one of the coefficients can be chosen = 1

www.mbr-network.eu

Process kinetics
A reaction will not occur (reaction rate = 0) when its sources (substrates) are absent compounds on the left of the reaction arrow A reaction will have a maximum rate - when all sources are in excess - inhibition by sources/products may affect max. rate

Monod (S)= *(S).S/(KS+S) S

Andrews (S)= *(S).S/(KS+S+S2/KI) S

www.mbr-network.eu

Conversion rates
Take the conversion above a C18H19O9N + b O2 + c H+ Suppose the reaction kinetics: (S)=*.X.S/(KS+S)
Monod kinetics in the substrate first order in the biomass concentration

1 C5H7O2N + d CO2 + e H2O

The conversion of each component is then: C18H19O9N : - a. (S) C5H7O2N O2 : - b. (S) CO2 H+ : - c. (S) H2O

: + 1. (S) : + d. (S) : + e. (S)

www.mbr-network.eu

Conversion rates (contd)

Conversion rate of a compound consists of 3 parts:
sign (+/-) dependent on whether it is used or produced stoichiometric coefficient in the reaction rate of the reaction

What if parallel reactions with same compounds ?

a C18H19O9N + b O2 + c H+ 1 C5H7O2N + d CO2 + e H2O 1 C5H7O2N + i NO3 + j H2O + kH+ f CO2 + g O2 + h NH4+

C5H7O2N, O2, CO2, H+, H2O occur more than once C5H7O2N : + 1. 1 + 1. 2 O2 : - b. 1 - g. 2 CO2 : + d. 1 - f. 2 H+ : - c. 1 + k. 2
www.mbr-network.eu

General conversion model

For the i-th compound, Si:

r(Si) = j sign(ji) vji.j

where j = the rate of the j-th reaction in which Si participates vji = the stoichiometric coefficient for Si in the j-th reaction sign(ji) = sign (+/-) indicating whether Si is substrate or product in the j-th reaction

www.mbr-network.eu

Temperature effect on conversion rate

(T)= (20).ek(T-20)

Rule of thumb: Doubling of reaction rate for temperature increase with 10oC
www.mbr-network.eu

Yearly temperature evolution

Winter period is critical for process performance, especially for nitrification, since this is the slowest
www.mbr-network.eu

pH effect on conversion rates

(pH)= (pHopt).KpH/(KpH -1+10| pH-pHopt | )

Process is changing the system pH by production of H+ (e.g. nitrification, digestion) or OH- (denitrification)
www.mbr-network.eu

Treatment Processes (1)

Aerobic organic substrate removal
in the presence of O2 (aerobic) heterotrophic organisms (i.e. C-source is organic) C18H19O9N + O2 (+ H+) + NH4 --> C5H7O2N + CO2 + H2O high yield (1 g substrate-COD --> 0.4 g biomass-COD)

www.mbr-network.eu

Treatment Processes (2)

Nitrification
in the presence of O2 (aerobic) autotrophic organisms (i.e. C-source is inorganic: CO2) NH4 + CO2 + O2 + --> C5H7O2N + NO3 + H2O + H+ low yield (0.24 g COD/g N oxidised) slow growth rate highly sensitive to lots of disturbances (pH, T, inhibitors) in fact: two-step process (NH4 -> NO2 -> NO3)

www.mbr-network.eu

Treatment Processes (3)

Denitrification
in the absence of O2 (anoxic) in the presence of NO3 and COD heterotrophic organisms C18H19O9N + NO3 + H+ + NH4 --> C5H7O2N + CO2 + H2O

+ N2

relatively high yield (0.3 g biomass-COD/g COD) performs both nitrogen and COD removal ! recuperates O2 invested in nitrification !

www.mbr-network.eu

Treatment Processes (4)

Excess Biological Phosphorus Removal (EBPR)

www.mbr-network.eu

Treatment Processes (4)

Excess Biological Phosphorus Removal (EBPR)
requires a sequence of anaerobic (absence of O2 & NO3 )

and (anoxic or) aerobic conditions COD as volatile fatty acids (acetic acid) heterotrophic organisms VFA + PO4 + NH4 + O2 (+NO3) --> C5H7O2N + CO2 + H2O + Poly-P (+ N2)
relatively high yield (0.3 g biomass-COD/g COD) can perform nitrogen, phosphate and COD removal ! complex process

www.mbr-network.eu

Treatment Processes (5)

Anaerobic Digestion
consortium of anaerobic organisms

(acidogens, methanogens) slow growers delicate balance H2 is an inhibitory intermediate C18H19O9N + NH4 C5H7O2N + CO2 + H2O + CH4 Biogas

www.mbr-network.eu

Treatment Processes (5)

Anaerobic Digestion: acidification step

www.mbr-network.eu

Treatment Processes (5)

Anaerobic Digestion: methanation step

www.mbr-network.eu

Typical activated sludge configurations

Only aerobic reactors aerobic COD removal nitrification (if biomass retention is sufficiently long)

www.mbr-network.eu

Typical layout of N/P-removal ASP

Aerobic, anoxic and anaerobic reactors aerobic COD removal N removal (nitrification and denitrification) Bio-P removal

www.mbr-network.eu

Biofilm reactors

An increasingly popular alternative for ASP

www.mbr-network.eu

Why growth in flocs / biofilms ?

If the residence time of organisms in the process < inverse of their growth rate

WASH OUT By growing in

flocs that settle in the clarifier biofilms that attach to surfaces

they can stay sufficiently long in the system

www.mbr-network.eu

Trickling Filters

Simple, reasonably performing, old technology

www.mbr-network.eu

Trickling Filters

Disadvantages: Clogging + flies Advantages: Cheap aeration

www.mbr-network.eu

Biofilm processes
Conversion + DIFFUSION Principle:
Biofilm Bulk liquid S0
J

L
Substratum

J+dJ/dz*dz

dz

www.mbr-network.eu

Biofilm processes
Active fraction concept
Substrate Conc. L SO r O , rS L

Bulk liquid 0
SS
penetr. depth bSo = active fraction f

zSo

zSo

inactive fraction

www.mbr-network.eu

Biofilm processes
The active fraction concept leads to the interpretation of biofilms as systems in which layers exist with different conversion processes taking place The layers change in size as the process conditions change

www.mbr-network.eu

Sludge treatment

www.mbr-network.eu

Conversion process modelling The starting point:

Henze M., Gujer W., Mino T. and van Loosdrecht M. (2000) Activated Sludge Models ASM1, ASM2, ASM2D & ASM3 Scientific and Technical Report No. 9 IWA Publishing, London.
www.mbr-network.eu

The Gujer matrix presentation

Compact presentation of a biokinetic model (ASM0)

www.mbr-network.eu

Mass balancing
Vertical summation of Stoichiometry term * Kinetics terms gives total conversion

r(Si) = j sign(ji) vji.j

Add the transport terms the mass balance !

www.mbr-network.eu

The Gujer matrix presentation Standard models in use today

Model ASM0 ASM1 ASM2 ASM2d ASM3 ASM3P ADM1 ASDM
#C=

Year 1987 1987 1994 1999 1999 2001 2002 2007

WWTP# C C,N C,N,P C,N,P C,N C,N,P X C,N,P,X

Processes 2 9 20 22 13 24 28 51

Components 3 13 19 19 13 17 36 62

Parameters* 4 19 65 67 36 71 96 470

COD-removal, N= N-removal, P= P-removal, X = Sludge digestion *Parameters: stoichiometric, kinetic & composition parameters

www.mbr-network.eu

The Gujer matrix presentation

ASDMs model matrix

www.mbr-network.eu

Gujers law
Model complexity measure:
Complexity = #components x #processes x #compartments
1010 108 106 104 100 1 1960 ASM Simulation Slope 0.35 yr-1 1970 1980 1990 2000 2010 Year
www.mbr-network.eu

Transistors per square inch Moores law Slope 0.35 yr-1 Relative model complexity C

Activated Sludge Model No 1

Henze et al. (1987) Innovations:
Nomenclature: Solubles: symbol S Focus on:

Particulates: symbol X

- Sludge production - Oxygen consumption - Nitrogen removal Mass balancing

COD based modelling Petersen matrix

www.mbr-network.eu

ASM1: COD-components
Total COD

Biodegradable COD

Nonbiodeg. COD

Active mass COD

Soluble SS

Particulate Xs

Heterotrophs XB,H

Autotrophs XB,A

Soluble SI

Particulate XI & XP
www.mbr-network.eu

ASM1: N-components
Nitrate/nitrite N SNO Total Kjeldahl N TKN

Free & saline ammonia SNH

Organically bound N

Active mass N XNB

Soluble organic N

Particulate organic N

Nonbiodeg. N SNI

Biodeg. N SND

Biodeg. N XND

Nonbiodeg. N XNI & XNP

www.mbr-network.eu

ASM1: Processes
1) Growth of biomass
heterotrophs - aerobic - anoxic autotrophs (nitrification)

2) Decay of biomass
heterotrophs autotrophs

3) Ammonification of organic nitrogen 4) Hydrolysis of particulate organic matter

www.mbr-network.eu

SO
Growth

SNH SNO

Flow of COD in ASM1

XA
Decay

XI
Hydrolysis

XS SS SO
Growth

Decay

XH
www.mbr-network.eu

Continuity check
Horizontal summation of stoichiometric coefficients should equal 0 !

i vji.ici = 0
Provided: - consistent units have been used - all substrates/products are included

This can be done for COD, N, P, Charge, Mass Example: ASM3 !

www.mbr-network.eu

Component i > j 1 2 3 4 5 6 7 8 9 Process Hydrolysis

1 SO expressed as > O2

2 3 4 SI SS SNH COD COD N fSI 1-fSI -1 -1 y1 y2 y3 y4 -iNBM y6 y6

5 SN2 N

6 7 8 9 10 11 12 13 SNO SHCO XI XS XH XSTO XA XTS N Mole COD COD COD COD COD TSS z1 z2 z3 z4 z5 z6 z7 z9 z10 z11 z12 -1 YSTO YSTO -1/YH -1/YH -iXS t2 t3 t4 t5 t6 t7 -0.60 -0.60 1 -1 -1 iTSBM t11 t12

Heterotrophic organisms, denitrification Aerobic storage of COD x2 Anoxic storage of COD Aerobic growth x4 Anoxic growth (denitrification) Aerobic endog. respiration -(1-fI) Anoxic endog. respiration Aerobic respiration of PHA -1 Anoxic respiration of PHA x8 -(1-fI) -x3 -x5 -x7 -x9 y10 y11 y12 x3 x5 x7 x9 1/YA -y12 y12

fI fI

1 1 -1 -1

-1 -1

Autotrophic organisms, nitrification 10 Nitrification 11 Aerobic endog. respiration 12 Anoxic endog. respiration Composition matrix k,I k 1 2 3 4 Conservatives COD Nitrogen Ionic charge Observables TSS g COD gN Mole + g TSS -1 1 iNSI 1 iNSS -1.71 -4.57 1 1 1 1/14 -1/14 1 iNXI -1 iTSXI iTSXS iTSBM 0.60 iTSBM 1 iNXS 1 iNBM 1 1 iNBM fI fI

www.mbr-network.eu

Consistency Check ASM3

Conservation eq.:

j,i

k ,i = 0 for i = 1 to 12

for k = COD, N and ionic charge yields j k = 36 equations which allow to easily predict all xj, yj, zj TSS Composition eq.:

t j = j,TSS =

j,i

TSS,i for i = 8 to 12

yields j = 12 equations which allow to easily predict all tj

www.mbr-network.eu

Overall modelling principle: Mass balancing

Mass Balance for compound:

dM d (VC ) = = Qin Cin Qout Cout + rV dt dt

transport
with M: C: V: Q: r:

conversion

Mass of compound in system (g) Concentration of compound (g/m3) Volume of system (m3) Flow rate (m3/h) Volumetric conversion rate (g/m3.h)
www.mbr-network.eu