Silva (2016)

UNIVERSIDADE FEDERAL DA BAHIA
ESCOLA POLITÉCNICA
PROGRAMA DE PÓS-GRADUAÇÃO EM ENGENHARIA INDUSTRIAL
CARLOS EDUARDO PEREIRA MENDES DA SILVA
MODELAGEM MATEMÁTICA E SIMULAÇÃO DA

REMOÇÃO SIMULTÂNEA DE CARBONO E NITROGÊNIO
EM REATOR ANAERÓBIO-ANÓXICO:
UMA APLICAÇÃO DO ADM1
Salvador
2016
CARLOS EDUARDO PEREIRA MENDES DA SILVA
MODELAGEM MATEMÁTICA E SIMULAÇÃO DA

REMOÇÃO SIMULTÂNEA DE CARBONO E NITROGÊNIO
EM REATOR ANAERÓBIO-ANÓXICO:
UMA APLICAÇÃO DO ADM1
Tese apresentada ao programa de Pós-Graduação

em Engenharia Industrial, Universidade Fe-
deral da Bahia, como requisito parcial para
obtenção do grau de Doutor em Engenharia
Industrial.
Orientadores:
Profa . Dra . Karla Patrı́cia S. O. R. Esquerre
Prof. Dr. Luciano Matos Queiroz
Salvador
2016
S586 Silva, Carlos Eduardo Pereira Mendes da.
Modelagem matemática e simulação da remoção
simultânea de carbono e nitrogênio em reator anaeróbio-
anóxico: uma aplicação do ADM1 / Carlos Eduardo Pereira
Mendes da Silva. – Salvador, 2015.
139 f. : il. color.
Orientadores: Profa. Dra..Karla Patrícia S. O. R. Esquerre;

Prof. Dr. Luciano Matos Queiroz.
Tese (doutorado) – Universidade Federal da Bahia. Escola

Politécnica, 2015.
1. Modelagem matemática. 2. Desnitrificação. 3. Carbono -

Remoção. 4. Nitrogênio – Remoção. 5. ADM1. I. Esquerre,
Karla Patrícia S. O. R. II. Queiroz, Luciano Matos. III.
Universidade Federal da Bahia. III. Título.
CDD.: 511.8
Dedico esta tese à minha famı́lia, em especial à minha esposa Carol e aos meus pais
Eduardo e Paixão
Agradecimentos
O doutorado possibilitou o aprofundamento do conhecimento adquirido na graduação em Engenha-

ria Ambiental, amadurecimento como pesquisador, assim como o contato com novas áreas do conhe-
cimento. Muitas pessoas contribuíram para a superação dos desafios, assim como para a realização
das atividades do dia a dia.
Antes de tudo, gostaria de registrar meus sinceros agradecimentos ao Deus eterno imortal, invisível,
mas real. Criador dos céus, da terra e tudo que nela há, a saber, Jesus Cristo. Agradeço a Ele por
todas as bênçãos em mim derramadas e pelo constante auxílio em tempos de tribulação.
À minha amada esposa, Carolina Mendes, por ter sido paciente durante o tempo em que precisava
para dedicar a pesquisa. Sem seu amor, apoio, carinho, cumplicidade, incentivo e orações esta
pesquisa não teria sido possível.
Aos meus pais, Eduardo Mendes da Silva e Maria da Paixão da Silva, pelo amor incondicional, pelo
carinho e preocupação constante, ensinando-me os valores de respeito e de responsabilidade. Por
estimular-me a seguir no doutorado e por sempre acreditar no meu potencial e mostrá-lo para mim.
Por simplesmente serem meus pais e fornecerem todos os subsídios para minha formação.
Ao meu irmão, Alexandre Mendes, pela amizade, pelo companheirismo, pelo incentivo e pela torcida
nesta tarefa tão árdua de execução da tese.
Aos meus orientadores e também meus amigos, professora Dra . Karla Esquerre e o professor Dr.
Luciano Matos, por sua paciência, dedicação e atenção dada ao longo desses anos de convivência.
Pelos ensinamentos e contribuições para o meu crescimento intelectual.
À minha Tia Kátia e toda sua família pela hospedagem em sua residência e pelo cuidado durante o
período do doutorado sanduíche na Universidade de São Paulo.
Aos membros da banca examinadora de qualificação e defesa, professores Dr. Mario Kato, Dr. Theo
Syrto Octavio de Souza, Dr. Cláudio Leite de Souza e Dr. Márcio André Fernandes Martins, pelos
pertinentes apontamentos que engradeceram essa pesquisa.
Ao professor Dr. Márcio Martins, pelo exemplo de pesquisador, amizade, solidariedade e importan-
tes discussões e sugestões no desenvolvimento desta pesquisa.
Aos professores Dr. Damien Batstone do Advanced Wastewater Management Centre da Univer-
sidade de Queensland/Austrália e Dr. César Huiliñir Universidad de Concepción, Chile pela opor-
tunidade de aprendizado no tema de modelagem matemática de digestão anaeróbia e importantes
sugestões no desenvolvimento desta pesquisa.
Aos colegas do Programa de Pós-Graduação em Engenharia Industrial da Escola Politécnica -

UFBA, pelo convívio durante estes anos e por manterem o ambiente propício à pesquisa.
A Universidade Federal da Bahia, por fornecer infraestrutura para o desenvolvimento de pesquisas,

aprendizado e lazer, e por ser meu “lar” durante este tempo.
A Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (Capes) por conceder auxílio finan-
ceiro, possibilitando a formação de mão-de-obra altamente qualificada, esforçando-se em colocar
e manter o Brasil como gerador e desenvolvedor de ciência e tecnologia, um dos passos rumo ao
desenvolvimento e reconhecimento da nossa nação.
A todos aqueles que contribuíram, direta ou indiretamente, para realização desta pesquisa.
“Eu sou a Sabedoria! Em mim habita todo o conhecimento, o discernimento e o ensino”
Bı́blia Sagrada - Provérbios 8:12
“Ó profundidade da riqueza, da sabedoria e do conhecimento de Deus! Quão insondáveis

são os seus juı́zos, e quão inescrutáveis os seus caminhos! ”
Bı́blia Sagrada - Romanos 11:33

Resumo
Esta tese tem como objetivo principal avaliar a remoção simultânea de carbono e nitrogênio
(RSCN) em reator anaeróbio-anóxico por meio da modelagem matemática e simulação. Na
primeira etapa da pesquisa, o Anaerobic Digestion Model No. 1 (ADM1) foi implementado
no software Matlab/Simulink e ajustado aos dados experimentais obtidos a partir de um
digestor anaeróbio utilizado para estabilização do lodo de esgoto doméstico. Os resultados
mostram uma boa concordância entre os dados experimentais e os simulados para as con-
centração de propionato, acetato, produção e composição de biogás e pH, corroborando
a potencialidade de utilização do modelo para prever o comportamento dos processos
biológicos de digestão anaeróbia. Na segunda etapa, foi desenvolvida uma extensão do
ADM1 incluindo o processo de desnitrificação, para simular o efeito do tempo de detenção
hidráulica (TDH) e da relação DQO/NO−
3 na RSCN em um reator anaeróbio-anóxico
tratando águas residuárias doméstica. O modelo foi calibrado com dados experimentais
previamente publicados e obtidos a partir de ensaios em bateladas. Os experimentos foram
conduzidos em reatores operados em bateladas, alimentados com substrato sintético (1500
mg DQO.L−1 ) e inoculados com lodo floculento (500 mg SSV.L−1 ) proveniente de um rea-
tor UASB. Diferentes concentrações de nitrato de potássio foram adicionadas nos reatores
a fim de obter relações de DQO/NO−
3 de 40 e 150, respectivamente. Adicionalmente, foram
realizadas simulações para avaliar a RSCN em um reator anaeróbio-anóxico operado em

regime contı́nuo considerando 6 cenários diferentes. Os cenários propostos foram baseados
na caracterização afluente tı́pica de águas residuárias doméstica. O modelo foi calibrado
satisfatoriamente e os resultados simulados revelaram que o melhor cenário para ocorrência
da RSCN foi com a relação DQO/NO−
3 e TDH iguais a 10 e 15 horas, respectivamente.
Por fim na terceira etapa, o modelo ADM1 foi adaptado para avaliar a RSCN em um
reator anaeróbio-anóxico tratando águas residuárias sintética simulando efluente de pro-
cessamento de pescado. O modelo foi calibrado e validado a partir de dados experimentais
obtidos em ensaios em bateladas. Os resultados simulados mostraram boa aderência aos
dados experimentais, com um erro médio absoluto variando entre 15 e 38 % na calibração e
19 e 36 % na validação. Os resultados obtidos nessa pesquisa mostraram que as adaptações
do ADM1 para RSCN de águas residuárias de origens doméstica e de processamento de
pescado podem ser utilizadas como uma ferramenta para melhor compreensão da RSCN,
permitindo a definição de melhores condições de operação e processo.
Palavras-chave: Anaerobic digestion model - ADM1, Desnitrificação, Remoção simultânea
de carbono e nitrogênio, Simulação, Modelagem matemática.
Abstract
The aim of this thesis is to evaluate the simultaneous carbon and nitrogen removal (SCNR)
in anaerobic-anoxic reactors through mathematical modeling and simulation. In the first
stage of the research, the Anaerobic Digestion Model No. 1 (ADM1) model was coded
using the Matlab/Simulink and adjusted to the experimental data obtained from an ana-
erobic digester used for the stabilization of sewage sludge. The simulation results indicate
good agreement with the experimental data of effluent propionate, acetate, composites
and biogas flows and pH, thus corroborating the potential use of this model to predict the
behavior of anaerobic digestion processes. In the second stage, ADM1 was extended by
the addition of the denitrification process to simulate the effects of nitrate concentration
and hydraulic retention time (HRT) on SCNR in anaerobic-anoxic reactor treating domes-
tic wastewater. The model was calibrated using previously published experimental data
obtained from anaerobic batch tests. The experiments were conducted in batch reactors
fed with synthetic substrate (1500 mg COD.L−1 ) and inoculated with flocculent sludge
(500 mg VSS.L−1 ) from a UASB reactor. Different amounts of potassium nitrate were ad-
ded into the reactors to achieve COD/NO−
3 ratios 40 and 150, respectively. Furthermore,
model simulations were performed to evaluate the SCNR in a continuous stirred tank re-
actor (CSTR) operating under mesophilic conditions considering six different scenarios.
Each scenarios was evaluated to investigate the performance of the SCNR based on typical
effluent characteristics of domestic wastewater. The model was calibrated satisfactorily
and the simulated results indicate that the best scenario for the occurrence of SCNR was
with COD/NO−
3 ratio and HRT equal to 10 and 15 hours, respectively. Finally in the
third stage, the ADM1 model was adapted for SCNR in an anaerobic-anoxic reactor tre-
ating synthetic wastewater simulating fish processing effluent. The model was calibrated
and validated from experimental data obtained in tests on batch. The simulated results
showed good agreement with experimental data, with a mean absolute error between 15
and 38% in the calibration and 19 and 36% in the validation. The results obtained in
this research showed that ADM1 adaptations for SCNR domestic wastewater and salmon
fishery wastewater can be used as a tool to better understand the SCNR and to define
best operating and process conditions.
Keywords: Anaerobic digestion model - ADM1, Denitrification, Simultaneous carbon and
nitrogen removal, Simulation, Mathematical modeling.
Lista de Figuras
2.1 Sequência de processos metabólicos durante a digestão anaeróbia. Fonte:

Adaptado de Batstone et al. (2015) . . . . . . . . . . . . . . . . . . . . . . . 34
2.2 Inter-relação da digestão anaeróbia com o ciclo do nitrogênio . . . . . . . 39
2.3 Ciclo biológico do nitrogênio. Adaptado de Metcalf e Eddy (2007) e Henze
et al. (2008). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.4 Variação do número de oxidação do nitrogênio. Adaptado de Van Haandel
(1999). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.5 Diagrama esquemático: (a) método convencional (sistema A2 O); e (b) método
alternativo (RSCN) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.6 Modelo ADM1 Fonte: Adaptado de Batstone et al. (2002a) . . . . . . . . . . 55
3.1 Estimation procedure to identify most sensitive parameters. . . . . . . . . 74

3.2 Sensitivity indices (SI) calculated for different ranges of (a) hydrolysis cons-
tant rates; (b) Monod maximum specific uptake rates; (c) half-saturation
constants; and (d) yields of biomass on substrate. . . . . . . . . . . . . . . 75
3.3 Comparison between simulations and experimental data with and without
optimization: (a) acetate and (b) propionate. . . . . . . . . . . . . . . . . . 76
3.4 Comparison between simulations and experimental data after parameter
estimation: (a) biogas production, (b) biogas composition and (c) pH . . . 77
3.5 Model simulations for a continuous flow system with shock loading of 5
kg/m3 day experiencing a step increase in organic loading rate to 35 kg/m3
day on day 30 and to 70 kg/m3 day on day 60. (a): pH and (b) Biogas
composite (c) VFAs (18 days HRT). . . . . . . . . . . . . . . . . . . . . . . 79
4.1 Conceptual model of the simultaneous AD/DN processes of domestic was-

tewater. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.2 Schematic representation of the simplified COD fractionation. Percentages
correspond to COD fractions immediately before each compound. . . . . . 90
13
4.3 Model calibration for a batch reactor inoculated with strict methanogenic
sludge after the addition of different concentrations of nitrate 0 (control),
10 and 37.5 mg N/L: (a) methane production; (b) production/consumption
profiles of acetate; (c) consumption of nitrate; and (d) (DP) and (MP) % of
COD used for the denitrification and methanogenic process, respectively. . 94
4.4 Results of the model simulation for continuous flow systems treating domes-
tic wastewater: (a) removal COD; (b) methane production; and (c) VFA . 95
tic wastewater: (a) consumption of nitrate; (b) consumption of nitrite; and
(c) nitrogen gas production . . . . . . . . . . . . . . . . . . . . . . . . . . 96
tic wastewater: (a) pH; (b) (DP) % of COD used for denitrification; and (c)
(MP) % of COD used for methanogenic. . . . . . . . . . . . . . . . . . . . 97
4.7 Results of the model simulation for continuous flow systems treating strong
domestic wastewater experiencing three step decrements in the HRT from
15 to 4 hours: (a) removal COD; (b) methane production and (c) VFA . . 99
15 to 4 hours: (a) consumption of nitrate; (b) consumption of nitrite; and
(c) nitrogen gas production . . . . . . . . . . . . . . . . . . . . . . . . . . 100
15 to 4 hours: (a) pH; (b) (DP) % of COD used for denitrification; and (c)
(MP) % of COD used for methanogenic . . . . . . . . . . . . . . . . . . . . 101
5.1 Conceptual model of the simultaneous AD/DN processes of salmon fishery

wastewater. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.2 Sensitivity indices (SI) of the organic matter calculated for different ranges
of (a) hydrolysis constant rate; (b) Monod maximum specific uptake rates;
(c) half-saturation constants; and (d) parameters denitrification . . . . . . 113
5.3 Sensitivity indices (SI) of the nitrate reduction calculated for different ranges
of (a) hydrolysis constant rate; (b) Monod maximum specific uptake rates;
(c) half-saturation constants; and (d) parameters denitrification . . . . . . 114
5.4 Comparison between simulations and experimental data for D1 set after
calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5.5 Validation of simulations with experimental results of D2 set . . . . . . . 117
Lista de Tabelas
2.1 Reações do nitrogênio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

2.2 Matriz estequiométrica para compostos solúveis (i = 1-12, j = 1-19) . . . . 52
2.3 Matriz estequiométrica para compostos solúveis (i = 13-26, j = 1-19) . . . 53
2.5 Processos e taxas de reações correspondentes do ADM1 . . . . . . . . . . . 54
2.6 Extensões e aplicações do ADM1. . . . . . . . . . . . . . . . . . . . . . . . 61
2.7 Comparação da complexidade entre o ADM1 e o modelo simplificado (Garca-
Diéguez et al., 2013) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.1 Composition of the wastewater sludge . . . . . . . . . . . . . . . . . . . . . 71

3.2 Model inputs used for simulations . . . . . . . . . . . . . . . . . . . . . . . 72
3.3 Initial and estimated values of kinetic parameters . . . . . . . . . . . . . . 75
4.1 Disintegration and hydrolysis kinetic constants . . . . . . . . . . . . . . . . 85

4.2 Denitrification process included in the modified Anaerobic Digestion Model
no1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.3 Different scenarios to simulate the effects of nitrate concentration on the
SCNR in anaerobic-anoxic reactor for weak domestic wastewater (S1), me-
dium domestic wastewater (S2), and strong domestic wastewater (S3 to S6) 92
5.1 Denitrification process included in the modified Anaerobic Digestion Model

no1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5.2 Initial and estimated kinetic parameters of the implemented extended ADM1.114
16
Lista de Abreviaturas e Siglas
ACP Análise de Componentes Principais

ADM1 Anaerobic Digestion Model no 1
AGV Ácidos Graxos Voláteis
ASM1 Anaerobic Sludge Model 1
COD Chemical Oxygen Demand
CSTR Continuous Stirred Tank Reactor.
DBO Demanda Bioquı́mica de Oxigênio
DNRA Dissimilatory Nitrate Reduction to Ammonia
DQO Demanda Quı́mica de Oxigênio
EDO Equações Diferenciais Ordinárias
EGSB Expanded Granular Sludge Bed Reactor
HRT Hydraulic Retention Time
IWA International Water Association
LCFA Ácidos Graxos de Cadeia Longa
pH Potencial Hidrogeniônico
RSCN Remoção Simultânea de Carbono e Nitrogênio
SCNR Simultaneous Carbon and Nitrogen Removal
SRT Solids Retention Time
SSV Sólidos Suspensos Voláteis
TCO Total Organic Carbon
TDH Tempo de Detenção Hidráulica
UASB Upflow Anaerobic Sludge Blanket
VFA Volatile fatty acids
VSS Volatile Suspended Solids
Lista de Sı́mbolos
Csu Conteúdo de carbono presente nos monossacarı́deos [kmolC

kgDQO−1 ]
Caa Conteúdo de carbono presente nos aminoácidos [kmolC kgDQO−1 ]
Cf a Conteúdo de carbono presente nos ácidos graxos [kmolC kgDQO−1 ]
Cva Conteúdo de carbono presente no valerato [kmolC kgDQO−1 ]
Cbu Conteúdo de carbono presente no butirato [kmolC kgDQO−1 ]
Cpro Conteúdo de carbono presente no propionato [kmolC kgDQO−1 ]
Cac Conteúdo de carbono presente no acetato [kmolC kgDQO−1 ]
Cch4 Conteúdo de carbono presente no metano [kmolC kgDQO−1 ]
CsI Conteúdo de carbono presente nos inertes solúveis [kmolC
kgDQO−1 ]
Cxc Conteúdo de carbono presente nos compostos particulados [kmolC
kgDQO−1 ]
Cch Conteúdo de carbono presente nos carboidratos [kmolC kgDQO−1 ]
Cpr Conteúdo de carbono presente nas preteı́nas [kmolC kgDQO−1 ]
Cli Conteúdo de carbono presente nos lipı́deos [kmolC kgDQO−1 ]
Cbac Conteúdo de carbono presente nos microrganismos [kmolC
kgDQO−1 ]
CxI Conteúdo de carbono presente nos inertes particulados [kmolC
kgDQO−1 ]
fSi,xc Parcela de material inerte solúvel nos compostos [kgDQO
kgDQO−1 ]
fxI,xc Parcela de material inerte particulado nos compostos [kgDQO
kgDQO−1 ]
fbu,su Parcela de butirato proveniente dos monossacarı́deos [kgDQO
kgDQO−1 ]
fli,xc Parcela de lipı́deos nos compostos [kgDQO kgDQO−1 ]
fh2,su Parcela de hidrogênio proveniente dos monossacarı́deos [kDQO
kgDQO−1 ]
fpr,xc Parcela de proteı́nas nos compostos [kgDQO kgDQO−1 ]
fchxc Parcela de carboidratos nos compostos [kgDQO kgDQO−1 ]
ff a,li Parcela de ácidos graxos proveniente dos lipı́deos [kgDQO
kgDQO−1 ]
fh2,aa Parcela de hidrogênio proveniente dos aminoácidos [kgDQO
kgDQO−1 ]
fva,aa Parcela de valerato proveniente dos aminoácidos [kgDQO
kgDQO−1 ]
fpro,aa Parcela de propionato proveniente dos aminoácidos [kDQO
kgDQO−1 ]
fac,aa Parcela de acetato proveniente dos aminoácidos [kgDQO kgDQO−1 ]
fpro,su Parcela de propionato proveniente dos monossacarı́deos [kgDQO
kgDQO−1 ]
fac,su Parcela de acetato proveniente dos monossacarı́deos [adimensional]
fbu,aa Parcela de butirato proveniente dos aminoácidos [kgDQO
kgDQO−1 ]
kdis Constante cinética de 1a ordem para processo de desintegração do
composto particulado [d−1 ]
khyd,pr Constante cinética de 1a ordem para processo de hidrólise de
proteı́nas [d−1 ]
khyd,li Constante cinética de 1a ordem para processo de hidrólise de
lipı́deos [d−1 ]
kd Coeficiente de decaimento [d−1 ]
kdec,Xsu Constante cinética de 1a ordem para processo de decaimento celular
(monossacarı́deos) [d−1 ]
kdecXaa Constante cinética de 1a ordem para processo de decaimento celular
(aminoácidos) [d−1 ]
kdec,Xf a Constante cinética de 1a ordem para processo de decaimento celular
(ácidos graxos) [d−1 ]
kdec,Xc4 Constante cinética de 1a ordem para processo de decaimento celular
(degradadores de monossacarı́deos) [d−1 ]
kdec,Xpro Constante cinética de 1a ordem para processo de decaimento celular
(propionato) [d−1 ]
kdec,Xac Constante cinética de 1a ordem para processo de decaimento celular
(acetato) [d−1 ]
kdec,Xh2 Constante cinética de 1a ordem para processo de decaimento celular
(gás hidrogênio) [d−1 ]
KS,SIN Constante de meia saturação [kgDQO m−3 ]
KS Constante de saturação de substrato [kgDQO m−3 ]
km,su Velocidade máxima especı́fica de consumo de monossacarı́deos [d−1 ]
KS,su Constante de meia saturação para o consumo de monossacarı́deos
[kgDQO m−3 ]
km,aa Velocidade máxima especı́fica de consumo de aminoácidos [d−1 ]
KS,aa Constante de meia saturação [kgDQO m−3 ]
km,f a Velocidade máxima especı́fica de consumo de ácidos graxos de ca-
deia longa [d−1 ]
KS,f a Constante de meia saturação para o consumo de ácidos graxos de
cadeia longa [kgDQO m−3 ]
KIh2,f a Concentração de ácidos graxos que inibem a produção de hidrogênio
[kgDQO m−3 ]
km,c4 Velocidade máxima especı́fica de consumo de valerato e butirato
[d−1 ]
KS,c4 Constante de meia saturação para o consumo de valerato e butirato
[kgDQO m−3 ]
KIh2,c4 Concentração de ácidos graxos que inibem a produção de hidrogênio
[kgDQO m−3 ]
km,pro Velocidade máxima especı́fica de consumo de propionato [d−1 ]
KS,pro Constante de meia saturação para o consumo de propionato
[kgDQO m−3 ]
KIh2,pro Concentração de ácidos graxos que inibem a produção de hidrogênio
[kgDQO m−3 ]
km,ac Velocidade máxima especı́fica de consumo de acetato [d−1 ]
KS,ac Constante de meia saturação para o consumo de acetato [kgDQO
m−3 ]
KInh3 Concentração de ácidos graxos que inibem a produção de hidrogênio
[kgDQO m−3 ]
km,h2 Velocidade máxima especı́fica de consumo de hidrogênio [d−1 ]
KS,h2 Constante de meia saturação para o consumo de hidrogênio [kg
DQO m−3 ]
KH,co2 Coeficiente da Lei de Henry [kmol m−3 bar−1 ]
KH,ch4 Coeficiente da Lei de Henry [kmol m−3 bar−1 ]
KH,h2 Coeficiente da Lei de Henry [kmol m−3 bar−1 ]
kL a Coeficiente de transferência gás-lı́quido [d−1 ]
pHU L,aa Limite superior da inibição por pH no consumo de aminoácidos
[adimensional]
pHLL,aa Limite inferior da inibição por pH no consumo de aminoácidos [adi-
mensional]
pHU L,ac Limite superior da inibição por pH no consumo de acetato [adimen-
sional]
pHLL,ac Limite inferior da inibição por pH no consumo de acetato [adimen-
sional]
pHU L,h2 Limite superior da inibição por pH no consumo de hidrogênio [adi-
mensional]
pHLL,h2 Limite inferior da inibição por pH no consumo de hidrogênio [adi-
mensional]
Pgas,h2 Pressão parcial de gás hidrogênio [bar]
Pgas,ch4 Pressão parcial de gás metano [bar]
Pgas,co2 Pressão parcial de gás carbônico [bar]
Patm Pressão atmosférica [bar]
pgas,h2o Pressão de vapor de água [bar]
Ssu Concentração de monossacarı́deos na forma solúvel [kgDQO m−3 ]
Saa Concentração de aminoácidos na forma solúvel [kgDQO m−3 ]
Sf a Concentração de ácidos graxos na forma solúvel [kgDQO m−3 ]
Sva Concentração de ácido valérico na forma solúvel [kgDQO m−3 ]
Sbu Concentração de ácido butı́rico na forma solúvel [kgDQO m−3 ]
Spro Concentração de ácido propiônico na forma solúvel [kgDQO m−3 ]
Sac Concentração de ácido acético na forma solúvel [kgDQO m−3 ]
Sh2 Concentração de gás hidrogênio na forma lı́quida [kgDQO m−3 ]
Sch4 Concentração de gás metano na forma lı́quida [kgDQO m−3 ]
SIC Concentração de carbono inorgânico [kgDQO m−3 ]
SIN Concentração de nitrogênio inorgânico [kgDQO m−3 ]
SI Concentração de inertes [kgDQO m−3 ]
Y Coeficiente de conversão de biomassa [kgDQO X kgDQO S −1 ]
Yaa Produção de biomassa a partir de aminoácidos [kgDQO X kgDQO
S −1 ]
Yf a Produção de biomassa a partir de ácidos graxos de cadeia longa
[kgDQO X kgDQO S −1 ]
Ysu Produção de biomassa a partir de monossacarı́deos [kgDQO X
kgDQO S −1 ]
Yc4 Produção de biomassa a partir de valerato e butirato [kgDQO X
kgDQO S −1 ]
Ypro Produção de biomassa a partir de propionato [kgDQO X kgDQO
S −1 ]
Yac Produção de biomassa a partir de acetato [kgDQO X kgDQO S −1 ]
Yh2 Produção de biomassa a partir de hidrogênio [kgDQO X kgDQO
S −1 ]
Scat Concentração equivalente total de cátions [kmol m−3 ]
San Concentração equivalente total de ânions [kmol m−3 ]
SN H4 Concentração de ı́on amônio [kmolN m−3 ]
SH + Concentração de ı́on hidrogênio [kmolH+ m−3 ]
SHCO3− Concentração de ı́on carbonato [kmolC m−3 ]
Sac− Concentração de ı́on acetato [kgDQO m−3 ]
Spro− Concentração de ı́on propionato [kgDQO m−3 ]
Sbu− Concentração de ı́on butirato [kgDQO m−3 ]
Sva− Concentração de ı́on valerato [kgDQO m−3 ]
SOH − Concentração de ı́on hidróxido [kgDQO m−3 ]
SCO2 Concentração de gás carbônico [kgDQO m−3 ]
SN H 3 Concentração de amônia [kgDQO m−3 ]
SIC Concentração de carbono inorgânico [kgDQO m−3 ]
SIN Concentração de nitrogênio inorgânico [kgDQO m−3 ]
Xc Concentração de material particulado [kgDQO m−3 ]
Xch Concentração de carboidratos na forma particulada [kgDQO m−3 ]
Xpr Concentração de proteı́nas na forma particulada [kgDQO m−3 ]
Xli Concentração de lipı́deos na forma particulada [kgDQO m−3 ]
Xsu Concentração de decompositores de monossacarı́deos [kgDQO m−3 ]
Xaa Concentração de decompositores de aminoácidos [kgDQO m−3 ]
Xf a Concentração de decompositores de ácidos graxos de cadeia longa
[kgDQO m−3 ]
Xc4 Concentração de decompositores de valerato e butirato [kgDQO
m−3 ]
Xpro Concentração de decompositores de propionato [kgDQO m−3 ]
Xac Concentração de decompositores de acetato [kgDQO m−3 ]
Xh2 Concentração de decompositores de hidrogênio [kgDQO m−3 ]
Xi Concentração de inertes na forma particulada [kgDQO m−3 ]
Lista de Publicações
Principais publicações do autor durante o doutoramento.
Carlos Mendes, Leonardo de Souza, Ricardo Kalid, Karla Esquerre, Asher Kiperstok.
Assessment of the uncertainty associated with the energy indicator. Renewable and Sus-
tainable Energy Reviews, 15, 3156 - 3164, 2011. (Fator de impacto 5.627 - Qualis A1 da
CAPES Engenharias III).
Carlos Mendes, Márcio A.F. Martins, Karla Esquerre. Evaluating the uncertainty associ-
ated with the industrial effluent generation indicator and its impact in a decision-making
process. O artigo deve ser submetido para o Journal of Cleaner Production. (Fator de
impacto 4.16 - Qualis A1 da CAPES Engenharias III).
Carlos Mendes, Robson da Silva Magalhães, Karla Esquerre, Luciano Matos Queiroz. Ar-
tificial neural network modeling for predicting organic matter in a full-scale UASB reactor.
Environmental Modeling and Assessment, 20, 625 - 635, 2015. (Fator de impacto 0.980 -
Qualis B1 da CAPES Engenharias III).
Publicações associadas à tese

Carlos Mendes, Karla Esquerre, Luciano Matos Queiroz, Application of Anaerobic Diges-
tion Model No. 1 for simulating anaerobic mesophilic sludge digestion. Waste Manage-
ment, 35, 89 - 95, 2015. (Fator de impacto 3.522 - Qualis A2 da CAPES Engenharias
III).
Carlos Mendes, Karla Esquerre, Luciano Matos Queiroz, Modeling simultaneous carbon
and nitrogen removal in an anaerobic-anoxic reactor treating domestic wastewater. O
artigo encontra-se em revisão no Journal Environmental Management. (Fator de impacto
3.895 - Qualis A2 da CAPES Engenharias III).
Carlos Mendes, Karla Esquerre, Luciano Matos Queiroz, Cesar Huiliñir. Modeling of
simultaneous carbon and nitrogen removal in an anaerobic-anoxic reactor treating salmon
fishery wastewater. O artigo deve ser submetido para o Water Research. (Fator de impacto
6.279 - Qualis A1 da CAPES Engenharias III).
24
Sumário
1. Introdução 29
1.1 Motivação e justificativa . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.2 Objetivos da tese . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.3 Estrutura da tese . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2. Revisão da literatura 33
2.1 Processos biológicos de digestão anaeróbia . . . . . . . . . . . . . . . . . . 33
2.1.1 Fermentação e metanogênese . . . . . . . . . . . . . . . . . . . . . . 33
2.1.2 Redução de nitrato . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.2 Processo convencional de remoção biológica de nitrogênio . . . . . . . . . . 40
2.3 Remoção simultânea de carbono e nitrogênio em sistema integrado anaeróbio-
anóxico . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.4 Anaerobic Digestion Model No. 1 (ADM1) . . . . . . . . . . . . . . . . . . 49
2.4.1 Processos bioquı́micos . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.4.1.1 Modelo de inibição e toxicidade . . . . . . . . . . . . . . . 56
2.4.2 Processos fı́sico-quı́micos . . . . . . . . . . . . . . . . . . . . . . . . 56
2.4.2.1 Processo lı́quido-lı́quido . . . . . . . . . . . . . . . . . . . 57
2.4.2.2 Processos lı́quido-gás . . . . . . . . . . . . . . . . . . . . . 57
2.4.3 Processos excluı́dos do ADM1 . . . . . . . . . . . . . . . . . . . . . 58
2.4.4 Implementação do ADM1 em um reator CSTR . . . . . . . . . . . 58
2.4.4.1 Modelagem da fase lı́quida . . . . . . . . . . . . . . . . . . 58
2.4.4.2 Modelagem da fase gasosa . . . . . . . . . . . . . . . . . . 59
2.4.4.3 Estimativa do pH . . . . . . . . . . . . . . . . . . . . . . . 60
2.5 Extensões do modelo ADM1 . . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.5.1 Extensão do ADM1 para processo de desnitrificação . . . . . . . . . 62
2.6 Metodologias e estratégias para aplicação do modelo ADM1 . . . . . . . . 63
2.6.1 Simplificações do ADM1 . . . . . . . . . . . . . . . . . . . . . . . . 64
2.6.2 Métodos para caracterização do afluente . . . . . . . . . . . . . . . 65
2.7 Considerações finais . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3. Application of Anaerobic Digestion Model No. 1 for simulating anae-

robic mesophilic sludge digestion 69
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.2.1 Experimental data . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.2.2 Input characterization: COD fractioning . . . . . . . . . . . . . . . 71
3.2.3 Implementation of ADM1 in Matlab/Simulink . . . . . . . . . . . . 71
3.2.4 Sensitivity analysis and parameter estimation . . . . . . . . . . . . 72
3.2.5 Effect of organic shock loads on the stability of the anaerobic reactor 73
3.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.3.1 Sensitivity analysis and parameter estimation . . . . . . . . . . . . 74
3.3.2 Effect of organic shock loads on the stability of the anaerobic reactor 78
3.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
4. Modeling simultaneous carbon and nitrogen removal in an anaerobic-

anoxic reactor treating domestic wastewater 81
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.2.1 ADM1 extension modifications and implementation . . . . . . . . . 84
4.2.2 Batch model calibration . . . . . . . . . . . . . . . . . . . . . . . . 88
4.2.2.1 Experimental setup . . . . . . . . . . . . . . . . . . . . . . 88
4.2.2.2 Model calibration procedure . . . . . . . . . . . . . . . . . 89
4.2.3 Model simulation in a continuous reactor . . . . . . . . . . . . . . . 90
4.2.3.1 Model inputs and initial conditions . . . . . . . . . . . . . 90
4.2.3.2 Definition of scenarios . . . . . . . . . . . . . . . . . . . . 91
4.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.3.1 Batch reactor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.3.1.1 Model calibration . . . . . . . . . . . . . . . . . . . . . . . 92
4.3.2 Simulation continuous reactor . . . . . . . . . . . . . . . . . . . . . 93
4.3.2.1 Analysis of scenarios . . . . . . . . . . . . . . . . . . . . . 93
4.3.2.2 The effect of HRT on SCNR . . . . . . . . . . . . . . . . . 98
4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
27
5. Modeling of simultaneous carbon and nitrogen removal in an anaerobic-

anoxic reactor treating salmon fishery wastewater 103
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.2.1 Model description and implementation . . . . . . . . . . . . . . . . 105
5.2.2 Experimental setup and initial conditions . . . . . . . . . . . . . . . 110
5.2.3 Sensitivity analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.2.4 Model calibration and validation procedure . . . . . . . . . . . . . . 112
5.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
5.3.1 Sensitivity analysis and model calibration . . . . . . . . . . . . . . 112
5.3.2 Direct validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.3.3 Cross validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
6. Conclusões e sugestões para trabalhos futuros 119

6.1 Conclusões . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6.2 Sugestões para trabalhos futuros . . . . . . . . . . . . . . . . . . . . . . . . 120
Referências 123
28
Capı́tulo 1
Introdução
1.1 Motivação e justificativa
Nos últimos anos, um fator-chave nos processos de decisão e escolha de um sistema ade-
quado para tratamento de águas residuárias está relacionado ao investimento e a operação.
No contexto de pesquisa, tais perspectivas frequentemente estão associadas à inovação e de-
senvolvimento de tecnologias simplificadas e sustentáveis. Estudos que envolvem inovação
tecnológica e a compreensão dos mecanismos para o desenvolvimento de tecnologias sim-
plificadas dependem de estudos avançados, nos quais, a modelagem matemática é uma das
ferramentas que tem sido utilizada em estágio crescente no contexto de pesquisas brasilei-
ras.
Em relação às tecnologias simplificadas, os reatores anaeróbios quando aplicados como
unidades principais de tratamento de águas residuárias apresentam vantagens para a
remoção de matéria orgânica, podendo-se destacar: (i) menor produção de lodo, (ii) baixo
consumo energético (com potencial de produção de energia a partir do uso do biogás),
(iii) menor susceptibilidade a variação de carga orgânica, entre outras. As diversas carac-
terı́sticas favoráveis desse processo, aliadas às condições ambientais, principalmente nas
regiões onde há predominância de temperaturas mais elevadas, têm contribuı́do para a co-
locação dos sistemas anaeróbios de tratamento de águas residuárias em posição de destaque
em relação aos sistemas aeróbios. No entanto, dificilmente esses reatores produzem efluen-
tes que se enquadrem a padrões estabelecidos pela legislação ambiental vigente na maioria
dos paı́ses, tonando-se necessário a aplicação de um pós-tratamento visando complementar
à remoção de matéria orgânica, bem como o de proporcionar a remoção de nutrientes eu-
trofizantes (nitrogênio e fósforo) e organismos patógenos. Desse modo, o pós-tratamento
dos efluentes de sistemas anaeróbios se torna indispensável.
O sistema convencional de tratamento biológico de águas residuárias que visa promo-
30 Capı́tulo 1. Introdução
ver a remoção de matéria orgânica e de nutrientes é composto de três estágios, anóxico,

anaeróbio e aeróbio (anoxic-anaerobic-oxic - A2 O). No caso do nitrogênio, a remoção é re-
alizada por meio da nitrificação autotrófica seguida da desnitrificação heterotrófica. Essa
última é efetivada na presença de doadores de elétrons orgânicos, que podem ser proveni-
entes de fonte endógena ou exógena. No caso do efluente do estágio anaeróbio, a fração
de matéria orgânica remanescente não é facilmente biodegradável, o que dificulta sua uti-
lização como fonte endógena para o processo de desnitrificação. Portanto, se faz necessária
a adição de fontes de carbono exógenas, tais como, metanol e etanol, aumentando os custos
operacionais do processo.
Uma alternativa visando a simplificação e redução dos custos relacionados ao trata-
mento terciário é a integração dos sistemas anaeróbio-anóxico, cujo objetivo é a remoção
simultânea de carbono e nitrogênio (RSCN). No primeiro estágio, ocorrem os processos
simultâneos de digestão anaeróbia e desnitrificação e no segundo estágio a nitrificação e
a subsequente recirculação do efluente aeróbio nitrificado para o sistema anaeróbio. Essa
configuração pode levar à redução no consumo de energia, à produção de lodo e à possibili-
dade da remoção de compostos tóxicos. Além disso, o nitrato pode atuar como responsável
pelo aumento do potencial redox no reator anaeróbio, inibindo a formação do gás sulfı́drico,
responsável pela corrosão das estruturas fı́sicas dos reatores anaeróbios e emissão de odores.
Embora já se conheça alguns efeitos da exposição de microrganismos anaeróbios às
concentrações de amônia, os efeitos da introdução de óxidos de nitrogênio em sistemas de
tratamento de águas residuárias não estão completamente elucidados. No Brasil, o apro-
fundamento das pesquisas sobre esse tema justifica-se pela larga utilização da associação
de processos anaeróbios e aeróbios para o tratamento de águas residuárias concentradas e
diluı́das. Entretanto, deve-se destacar que a busca do equilı́brio entre o processo de redução
biológica de óxidos de nitrogênio e a digestão anaeróbia não é simples, dada a possibili-
dade de ocorrência de uma competição pelos compostos orgânicos doadores de elétrons
das reações e/ou de inibição da atividade metanogênica e, consequentemente, redução da
produção de gás metano.
Diante desse contexto, o comportamento da RSCN em reatores anaeróbio-anóxico vem
sendo estudado através de modelagem matemática e simulação. Assim, justificadamente,
nos últimos anos houve um progresso significativo na modelagem de processos de digestão
anaeróbia baseada no modelo matemático genérico desenvolvido por um grupo de especi-
alistas vinculados a International Water Association (IWA), o Anaerobic Digestion Model
- ADM1 (Batstone et al., 2002a). Uma extensão do ADM1 foi proposta por Tugtas et al.
(2006), a qual inclui o processo de desnitrificação, porém, sua aplicação na investigação da
Seção 1.2. Objetivos da tese 31
influência de variáveis de processo tais como: variação da vazão volumétrica na corrente

afluente, pH e o efeito da relação DQO/NO−
3 na remoção simultânea de carbono e ni-
trogênio de águas residuárias concentradas e diluı́das, ainda é incipiente e pouco explorada

o que justifica a realização de trabalhos nessa direção.
1.2 Objetivos da tese
O objetivo principal desta pesquisa é avaliar a remoção simultânea de carbono e ni-

trogênio (RSCN) em reator anaeróbio-anóxico no tratamento de águas residuárias por meio
da modelagem matemática e simulação.
Para atingir o objetivo principal desta pesquisa, foram definidos os seguintes objetivos
especı́ficos:
• Aplicar o ADM1 para simular o comportamento dinâmico de um reator anaeróbio

alimentado com lodo de esgoto doméstico em condições mesofı́licas;
• Modelar e simular os processos simultâneos de fermentação, metanogênese e des-

nitrificação para avaliar a RSCN em um reator anaeróbio-anóxico tratando águas
residuárias domésticas;
• Modelar e simular os processos simultâneos de fermentação, metanogênese e des-

nitrificação para avaliar a RSCN em um reator anaeróbio-anóxico alimentado com
substrato sintético para simular um efluente de indústria de processamento de pes-
cado;
1.3 Estrutura da tese
Esta tese está dividida em capı́tulos de tal forma que os assuntos são apresentados com
um aprofundamento gradativo no tema. Os capı́tulos são auto suficientes, contendo os
seus próprios resumos e objetivos especı́ficos, portanto, podem ser lidos separadamente.
Em consonância com a prática de uso intensivo e difundido do idioma inglês como meio de
comunicação e divulgação, usual na comunidade cientı́fica das áreas de engenharia e tecno-
logia, e respeitadas as normas do Programa de Pós-Graduação em Engenharia Industrial
da Universidade Federal da Bahia (PEI-UFBA), alguns dos capı́tulos que apresentam os
principais resultados e novas contribuições são apresentados em idioma inglês. Adicional-
mente, busca-se ampliar o acesso da comunidade cientı́fica internacional ao trabalho. Além
disso, também, buscando aderência com as normas de publicação cientı́ficas internacionais,
32 Capı́tulo 1. Introdução
é adotada a convenção de utilização do ponto como separador decimal, ao invés do uso da

vı́rgula.
O capı́tulo 2 apresenta uma revisão bibliográfica estruturada em três seções principais,
que norteou a consecução dos objetivos propostos para a elaboração desta tese. Na primeira
seção, buscou-se fornecer uma visão geral dos processos biológicos de digestão anaeróbia
incluindo a fermentação, a metanogênese, a redução de nitrato e a interação simultânea
desses processos em ambientes anaeróbios. Em seguida, apresenta-se a base conceitual do
processo convencional de remoção de nitrogênio e uma breve revisão da literatura sobre a
RSCN em reatores anaeróbio-anóxico. Na segunda, é apresentada a evolução dos modelos
matemáticos de digestão anaeróbia, a descrição detalhada do modelo ADM1 incluindo
a extensão do processo de desnitrificação. Na terceira, é apresentada uma revisão das
metodologias e estratégias para aplicar o ADM1.
O capı́tulo 3 tem como objetivo mostrar os resultados da implementação e ajuste do
ADM1 aos dados experimentais (fonte secundária) de um reator anaeróbio operado em es-
cala de laboratório alimentado com lodo de esgoto sanitário. Adicionalmente, foi realizada
uma simulação do efeito de choque de carga orgânica no desempenho desse reator.
O capı́tulo 4 mostra os resultados da modelagem matemática baseada no ADM1 para
avaliar a RSCN em um reator anaeróbio-anóxico tratando águas residuária doméstica. Os
resultados obtidos pelo modelo são comparados com os dados experimentais visando avaliar
o potencial desse modelo para a sua aplicação. O capı́tulo segue com a proposição de seis
cenários para simular efeito do TDH e da razão DQO/NO−
3 em um reator anaeróbio-anóxico
operado em regime contı́nuo.

O capı́tulo 5 mostra os resultados da modelagem matemática baseada no ADM1 para
avaliar a RSCN em um reator anaeróbio-anóxico alimentado com substrato sintético para
simular um efluente de indústria de processo de pescado. Foi realizada uma análise de
sensibilidade para identificar os parâmetros mais importantes para o processo. Após o
levantamento dos parâmetros mais significativos, o modelo foi calibrado e validado.
Este trabalho é encerrado no capı́tulo 6 com a apresentação dos comentários finais e a
descrição das atividades futuras.
Capı́tulo 2
Revisão da literatura
2.1 Processos biológicos de digestão anaeróbia
A digestão anaeróbia em reatores estrutura um ecossistema no qual micro-organismos

pertencentes aos domı́nios Archaea e Bacteria, trabalham interativamente para promover
a oxidação e fermentação estável e auto-regulada da matéria orgânica. Esse processo
ocorre na ausência de oxigênio, contando com dióxido de carbono (CO2 ), sulfato (SO4 2− ),
nitrato (NO3 − ), ferro trivalente (Fe3 + ) ou outro ı́on, como aceptores finais de elétrons.
Os produtos finais estão associados ao fluxo de elétrons, destacando-se o metano, gás
carbônico, sulfeto de hidrogênio e nitrogênio gasoso. No tratamento de águas residuárias e
resı́duo sólidos, esse processo é responsável pela conversão da matéria orgânica em metano,
dióxido de carbono, hidrogênio, como principais produtos gasosos, e biomassa por meio de
processos metabólicos que ocorrem em etapas sequenciais (Mccarty, 2001). A Figura 2.1
mostra a sequência de processos metabólicos na digestão anaeróbia incluindo, sulfetogênese,
desnitrificação, e o uso de compostos inorgânicos (sulfeto, enxofre elementar e hidrogênio)
como doadores de elétrons para o processo de desnitrificação autotrófica.
2.1.1 Fermentação e metanogênese
O processo de conversão da matéria orgânica em ambientes anaeróbios ocorre por meio

de processos metabólicos de fermentação e respiração. Durante a fermentação, a oxidação
da matéria orgânica é realizada na ausência de um aceptor final de elétrons, enquanto que
na respiração, são utilizados aceptores de elétrons inorgânicos, como o NO3 − (redução do
NO3 − ) e CO2 (formação do metano) (Chernicharo, 2007). Nesse caso, a metanogênese pode
ser considerada uma forma de respiração anaeróbia, na qual o gás carbônico é o aceptor
final de elétrons. Portanto, o processo de oxidação de matéria orgânica em ambientes
anaeróbios pode ser simplificado em duas fases, fermentação e metanogênese.
34 Capı́tulo 2. Revisão da literatura
Figura 2.1: Sequência de processos metabólicos durante a digestão anaeróbia. Fonte: Adaptado de
Batstone et al. (2015)
Para à ocorrência dos processos metabólicos de fermentação e metanogênese é necessária

a atividade conjunta de ao menos três grupos de micro-organismos distintos: as bactérias
fermentativas hidrolı́ticas ou acidogênicas e bactérias sintróficas acetogênicas e as Archaeas
metanogênicas.
Na primeira etapa do processo fermentativo, denominada acidogênese, as bactérias
fermentativas hidrolı́ticas excretam exoenzimas que são responsáveis pela degradação da
matéria orgânica complexa (polı́meros), em compostos dissolvidos menores (monômeros),
os quais podem atravessar as paredes celulares das bactérias fermentativas, uma vez que as
bactérias não são capazes de assimilar a matéria orgânica particulada (macromoléculas). As
proteı́nas são degradadas em aminoácidos (AA), os carboidratos em açúcares solúveis (MS)
e os lipı́deos, em ácidos graxos de cadeia longa (LCFA) (C15 e C17 ) e glicerol (van Haandel
e Lettinga, 1994). Os produtos solúveis oriundos da hidrólise são absorvidos e metaboli-
zados pelas bactérias fermentativas acidogênicas, que, por sua vez, excretam substâncias
orgânicas mais simples, como ácidos orgânicos (principalmente os ácidos acético, propiônico
(ProH), butı́rico (BuH) e valérico (ValH)), etanol (EtOH), ácido lático (LacH) e compostos
inorgânicos, tais como CO2 , H2 , NH3 e H2 S, além de formar novas células dos microrga-
nismos. Com exceção das novas células, esses produtos são importantes substratos para
Seção 2.1. Processos biológicos de digestão anaeróbia 35
os micro-organismos acetogênicos e metanogênicos. A maioria das bactérias fermentativas

acidogênicas são anaeróbias estritas, existindo, também, algumas espécies que metaboli-
zam o material orgânico pela via oxidativa. Esses micro-organismos são os que mais se
beneficiam energeticamente, pois são os primeiros a atuarem na degradação do material
orgânico. Dessa forma, essa etapa só será limitante do processo, se o material orgânico
não for facilmente hidrolisável (Henze et al., 2008).
Na segunda etapa do processo fermentativo, denominada acetogênese, as bactérias
sintróficas acetogênicas produzem acetato, CO2 e H2 , substratos que são metabolizados pe-
las Archaeas metanogênicas. Baseado no metabolismo, dois grupos distintos de bactérias
acetogênicas podem ser distinguidos (Westerholm et al., 2011).
O primeiro grupo é o de bactérias acetogênicas produtoras de hidrogênio obrigatórias,
também chamadas de acetogênicas redutoras de prótons, que produzem ácido acético, CO2
e H2 a partir de uma grande variedade de substratos (ácidos orgânicos). Segundo Hao et al.
(2011) a acetogênese faz a ligação entre ácidos graxos com três ou mais carbonos e o acetato.
As ligações carbono-carbono são quebradas pelas acetogênicas redutoras de prótons até
que sejam formados os principais produtos finais, acetato e hidrogênio. No entanto, essas
reações só passaram a ser exergônicas (∆G◦ < 0), se o hidrogênio for removido e sua pressão
parcial for mantida em nı́veis baixos, entre 10 - 400 Pa. Para que essas reações ocorram, é
necessário o crescimento de bactérias redutoras de prótons, além de depender do consumo
de hidrogênio por outros micro-organismos. Esse equilı́brio pode ser alcançado por meio de
micro-organismos consumidores de hidrogênio que estão presentes no sistema, tais como:
as Archaeas metanogênicas hidrogenotróficas, as bactérias redutoras de sulfato (BRS) e
as homoacetogênicas. A maioria dos ambientes metanogênicos mantém uma concentração
de H2 abaixo de 10−4 atm, que é suficiente para estimular o crescimento dessas bactérias
acetogênicas, evitando-se o acúmulo de ácidos (Chernicharo, 2007).
O segundo grupo de bactérias acetogênicas são as homoacetogênicas, que são estrita-
mente anaeróbias, catalisando a formação de acetato a partir de CO2 e H2 . A atividade
metabólica dessas bactérias independe das relações sintróficas já mencionadas, e, por isso,
são mais facilmente isoladas de amostras de lodo anaeróbio. Assim como as Archaeas me-
tanogênicas, essas bactérias podem também contribuir para a manutenção de um ambiente
com baixa pressão de H2 .
O acetato e o hidrogênio produzidos nas etapas anteriores do processo fermentativo
são convertidos em metano e dióxido de carbono, pelas Archaeas metanogênicas. Esses
micro-organismos são anaeróbios estritos e, geralmente, estão presentes em ambientes onde
os aceptores de elétrons como O2 , NO3 − , Fe3 + e SO4 2− estão ausentes ou apresentam-se
sob baixas concentrações. Como resultado de sua afinidade por substrato e magnitude
de produção de metano, podem ser divididas em dois grupos principais: um que forma o
metano a partir de acetato, e outro que produz metano a partir de hidrogênio e dióxido
de carbono. As Archaeas metanogênicas acetoclásticas usam o acetato como fonte de
carbono e energia. As Archaeas metanogênicas hidrogênotróficas utilizam o dióxido de
carbono como fonte de energia e aceptor final de elétrons, e o hidrogênio, como fonte de
energia. Ao contrário das Archaeas metanogênicas acetoclásticas, praticamente todas as
espécies conhecidas de Archaeas metanogênicas são capazes de produzir metano, a partir
do hidrogênio e dióxido de carbono, resultando em uma maior liberação de energia. No
entanto, embora poucas espécies sejam capazes de formar o metano a partir do acetato,
as Archaeas metanogênicas acetoclástica são responsáveis por cerca de 60 a 70% de toda
produção de metano em ambientes anaeróbios (Chernicharo, 2007). As Equações 2.1 e
2.2, mostram a comparação energética entre as Archaeas metanogênicas acetoclásticas e
as Archaeas metanogênicas hidrogenotrófica.
CH3 COOO− + H2 O → CH4 + HCO3− ∆G◦ = −31, 0 kJ (2.1)
1 1 1 3
H2 + HCO3− + H + → CH4 + H2 O ∆G◦ = −33, 9 kJ (2.2)
4 4 4 4
Além dos processos fermentativo e metanogênico que levam a produção do biogás,
podem se desenvolver outros processos biológicos em ambientes anaeróbios. Na presença
de oxidantes alternativos (aceptores de elétrons), como nitrato e sulfato, muitos compostos
intermediários formados por meio da fermentação anaeróbia, passam a ser utilizados pelas
bactérias redutoras de nitrato e sulfato, respectivamente, provocando alterações das rotas
metabólicas no ambiente anaeróbio. O nitrato pode ser usado como agente oxidante, sendo
reduzido a nitrogênio molecular (N2 ) ou a ı́on amônio (NH4 + ) e o sulfato pode ser reduzido
para sulfeto de hidrogênio (Batstone e Jensen, 2011).
2.1.2 Redução de nitrato
A redução do nitrato por meio de processos biológicos pode ser realizada de duas
formas: assimilativa e desassimilativa. Na redução assimilativa do nitrato, a catálise é feita
pela enzima redutase nitrato, convertendo nitrato a amônia, seguindo a rota metabólica
inversa da nitrificação, sendo a amônia resultante usada na sı́ntese de novas estruturas
celulares. Na redução desassimilativa, o nitrato é reduzido a nitrogênio gasoso na presença
de doadores de elétrons, que podem ser de origem orgânica ou inorgânica, caracterizando
a desnitrificação heterotrófica e autotrófica, respectivamente (desnitrificação) (Madigan

et al., 2004). O processo mais largamente empregada no tratamento de águas residuárias
é a desnitrificação heterotrófica.
No processo de desnitrificação heterotrófica, os micro-organismos utilizam uma fonte
de carbono orgânico como doador de elétrons, sendo de fonte endógena (proveniente da
própria água residuária ou do lodo) ou exógena (adicionada externamente). A sequência
do processo de desnitrificação pode ser visualizada na Equação 2.3.
N O3− → N O2− → N O(g) → N2 O(g) → N2 (g) (2.3)
A redução de cada óxido de nitrogênio é catalisada por uma enzima especı́fica: nitrato
redutase, nitrito redutase, óxido nı́trico redutase e óxido nitroso redutase, respectivamente.
Além do nitrato, os óxidos de nitrogênio intermediários servem também como aceptores de
elétrons na respiração, acompanhada da oxidação de compostos orgânicos como doadores
de elétrons para geração de energia e nova matéria celular (Philips et al., 2002).
A literatura reporta uma vasta quantidade de micro-organismos que são capazes de re-
alizar a desnitrificação, sendo a maioria bactérias. Algumas Archaeas também são capazes
de realizar esse processo, assim como alguns fungos filamentosos. Com poucas exceções,
os micro-organismos desnitrificantes também são capazes de respirar utilizando o oxigênio
como aceptor final de elétrons. Somente na ausência de oxigênio no meio, é que esses micro-
organismos utilizam nitrato e nitrito como aceptor final de elétrons. Dentre os gêneros de
bactérias envolvidas, podem ser citadas as Archromobacter, Aerobacter, Alcaligenes, Bacil-
lus, Flavobacterium, Micrococcus, Proteus e Pseudomonas (Shapleigh, 2006).
Nos ambientes anaeróbios, nos quais o nitrato está presente, uma gama de substra-
tos podem ser utilizados pelos micro-organismos desnitrificantes como fonte de energia
para crescimento, incluindo glicose, aminoácidos Marazioti et al. (2003), ácidos graxos
voláteis (AGV) Elefsiniotis e Wareham (2004) e hidrogênio Scheid et al. (2003). Alguns
estudos na literatura reportam que, dentre os AGV, o propionato tem sido a mais prefe-
rida fonte de carbono das bactérias desnitrificantes (Aboutboul et al., 1995; Elefsiniotis e
Wareham, 2004). Além disso, tem sido mostrado que para uma mistura de cultura fer-
mentativa/metanogênica, a presença de nitrato, nitrito e óxido nı́trico pode resultar em
uma diminuição da pressão parcial H2 para um nı́vel que não permite a ocorrência da
metanogênese (Kluber e Conrad, 1998a; Scheid et al., 2003). A Figura 2.2 mostra um
esquema da inter-relação dos processos de digestão anaeróbia com o ciclo de nitrogênio.
As equações estequiométricas das reações de alguns substratos considerados neste estudo
foram determinadas baseadas nos princı́pios bioenergéticos (Mccarty, 2001), a saber:
• Desnitrificação com glicose

1
CH O
24 6 12 6
+ 14 H2 O → 41 CO2 + H + + e− ∆G0 = −41, 35 kJ/eeq
1
CH O
24 6 12 6
+ 15 N O3− + 56 H + + e− → 1
N
10 2
+ 35 H2 O ∆G0 = −72, 20 kJ/eeq
Reação global:
1
CH O
24 6 12 6
+ 15 N O3− + 51 H + → 7
HO
20 2
+ 41 CO2 + 1
N
10 2
∆G0 = −113, 3 kJ/eeq
• Desnitrificação com propionato

1
24
CH3 CH2 COO− + 14
5
H2 O → 17 CO2 + 14
1
HCO3− + H + + e− ∆G0 = −27, 63 kJ/eeq
1
5
N O3− + 56 H + + e− → 1
N
10 2
+ 35 H2 O ∆G0 = −72, 20 kJ/eeq
Reação global:
1
14
CH3 CH2 COO− + 15 N O3− + 51 H + → 17
HO
10 2
+ 1
14
HCO3− + 1
N
10 2
∆G0 = −99, 83 kJ/eeq
• Desnitrificação com acetato

1
8
CH3 COO− + 38 H2 O → 81 HCO3− + H + + e− ∆G0 = −27, 40 kJ/eeq
1
5
N O3− + 56 H + + e− → 1
N
10 2
+ 35 H2 O ∆G0 = −72, 20 kJ/eeq
Reação global:
1
8
CH3 COO− + 15 N O3− + 51 H + → 9
HO
40 2
+ 18 HCO3− + 1
N
10 2
∆G0 = −99, 60 kJ/eeq
• Desnitrificação com H2
1
H
2 2
→ H + + e− ∆G0 = −39, 87 kJ/eeq
1
5
N O3− + 56 H + + e− → 1
N
10 2
+ 35 H2 O ∆G0 = −72, 20 kJ/eeq
Reação global:
1
H
2 2
+ 15 + N O3− 15 H + → 53 H2 O + 1
N
10 2
∆G0 = −112, 07 kJ/eeq
A utilização desses compostos orgânicos pode gerar uma competição entre as bactérias
fermentativas, Archaeas metanogênicas e desnitrificantes. No entanto, alguns pesquisado-
res apontam que a inibição por N-óxido é a principal causa para supressão da metanogênese
Composto orgânico complexo
2° etapa da 1° etapa da
nitrificação nitrificação
(nitratação) (nitritação)
Redução
desassimilatória
Carboidratos Proteínas Lipídeos do nitrato a íon
amônio (RDNA) Não
há
Assimilação pelas mud
células ança
Amonificação redo
assimilatória x
Açúcares Aminoácidos LCFA (Proteína) Não
Assimilação pelas há
(Monomer
células mud
os)
ança
Fixação redo
x
Ácidos orgânicos
Valerato, isovalerato, propionato,
butirato, etc
Desnitrificação
Acetato H2 autotrófica
Não há mudança
Redução Oxidação
redox
CH4 ,CO2
Figura 2.2: Inter-relação da digestão anaeróbia com o ciclo do nitrogênio
(Roy e Conrad, 1999; Tugtas e Pavlostathis, 2007a). Segundo Kluber e Conrad (1998a), o
efeito inibidor de óxidos de nitrogênio sobre a metanogênese foi relatada da seguinte forma:
NO > NO− −
2 > N2 O > NO3 . Além de NO ter o maior efeito inibidor, é também irreversı́vel.
Por outro lado, a inibição causada pelo N2 O foi parcialmente reversı́vel (Kluber e Conrad,
1998a).
Segundo Zumft (1997), a redução do nitrito é a etapa mais lenta do processo de des-
nitrificação, logo as taxas de redução de nitrato, óxido nitroso e óxido nı́trico são mais
elevadas. Como resultado, a acumulação de nitrito é frequentemente observada em siste-
mas desnitrificantes. Os valores da constante de meia velocidade para N-óxidos (KS,nox )
variam dependendo do tipo de micro-organismo e, geralmente, estão em uma faixa de
3 × 10−8 a 4 × 10−3 kg N/m3 , embora, um valor elevado como 0,018 kg N/m3 tenha sido
reportado (Zumft, 1997; Kluber e Conrad, 1998a).
A competição entre os micro-organismos desnitrificantes, fermentativos e metanogênicos
pelo substrato depende dos valores da taxa de utilização do substrato e dos valores da cons-
tante cinética de meia saturação de cada micro-organismo (KS,subs ). A taxa de utilização
de glicose e aminoácidos pelos micro-organismos desnitrificantes é dez vezes menor do que
a taxa de utilização desses substratos pelas bactérias fermentativas. Embora os valores da

constante cinética KS,subs para desnitrificantes sejam menores, esses micro-organismos não
são capazes de competir com as bactérias fermentativas por substratos como, glicose ou
aminoácidos, devido a maior taxa de utilização desses substratos pelas bactérias fermenta-
tivas. No entanto, os micro-organismos desnitrificantes podem competir com as Archaeas
metanogênicas por acetato.
2.2 Processo convencional de remoção biológica de nitrogênio
Nos sistemas convencionais, a remoção biológica de nitrogênio amoniacal (N-NH3 + e/ou

N-NH4 + ) de águas residuárias ocorre em duas etapas, envolvendo a nitrificação autotrófica
seguida da desnitrificação heterotrófica. A nitrificação autotrófica ocorre estritamente sob
condições aeróbias, sendo o nitrogênio amoniacal oxidado a nitrito e sequencialmente a
nitrato. Subsequentemente, na desnitrificação, o nitrato é reduzido a nitrogênio molecular
(N2 ) em um processo heterotrófico, sob condições anóxicas. Essas séries de reações de
oxidação e redução fazem parte do ciclo do nitrogênio (Figura 2.3). As reações que ocorrem
na Figura 2.3 estão resumidas na Tabela 2.1.
2° etapa da nitrificação 1° etapa da nitrificação

(nitratação) (nitritação)
Redução desassimilatória
do nitrato a íon amônio
(RDNA)
Não há
Assimilação pelas células mudanç
Amonificação assimilatória a redox
(Proteína)
Assimilação pelas células Não há
(Monomeros) mudanç
Fixação a redox
Desnitrificação autotrófica
Redução Oxidação Não há mudança redox
Figura 2.3: Ciclo biológico do nitrogênio. Adaptado de Metcalf e Eddy (2007) e Henze et al. (2008).
O processo de nitrificação é realizado em duas etapas, envolvendo dois grupos de micro-

organismos autotróficos: aqueles que oxidam ı́on amônio e/ou amônia a nitrito (nitritação),
predominantemente Nitrosomonas sp. e Nitrosospira sp. e os que oxidam nitrito a nitrato
Tabela 2.1 - Reações do nitrogênio
No Reação ∆G◦ (kJ/mol) Referência
Nitrificação (Nitritação e Nitratação)
1 N H4+ + 1, 5O2 → N O2− + 2H + + H2 O -277,68 Chen et al. (2009)
2 N O2− + 0, 5O2 → N O3− -74,14 Chen et al. (2009)
3 N H4+ + 2O2 → N O3− + 2H + + H2 O (I) -349 Bernet et al. (1996)
− + (II)
4 N H4+ + 2HCO3− 3 + 2O2 → N O3 + 2H + H2 O -349 Bernet et al. (1996)
5 N H4+ + 1, 381O2 + 1, 981HCO3− → 0, 981N O2− + 1, 89H2 CO3 + 1, 0363H2 O + 0, 0181C5 H7 N O2 (III) Bernet et al. (1996)
6 N O2− + 0, 0025N H4+ + 0, 01H2 CO3 + 0, 487O2 → N O3− + 0, 075H2 O + 0, 025C5 H7 N O2 (IV) Bernet et al. (1996)
7 N H4+ + 1, 863O2 + 0, 098CO2 → 0, 098N O3− + 1, 98H + + 0, 0941H2 O + 0, 0196C5 H7 N O2 (V) -349 Bernet et al. (1996)
Desnitrificação (Heterotrófica)
8 5H2 + 2N O3− + 2H + → 6H2 O + N2 (g) -224 Tugtas e Pavlostathis (2007a)
9 N O2− + 4H + + 3e− → 2H2 O + 0, 5N2 (g) 277 Gao et al. (2010)
10 5CH3 COO− + 8N O3− + 8H + → 9H2 O + 5CO2 + 5HCO3− + 4N2 (g) -797 Tugtas e Pavlostathis (2007a)
11 5CH3 CH2 COO− + 14N O3− + 14H + → 17H2 O + 10CO2 + 5HCO3− + 7N2 (g) -1398 Tugtas e Pavlostathis (2007a)
12 5C6 H12 O6 + 24N O3− + 24H + → 42H2 O + 30CO2 + N2 (g) -2657 Lengeler e Schlege (1999)
13 C6 H12 O6 + 8N O2− + 8H + → 10H2 O + 6CO2 + 4N2 (g) -3144 Lengeler e Schlege (1999)
14 5CH3 OH + 6N O3− → 3N2 + 5CO2 + 6OH − + 7H2 O Lengeler e Schlege (1999)
15 C10 H19 O3 + 10N O3 → 5N2 + 10CO2 + 3H2 O + N H3 + 10OH − Bernet et al. (1996)
Desnitrificação (Autotrófica)
16 8N O3− + 5M eS + H2 O → 2H + + 4N2 (g) + 5SO42− + M e(OH)2 Cardoso et al. (2006)
17 N O3− + H + + 2, 5H2 → 0, 5N2 (g) + 3H2 O -560,3 Banihani et al. (2009)
18 3N O3− + 5N H4+ → 4N2 (g) + 9H2 O + 2H + -297 Banihani et al. (2009)
RDNA (Heterotrófica)
19 4H2 + 2N O3− + 4H + → 6H2 O + 2N H4+ -150 Tugtas e Pavlostathis (2007a)
20 CH3 COO− + N O3− + 2H + → CO2 + HCO3− + N H4+ -500 Tugtas e Pavlostathis (2007a)
21 8CH3 CH2 COO− + 14N O3− + 28H + → 2H2 O + 16CO2 + 8HCO3− + 14N H4+ -878 Tugtas e Pavlostathis (2007a)
22 C6 H12 O6 + 3N O3− + 6H + → 3N H4+ + 3H2 O + 6CO2 -1767 Lengeler e Schlege (1999)
23 C6 H12 O6 + 12N O3− → 12N O2− + 6H2 O + 6CO2 -1767 Lengeler e Schlege (1999)
Seção 2.2. Processo convencional de remoção biológica de nitrogênio
24 C6 H12 O6 + 4N O2− + 8H + → 4N H4+ + 2H2 O + 6CO2 -1713 Lengeler e Schlege (1999)

Desamonificação
25 N H2 OH + N O2− → N2 O(g)
Fixação do nitrogênio
26 N2 + 8H + + 8e− + 16AT P → 2N H3 + H2 + 16ADP + 16P i Henze et al. (2008)
Oxidação anóxica do ı́on amônio
27 N O2− + N H4+ → N O(g) Henze et al. (2008)
Oxidação anaeróbia do ı́on amônio (ANAMMOX)
41
28 1, 3N O2− + N H3+ → 1, 02N2 (g) + 0, 26N O3− + 2H2 O -357 Henze et al. (2008)
(nitratação), predominantemente Nitrobacter sp. e Nitrospira sp. (Metcalf e Eddy, 2007).

Durante esse processo, os micro-organismos utilizam nitrogênio amoniacal ou nitrito como
fonte de energia e dióxido de carbono ou carbonato como fonte de carbono inorgânica,
enquanto o oxigênio molecular é o aceptor de elétrons. As equações estequiométricas
referentes às duas etapas do processo de oxidação do nitrogênio amoniacal a nitrato podem
ser descritas segundo as Equações 1 e 2 na Tabela 2.1.
O coeficiente de sı́ntese celular dos micro-organismos nitrificantes é baixo, pois a fixação
do CO2 consome cerca de 80% da energia gerada pelas bactérias que oxidam amônia, e
para cada átomo de carbono fixado, as bactérias nitrificantes precisam oxidar 35 moléculas
de NH3 ou 100 moléculas de nitrito, portanto o coeficiente de sı́ntese celular dos micro-
organismos nitrificantes é baixo. Além disso, a taxa de crescimento é lenta comparada aos
micro-organismos heterotróficos (Philips et al., 2002). Considerando-se as Equações de 1 a
7 da Tabela 2.1 e a Figura 2.4, que mostra a variação do número de oxidação do nitrogênio
nos processos de nitrificação e desnitrificação, pode-se estimar o consumo de oxigênio para
o processo de nitrificação.
Figura 2.4: Variação do número de oxidação do nitrogênio. Adaptado de Van Haandel (1999).
Assim, o consumo de O2 para nitrificação de 1 mol de N-NH3 = 64/14 = 4,57 mgO2 .mg
N−1 . Deve-se considerar o fato que o processo de nitrificação produz efeito sobre a alcali-
nidade da água residuária, estequiometricamente, nota-se que há produção de 2 moles de
H+ por mol de nitrato formado. Levando em conta que a produção de 1 mol de H+ é equi-
valente ao consumo de 1 mol de alcalinidade ou 50 gCaCO3 para nitrificação, a variação
da alcalinidade (VA ) é dado por:
−100
VA = = −7, 14 mgCaCO3 .mgN−1 (2.4)
14
A segunda etapa do processo de remoção biológica de nitrogênio envolve a desnitri-
ficação heterotrófica. A Figura 2.4 mostra que no processo de desnitrificação heterotrófica,
há uma transferência de 5 elétrons por átomo de nitrogênio na redução do oxidante do ni-
trato para nitrogênio molecular, ou seja, dos oito elétrons liberados durante a oxidação do
Seção 2.2. Processo convencional de remoção biológica de nitrogênio 43
N-NH3 a NO−
3 , cinco são recuperados na redução. Assim, observa-se que uma fração igual a
5/8 do oxigênio necessário para nitrificação pode ser recuperada durante a desnitrificação.
Ou seja:
5
Or = × 4, 57 = 2, 86 mgO2 .mgN−1 (2.5)
8
onde Or é a quantidade de oxigênio recuperado durante a desnitrificação. Desta maneira,
o consumo lı́quido de oxigênio (CLO ) para remoção de nitrogênio de uma água residuária,
utilizando o processo biológico de nitrificação/desnitrificação, é dado por:
CLO = 4, 57 − 2, 86 = 1, 71 mgO2 .mgN−1 (2.6)
Esse fato traz importantes repercussões para o dimensionamento de um sistema de

tratamento de águas residuárias sob o ponto de vista econômico. Estima-se que o consumo
de oxigênio para nitrificação seja em torno de 33 por cento do consumo total. Observando as
relações anteriores, constata-se que um sistema projetado para nitrificação e desnitrificação
consumiria 5/8 × 33% ∼
= 20 por cento a menos que um sistema somente com nitrificação.
Como o consumo de oxigênio é o fator principal dos custos operacionais em estações de
tratamento, a redução obtida é bastante significativa (Van Haandel, 1999; Metcalf e Eddy,
2007).
As reações de desnitrificação, independente da fonte de carbono, apontam a recuperação
de 1 mol de alcalinidade por mol de nitrato reduzido. Conclui-se que o processo de nitri-
ficação/desnitrificação apresenta um consumo de alcalinidade de 3,57 mgCaCO3 .mgN−1 ;
diferente de sistemas apenas nitrificantes, nos quais o consumo de alcalinidade é de 7,14
mgCaCO3 .mgN−1 . Esse desempenho também implica em conseqüências econômicas, uma
vez que reduz a necessidade de adição de alcalinizante, com a finalidade de manter o pH
do sistema na faixa ótima para o desenvolvimento dos micro-organismos (Van Haandel,
1999).
Um processo diferente da desnitrificação pode ocorrer em sistemas de remoção de ni-
trogênio, sendo o mesmo conhecido como redução desassimilatória do nitrato a ı́on amônio
(RDNA). A RDNA é um processo que ocorre em duas etapas: a primeira envolve a redução
do nitrato a nitrito e é denominada respiração do nitrato. Esse passo inicial é acoplado à
produção de energia na maioria dos micro-organismos e, apesar de necessário, não é um
passo limitante. O passo seguinte é a redução do nitrito a amônia, que é a etapa crı́tica
da reação global. Assume-se que a RDNA é favorecida em ambientes nos quais a presença
do aceptor de elétrons é limitada em relação ao doador de elétrons, ou seja, concentrações
baixas de nitrito ou nitrato. O processo de RDNA é realizado por bactérias fermentativas,

por exemplo: Aeromonas sp, Vibrio sp, Clostridium sp (Metcalf e Eddy, 2007). A Tabela
2.1 mostra as Equações 19 a 24 que caracterizam a reação ocorrida durante o processo.
2.3 Remoção simultânea de carbono e nitrogênio em sistema integrado

anaeróbio-anóxico
Embora o método mais aplicado para remoção de nitrogênio de águas residuárias seja os
processos de nitrificação seguido da desnitrificação, existem algumas limitações econômicas
para sistemas que o adotam, principalmente, com relação ao alto consumo de oxigênio, ao
requerimento de doadores de elétrons exógenos e à geração de lodo. Para contornar essas
limitações, uma alternativa é a remoção simultânea de carbono e nitrogênio (RSCN) em um
sistema integrado anaeróbio-anóxico. Nesse caso, o efluente aeróbio nitrificado é recirculado
para o sistema integrado anaeróbio-anóxico, onde ocorrem os processos simultâneos de
desnitrificação e digestão anaeróbia. As principais vantagens são: a redução dos custos com
tratamento devido a menor produção de lodo, menor consumo de energia e possibilidade
de redução dos efeitos do gás sulfı́drico, responsável por corrosão das estruturas fı́sicas dos
reatores anaeróbios, emissão do odores, etc. As Figuras 2.5a e b mostram os métodos de
tratamento biológico convencional (sistema A2 O) e o alternativo (RSCN) aplicado para
remoção de matéria orgânica e nitrogênio de águas residuárias.
Diversas pesquisas investigaram a RSCN em reatores anaeróbios tratando águas re-
siduárias concentradas (Hendriksen e Ahring, 1996; Bernet et al., 2001; Lee et al., 2004;
Baloch et al., 2006; Ruiz et al., 2006; Huang et al., 2007; Chen et al., 2009; Huiliñir et al.,
2012; Xie et al., 2012) .
Hendriksen e Ahring (1996) investigaram o potencial da RSCN em um reator UASB
tratando água residuária sintética. O reator foi inicialmente inoculado com lodo granular
metanogênico e foi gradualmente adaptado ao nitrato por meio do aumento das concen-
trações de nitrato afluente. Durante uma carga de 336 mg NO− −
3 / L.dia e 6600 mg NO3 /
L.dia, cerca de 99% da matéria orgânica e do nitrogênio foram removidos. Os autores, ci-
tam, ainda, que no experimento realizado em batelada, cerca de 90% do nitrato foi reduzido
a nitrogênio gasoso indicando que ocorreu a desnitrificação completa.
Bernet et al. (2001) estudaram a RSCN em um reator anaeróbio seguido de um rea-
tor aeróbio, operado em bateladas sequenciais tratando águas residuárias de suinocultura.
Nesse estudo, o efluente nitrificado foi recirculado para o reator anaeróbio para ocorrência
do processo de remoção simultânea de carbono e nitrogênio. Três razões de recirculação
Seção 2.3. Remoção simultânea de carbono e nitrogênio em sistema integrado anaeróbio-anóxico 45
a) Estágios: Anaeróbio Anóxico Aeróbio
Efluente
Bomba
Afluente
Ar
Recirculação do nitrato
Recirculação do lodo
Lodo residual
b) Estágios: Anaeróbio/Anóxico Aeróbio
Efluente
Bomba
Afluente
Ar
Recirculação do nitrato
Recirculação do lodo
Lodo residual
Figura 2.5: Diagrama esquemático: (a) método convencional (sistema A2 O); e (b) método alternativo
(RSCN)
foram investigadas, 100%, 200% e 300%, respectivamente. As eficiências de remoção de

carbono e nitrogênio no sistema anaeróbio-anóxico foram 81 - 91% e 85 - 91%, respecti-
vamente. Os autores observaram que quanto maior a razão de recirculação menor foi à
concentração de N-óxidos no efluente final.
Lee et al. (2004) avaliaram o desempenho de um reator UASB para RSCN alimentado
com água residuária sintética à base de acetato. A concentração de matéria orgânica
e nitrato no reator foi de 3.300 mg DQO/L e 600 mg NO−
3 / L. O sistema demonstrou
eficiência de remoção de matéria orgânica e nitrogênio de 99 e 95%, respectivamente. Além

disso, os pesquisadores observaram que a taxa especı́fica de desnitrificação aumentou com
a redução da razão DQO/NO−
3.
Baloch et al. (2006) investigaram a RSCN em um reator anaeróbio de leito granular

seguido de um reator aeróbio tratando águas residuárias de cervejaria. O reator anaeróbio
foi operado sob carga orgânica volumétrica de 3,57 kg DQO/m3 .dia e carga de nitrogênio
amoniacal igual a 0,13 kg NH4 -N/m3 .dia no reator aeróbio. O processo de RSCN ocorreu
com razões de recirculação variando entre 100 a 200% com carga de nitrato e nitrito
variando entre 0,04 a 0,05 kg NOx-N/m3 .dia. Os autores observaram para todas as razões
de recirculação, a eficiência de remoção de carbono foi de 97% no reator anaeróbio e 99%
nos dois reatores. Ocorreu a produção de metano, com sua composição variando de 25 a
50%, evidenciando que ambos os processos simultâneos de metanogênese e desnitrificação
ocorreram no reator anaeróbio.
Ruiz et al. (2006), avaliaram o efeito da razão DQO/NO−
3 na RSCN em 5 reatores
UASB operados em paralelo. Os reatores foram inoculados com lodo granular, proveni-
ente de um reator UASB em escala piloto tratando águas residuárias de cervejaria. A
carga orgânica volumétrica aplicada foi 7,5 kg DQO/m3 .dia. Os autores observaram que
a razão DQO/NO−
3 tem uma forte influência na via metabólica para utilização de nitrato
e matéria orgânica. Baixos valores dessa relação geraram altas atividade desnitrificantes,
e altos valores, elevada atividades metanogênica. A estrutura granular não pode ser man-
tida em reatores desnitrificantes com razões abaixo de DQO/NO−
3 de 5. Sendo assim, o
desenvolvimento do lodo com baixa velocidade de sedimentação, dificultou a operação dos

reatores. Por outro lado, durante a operação com uma razão DQO/NO−
3 maior que 10 as
propriedades dos grânulos mantiveram-se estável, com boa capacidade de sedimentação.

Huang et al. (2007) estudaram o efeito da razão de recirculação na RSCN em um sistema
anaeróbio-aeróbio (reator UASB seguido de lodo ativado) tratando águas residuárias de
cervejaria. O sistema foi continuamente alimentado com DQO variando de 1.600 a 2.700
mg/L, nitrogênio total variando de 320 a 416 mg N/L e razões de recirculação iguais a 100,
200 e 300%, respectivamente. Esse sistema mostrou-se bastante eficiente, com remoção de
carbono variando de 96 a 97 %, e nitrogênio total variando de 54 a 77%, respectivamente.
Os pesquisadores observaram que uma maior razão de recirculação resultou em uma alta
atividade de bactérias desnitrificantes, e, portanto, uma maior eficiência de remoção de
nitrogênio total.
Chen et al. (2009) estudaram efeito da razão de recirculação na RSCN em dois rea-
tores com biomassa imobilizada conectados em série, um reator anaeróbio seguido de um
reator aeróbio, com recirculação interna, tratando águas residuárias da indústria têxtil
(azo-compostos). Os resultados indicaram que a razão de recirculação tem pouco impacto
no desempenho global na remoção de carbono e na atividade de bactérias desnitrificantes
no reator anaeróbio. Para qualquer uma das razões estudadas, a eficiência de remoção de
carbono esteve entre 96 - 98%, com concentrações efluentes de DQO abaixo de 200 mg
O2 /L. No entanto, a taxa de nitrificação e remoção de nitrogênio total foram fortemente
impactada pela razão de recirculação. A taxa de nitrificação atingiu um valor máximo
igual a 0,48 kg N/m3 .dia para uma razão de recirculação igual a 1, enquanto que a com-
Seção 2.3. Remoção simultânea de carbono e nitrogênio em sistema integrado anaeróbio-anóxico 47
pleta nitrificação foi obtida com uma razão de recirculação igual a 2. Os autores também
observaram um aumento na eficiência de remoção de nitrogênio total e uma redução da
produção de lodo no sistema com um aumento da razão de recirculação.
Huiliñir et al. (2012) avaliaram o efeito da razão DQO/NO−
3 e da carga orgânica sobre
a RSCN usando águas residuárias do processamento de pescado. O estudo foi realizado

em um reator operado em bateladas com biomassa em suspensão e em um reator contı́nuo
de fluxo tubular alimentado com uma mistura de substrato sintético e efluente de uma
estação de tratamento de águas residuárias de processamento de pescado. Os resultados
dos experimentos em bateladas e contı́nuo mostraram que a desnitrificação foi maior que
95% para todas as razões DQO/NO−
3 estudadas sem efeitos na transformação de nitrato
e produção de amônia. Um aumento da concentração de nitrato, resultou em acréscimo

do consumo de matéria orgânica e consequentemente da taxa de hidrólise. A eficiência
de remoção de matéria orgânica variou entre 88 e 40% no processo contı́nuo. Os autores
observaram que para o processo contı́nuo o aumento da carga reduz a remoção da matéria
orgânica.
Xie et al. (2012) realizaram um estudo para avaliar o efeito da fonte de carbono e
da relação DQO/NO−
3 na RSCN em um reator anaeróbio tratando águas residuárias de
vinı́cola. Os resultados experimentais mostraram que para razões DQO/NO−

3 entre 3 e
55, a desnitrificação foi a principal via de redução de nitrato. O melhor resultado, em ter-
mos, de eficiência de remoção simultânea de carbono e nitrogênio, foi obtido para relação
DQO/NO−
3 de 7. Os autores observaram o acúmulo de intermediários da desnitrificação
com razão DQO/NO−

3 abaixo de 7 para ambas as fontes de carbono utilizadas, resultando
na inibição da atividade metanogênica. Esse acúmulo foi consequência da pouca disponibi-

lidade de matéria orgânica para realização do processo de desnitrificação. A principal rota
de consumo de matéria orgânica para razão DQO/NO−
3 acima de 7 foi a metanogênese
para os dois substratos utilizados. Segundo os autores, propionato e butirato foram os

principais AGVs utilizados pelas bactérias desnitrificantes.
A principal vantagem da RSCN em reatores anaeróbios tratando águas residuárias
concentradas em relação a águas residuárias diluı́das, é a maior disponibilidade de matéria
orgânica para realização dos processos simultâneos de digestão anaeróbia e desnitrificação.
Entretanto, alguns estudos já foram realizados buscando a RSCN em reatores anaeróbios
tratando águas residuárias diluı́das (Tai et al., 2006; Banihani et al., 2009; Kassab, 2009).
Tai et al. (2006) investigou o efeito do tempo de detenção hidráulica (TDH) e da razão
de recirculação na RSCN em um reator UASB seguido de um reator aeróbio tratando água
residuária sintética com caracterı́sticas semelhantes ao esgoto doméstico. A eficiência de
remoção de matéria orgânica no reator UASB aumentou de 87 para 92%, quando a razão de
recirculação aumentou de 1 para 4, para valor de TDH de 44 horas. A eficiência de remoção
de nitrogênio do sistema integrado aumentou de 49 a 86%, quando a razão de recirculação
aumentou de 1 para 4, para todos os TDH testados. Os autores observaram elevada
presença de nitrogênio no biogás com o aumento da razão de recirculação, enquanto, que
o conteúdo de metano no biogás apresentou uma tendência inversa, independentemente do
TDH. Os autores concluı́ram que a RSCN pode ocorrer no reator anaeróbio manipulando
o TDH e a razão de recirculação.
Banihani et al. (2009) estudaram a toxicidade de nitrato e nitrito nas Archaeas meta-
nogênicas a partir de um inóculo de lodo granular, oriundo de um reator UASB operado
em escala real, tratando águas residuárias domésticas e um lodo oriundo de um digestor
anaeróbio de uma estação de tratamento de águas residuárias domésticas. O ensaio de to-
xicidade foi conduzido em dois experimentos com diferentes doadores de elétrons: acetato
(2.000 mg O2 /L) e hidrogênio (1.500 mg O2 /L). O lodo do UASB e o lodo do digestor
foram adicionados em frascos de vidro de 160 mL com concentrações de 1.500 mg SSV/L
e 12.100 mg SSV/L, respectivamente. Os resultados obtidos na pesquisa indicaram que
o tipo de substrato teve um forte impacto na taxa de utilização de nitrato. A taxa de
utilização de hidrogênio pelas desnitrificantes foi mais lenta comparada com a taxa de uti-
lização de acetato pelas desnitrificantes. Consequentemente, houve um maior acúmulo de
intermediários da desnitrificação e maior inibição da atividade metanogênica. A produção
final de metano foi inversamente proporcional à quantidade de nitrato e nitrito adiciona-
dos, indicando que esses foram os aceptores de elétrons preferenciais comparando com a
metanogênese.
Kassab (2009) estudou a RSCN em um reator UASB operado com lodo floculento,
simulando um reator recebendo recirculação de um efluente nitrificado. O reator UASB
foi operado em escala de laboratório tratando águas residuárias diluı́das simulando es-
goto doméstico. A razão DQO/NO−
3 utilizada foi de 23, resultando em uma eficiência
de remoção de matéria orgânica e nitrogênio de 92 e 97%, respectivamente. A desnitri-

ficação foi a principal via de redução de nitrato. No entanto, 12% de nitrato aplicado,
foi reduzido via redução desassimilatória de nitrato a amônia. A integração dos proces-
sos de digestão anaeróbia e desnitrificação no reator UASB, resultou no consumo de 18%
da DQO para o processo de desnitrificação enquanto que o restante da DQO foi conver-
tido a metano. Na segunda parte da pesquisa, o autor avaliou o desempenho da RSCN
em reatores anaeróbio de leito granular expandido (EGSB) para o tratamento de esgoto
doméstico. O reator EGSB simulou o segundo estágio anaeróbio recebendo parte do eflu-
Seção 2.4. Anaerobic Digestion Model No. 1 (ADM1) 49
ente nitrificado do estágio aeróbio, a partir de um sistema de tratamento consistindo de

dois estágios anaeróbios seguido de um estágio aeróbio. Dois reatores EGSB em escala
de laboratório foram utilizados. O primeiro foi operado sob condições anaeróbia/anóxica
e o segundo foi operado como um reator controle sob condições estritamente anaeróbias.
O reator com os processos integrados foi operado com uma razão DQO/NO−
3 de 20. No
reator integrado, cerca de 67% da DQO foi utilizada para o processo da metanogênese.
Os autores atribuı́ram que a baixa utilização de DQO pelas metanogênese foi atribuı́do a
perda de biomassa no reator.
2.4 Anaerobic Digestion Model No. 1 (ADM1)
A modelagem do processo de digestão anaeróbia começou motivada pela demanda

da operação eficiente de sistemas anaeróbios no inı́cio década de 1970. Os modelos ma-
temáticos desenvolvidos inicialmente eram muito simples, devido ao conhecimento limitado
do processo (Andrews, 1969, 1971; Graef e Andrews, 1974). Com o avanço no conhecimento
computacional e do processo, foram desenvolvidos modelos mais complexos, tais como os
apresentados por Angelidaki et al. (1999); Tartakovsky et al. (2008); Bernard et al. (2001);
Siegrist et al. (2002). Pavlostathis e Giraldo-Gomez (1991); Lyberatos e Skiadas (1999);
Gavala et al. (2003) são alguns autores que realizaram trabalhos de revisão reportando a
evolução de modelos de digestão anaeróbia.
Os primeiros modelos foram baseados na metanogênese, considerada por alguns pes-
quisadores como a etapa limitante do processo Andrews (1969, 1971) e Graef e Andrews
(1974). Outros modelos consideraram a conversão de ácidos graxos em biogás ou a hidrólise
de sólidos em suspensão como etapa limitante (Eastman e Ferguson, 1981). Esses primeiros
modelos baseados em etapas limitantes eram simples e fáceis de aplicar, mas, não foram
capazes de descrever bem o processo, especialmente, sob condições de regime transiente.
A segunda geração de modelos considerou a concentração de ácidos orgânicos como
principal parâmetro, propondo cinco grupos de micro-organismos: um acidogênico (a partir
da glicose), dois acetogênicos (a partir do propionato e butirato) e dois metanogênicos (a
partir do acetato e do hidrogênio). Mosey (1983) incluiu duas formas de inibição biológica:
pelo pH (sofrida por todos os grupos de bactérias quando o pH < 6) e pela pressão parcial
de hidrogênio. O modelo predizia a instabilidade do processo anaeróbio quando ocorria
sobrecarga orgânica acarretando no acúmulo de ácidos graxos voláteis. A consequente
queda no valor do pH inibe as Archaeas metanogênicas hidrogenotróficas e aumenta a
pressão parcial do gás carbônico e do gás hidrogênio, provocando o acúmulo de propionato
e butirato. Além disso, a maior contribuição do seu modelo foi colocar, pela primeira
vez, a importância da pressão parcial de hidrogênio no meio como elemento regulador da
digestão anaeróbia.
Com base na abordagem proposta por Mosey (1983), foram desenvolvidos diversos
modelos matemáticos incluindo os de Costello et al. (1991); Angelidaki et al. (1993); Siegrist
et al. (1993); Vavilin et al. (1994, 1995); Kalyuzhnyi (1997a,b); Kalyuzhnyi et al. (1998);
Von Munch et al. (1999); Angelidaki et al. (1999); Batstone et al. (2000); Bernard et al.
(2001); Tartakovsky et al. (2002). Esses modelos incorporam o conhecimento sobre o
processo de digestão anaeróbia, incluindo mais subprocessos, espécies de micro-organismos
e equações cinéticas mais complexas considerando diferentes substratos.
Apesar do progresso significativo na modelagem de processos de digestão anaeróbia, no
ano de 1998 a International Water Association (IWA) criou um grupo de trabalho (TG-
MMADP) com a responsabilidade de propor um modelo de digestão anaeróbia abrangente,
que servisse como base para simulações dinâmicas do processo e uniformizar os parâmetros
e as variáveis dos modelos de digestão anaeróbia propostos anteriormente. O ADM1 foi pri-
meiro apresentado na 9a IWA Conference on Anaerobic Digestion, realizada na Antuerpia
no ano de 2001 e publicado pela IWA no seu relatório técnico e cientı́fico no ano 2002 (Bats-
tone et al., 2002a,b). O ADM1 foi proposto com o objetivo de ampliar o conhecimento,
aplicação, operação e otimização do processo de digestão anaeróbia e engloba fenômenos
biológicos, quı́micos, bioquı́micos e fı́sico-quı́micos decorrentes da digestão anaeróbia dos
compostos orgânicos complexos até produtos finais (CH4 e CO2 ) (Batstone et al., 2002a).
O ADM1 tem sido aplicado para controlar o processo de digestão anaeróbia de resı́duos
sólidos, lodo de esgoto (Batstone, 2006a,b; Huete et al., 2006; Kerroum et al., 2010) e
compostos orgânicos presentes em águas residuárias (Blumensaat e Keller, 2005; Mu et al.,
2007; Fezzani e Cheikh, 2008; Mu et al., 2008; Tartakovsky et al., 2008; Tejasen e Taruya-
non, 2010; Dereli et al., 2010; Garca-Diéguez et al., 2013).
O ADM1 é um modelo generalizado de digestão anaeróbia para diversos tipos de pro-
cessos, considerando os bioquı́micos como processos irreversı́veis e os fı́sico-quı́micos como
reversı́veis. Esse modelo inclui: 19 processos bioquı́micos de taxas e de transferência da
fase gasosa para lı́quida, com 105 parâmetros cinéticos e estequiométricos; e 6 proces-
sos cinéticos adicionais do equilı́brio ácido-base, quando considerado como conjunto de
equações diferenciais ordinárias.
As Tabelas 2.2 e 2.4 apresentam as variáveis de estado estruturadas na forma matricial
com coeficientes para reações bioquı́micas e cinéticas. Na primeira coluna da esquerda
têm-se os processos j de 1 a 19. Os processos de 1 a 12 foram descritos no item 2.1, e os
processos de 13 a 19 representam os decaimentos dos sete grupos de bactérias. Na primeira

linha das Tabelas 2.2 e 2.4 estão as variáveis i da fase lı́quida. O termo da cinética de
cada estado i é obtido com a somatória dos termos de sua coluna i (νi,j ), multiplicados
por suas respectivas taxas (ρj ), Tabela 2.5 . Os parâmetros f representam quais as frações
formadas numa determinada etapa da degradação de compostos orgânicos; e N significa a
percentagem de nitrogênio presente nos aminoácidos (Naa ) e na composição dos 7 grupos
de bactérias (Nbac ). No estado i=10 (SIC ), cada somatória corresponde a quantidade de
carbono inorgânico envolvida no respectivo processo j.
2.4.1 Processos bioquı́micos
O modelo ADM1 abrange as etapas de solubilização extracelular e intracelular. As eta-

pas de desintegração e hidrólise são processos extracelulares, e as etapas da acidogênese,
acetogênese e metanogênese, são processos intracelulares. Os processos extracelulares
são responsáveis pela quebra e solubilização da matéria orgânica complexa em substrato
solúvel. Nos processos intracelulares, os compostos solúveis gerados na etapa da hidrólise
são convertidos em metano, gás carbônico, água, gás sulfı́drico, amônia e novas células
bacterianas. O modelo conceitual proposto pelo TG-MMADP para processos da digestão
anaeróbia é mostrado na Figura 2.6. O modelo ADM1 considerou os seguintes processos
bioquı́micos: (1) acidogênese de açúcares; (2) acidogênese de aminoácidos; (3) acetogênese
de LCFA; (4) acetogênese de propionato; (5) acetogênese de butirato (Hbu) e valerato
(Hva); (6) metanogênese acetoclástica; e (7) metanogênese hidrogenotrófica.
A desintegração possibilita a conversão de substrato particulado e agregado, em iner-

tes, carboidratos, proteı́nas e lipı́dios. A inclusão dessa etapa no ADM1 foi considerada
inovadora em relação aos outros modelos apresentados anteriormente, pois permite diversi-
ficar a aplicação do modelo para diversas unidades de tratamento, representar à quebra de
moléculas complexas presentes no lodo biológico e do material orgânico complexo (em car-
boidratos, proteı́nas e lipı́deos) e recircular a biomassa anaeróbia inerte. A desintegração
pode ser aplicada em unidades de tratamento com matéria orgânica de difı́cil degradação
como é o caso de lodo biológico, resı́duos sólidos, etc (Batstone, 2006c).
O segundo processo de solubilização extracelular é a hidrólise enzimática. Nesse pro-
cesso, as bactérias fermentativas hidrolı́ticas excretam exoenzimas que são responsáveis
pela degradação da matéria orgânica complexa (polı́meros), em materiais dissolvidos me-
nores (monômeros), os quais podem atravessar as paredes celulares das bactérias fermenta-
tivas. As proteı́nas são degradas para formar aminoácidos, os carboidratos se transformam
Tabela 2.2 - Matriz estequiométrica para compostos solúveis (i = 1-12, j = 1-19). * kgDQO.m−3 ** kmolC.m−3 *** kmolN.m−3
Componentes → i 1 2 3 4 5 6 7 8 9 10 11 12
j Processos ↓ Ssu Ssa Sfa Sva Sbu Spro Sac Sh2 Sch4 Sic Sin Si
1 Desintegração fsi,xc
2 Hidrólise de carboidrato 1
3 Hidrólise de proteı́na 1
4 Hidrólise de lipı́deo 1- ff a,li ff a,li
P
5 Consumo de Ssu -1 (1-Ysu )fbu,su (1-Ysu )fpro,su (1-Ysu )fac,su (1-Ysu )fh2,su i=9,11−24 Ci νi,5 -(Ysu ) Nbac
P
6 Consumo de Saa -1 (1- Yaa ) fva,aa (1-Yaa ) fbu,aa (1-Yaa )fpro,aa (1-Yaa ) fac,aa (1-Yaa )fh2,aa i=9,11−24 Ci νi,6 Naa - (Yaa )Nbac
7 Consumo de Sfa -1 (1-Yf a )0,7 (1-Yf a )0,3 -(Yf a ) Nbac
Capı́tulo 2. Revisão da literatura
8 Consumo de Sva -1 (1-Yc4 )0.54 (1-Yc4 )0,31 (1-Yc4 )0,15 -(Yc4 ) Nbac
9 Consumo de Sbu -1 (1-Yc4 )0,8 (1-Yc4 )0,2 -(Yc4 ) Nbac
P
10 Consumo de Spro -1 (1-Ypro )0,57 (1-Ypro )0,43 i=9,11−24 Ci νi,10 -(Ypro ) Nbac
P
11 Consumo de Sac -1 (1-Yac ) i=9,11−24 Ci νi,11 -(Yac ) Nbac
P
12 Consumo de Sh2 -1 (1-Yh2 ) i=9,11−24 Ci νi,12 -(Yh2 ) Nbac
13 Decaimento de Xsu
14 Decaimento de Xaa
15 Decaimento de Xfa
16 Decaimento de Xc4
17 Decaimento de Xpro
18 Decaimento de Xac
19 Decaimento de Xh2
*** Nitrogênio inorgânico

** Carbono inorgânico
* Monossacarı́eo
* Ácidos graxos
* Solúvel inerte
* Aminoácidos
* Propionato
* Hidrogênio
* Butirato
* Valerato
* Acetato
* Metano
52
Tabela 2.3 - Matriz estequiométrica para compostos solúveis (i = 13-26, j = 1-19). * kgDQO.m−3 ** kmol.m−3
Tabela 2.4 -
Componentes → i 13 14 15 16 17 18 19 20 21 22 23 24 25 26
j Processos ↓ Xc Xch Xpr Xli Xsu Xaa Xfa Xc4 Xpro Xac Xh2 Xi Scat San
1 Desintegração -1 fch,xc fpr,xc fli,xc fxi,xc
2 Hidrólise de carboidrato -1
3 Hidrólise de proteı́na -1
4 Hidrólise de lipı́deo -1
5 Consumo de Ssu Ysu
6 Consumo de Saa Yaa
7 Consumo de Sfa Yaa
8 Consumo de Sva Yc4
9 Consumo de Sbu Yc4
10 Consumo de Spro Ypro
11 Consumo de Sac Yac
12 Consumo de Sh2 Yh2
13 Decaimento de Xsu 1 -1
14 Decaimento de Xaa 1 -1
15 Decaimento de Xfa 1 -1
Seção 2.4. Anaerobic Digestion Model No. 1 (ADM1)
16 Decaimento de Xc4 1 -1
17 Decaimento de Xpro 1 -1
18 Decaimento de Xac 1 -1
19 Decaimento de Xh2 1 -1
de
de
de
de
de
Degradadores
Degradadores
Degradadores
Degradadores
Degradadores
53
* Compostos
* Carboidrato
* Proteı́na
* Lipı́deos
*
monossacarı́deo
*
aminoácidos
*
AGCL
*
valerato e butirato
*
propionato
* Degradadores de
acetato
* Degradadores de hi-
drogênio
* Particulados inerte
** Cation
** Anion
Tabela 2.5 - Processos e taxas de reações correspondentes do ADM1

j Processos ↓ Taxa cinética (ρj , kgDQO.m−3 .d−1 )
1 Desintegração kdis Xc
2 Hidrólise de carboidrato khyd,ch Xch
3 Hidrólise de proteı́na khyd,pr Xpr
4 Hidrólise de lipı́deo khyd,li Xli
5 Consumo de Ssu km,su KsS+S
su
su
Xsu I1
6 Consumo de Saa km,aa KsS+S
aa
aa
Xaa I1
Sfa
7 Consumo de Sfa km,f a Ks +Sfa Xfa I2
8 Consumo de Sva km,c4 KsS+S
va
va
Xc4 1+Sbu1 /Sva I2
9 Consumo de Sbu km,c4 KsS+S
bu
bu
Xc4 1+Sva1 /Sbu I2
Spro
10 Consumo de Spro km,pr Ks +S pro
Xpro I2
Sac
11 Consumo de Sac km,ac Ks +Sac Xac I3
12 Consumo de Sh2 km,h2 KsS+S
h2
h2
Xh2 I1
13 Decaimento de Xsu kdec,Xsu Xsu
14 Decaimento de Xaa kdec,Xaa Xaa
15 Decaimento de Xfa kdec,Xf a Xfa
16 Decaimento de Xc4 kdec,Xc4 Xc4
17 Decaimento de Xpro kdec,Xpro Xpro
18 Decaimento de Xac kdec,Xac Xac
19 Decaimento de Xh2 kdec,Xh2 Xh2
Fator de inibição:
I1 = IpH IIN,lim
I2 = IpH IIN,lim Ih2
I2 = IpH IIN,lim INH3,Xac
Composto orgânico complexo e biomassa ativa
Inertes insolúveis
Desintegração
Carboidratos Proteínas Lipídeos Inerte solúvel
Hidrólise
Monossacaríeos Aminoácidos LCFA

1 2 3
Acidogênse
Propianato Hva, HBu

5
Acetogênese 4
Acetato H2
Morte, decaimento
Metanogênese 6 7
CH4 ,CO2
Figura 2.6: Modelo ADM1 Fonte: Adaptado de Batstone et al. (2002a)
em monossacarı́deos e os lipı́deos são convertidos em ácidos graxos de cadeia longa e gli-

cerina.
O ADM1 considerou dois grupos de micro-organismos acidogênicos que fermentam mo-
nossacarı́deo e aminoácidos e excretam substâncias orgânicas mais simples (propionato,
butirato, valerato, acetato e hidrogênio), além de novas células bacterianas sem a presença
de um aceptor adicional de elétrons. A degradação de ácidos graxos de cadeia longa é uma
reação de oxidação com aceptor externo de elétrons e, portanto, foi inclusa na etapa de ace-
togênese. O modelo ADM1 adotou apenas os ácidos, acetato, propionato e butirato como
produtos finais da acidogênese de monossacarı́deos, devido a sua degradação em vias dife-
rentes no processo de digestão anaeróbia e são medidos simultaneamente por cromatografia
gasosa. O lactato e o etanol, produtos da acidogênese da glicose, não foram contemplados
no ADM1 devido às baixas concentrações encontradas nos reatores anaeróbios (Batstone
et al., 2002a).
O modelo ADM1 considerou butirato, valerato, propionato e LCFA como substratos
utilizados pelos micro-organismos acetogênicos. Butirato e valerato são substratos de-
gradados pelo mesmo micro-organismo, enquanto que LCFA é degradado por um micro-
organismo especı́fico, devido às dificuldades de transporte e diferenças das caracterı́sticas
fı́sico-quı́micas dessas macromoléculas. A degradação de ácidos graxos de cadeia longa em
acetato pelos micro-organismos acetogênicos é um processo de oxidação sem a presença de
aceptor interno de elétrons. Portanto, esses micro-organismos necessitam de aceptor adici-

onal de elétrons (ı́ons de hidrogênio e dióxido de carbono) para produzir gás hidrogênio ou
formiato. O modelo cinético de inibição não-competitiva pela presença de hidrogênio foi
adotado na acetogênese para reduzir a complexidade e aumentar a flexibilidade do ADM1.
O ADM1 considerou dois subgrupos principais em função da magnitude na produção
de metano e de sua afinidade por substrato: as Archaeas metanogênicas acetoclásticas e as
Archaeas metanogênicas hidrogenotróficas. O decaimento celular é descrito como produção
de material inerte e de matéria orgânica degradada, o qual retorna a etapa de desintegração.
O modelo cinético de primeira ordem foi adotado para as etapas extracelulares de desin-
tegração e hidrólise enzimática (Batstone et al., 2002a). Para os processos intracelulares,
o modelo cinético é baseado na cinética clássica desenvolvida por Monod para possibilitar
a separação dos termos de crescimento e consumo de substrato. A respiração endógena
ou decaimento de micro-organismos foi representado, também, pelo modelo cinético de
primeira ordem.
2.4.1.1 Modelo de inibição e toxicidade
Do ponto de vista de engenharia, a inibição afeta a velocidade de crescimento e decai-

mento celular por substâncias que entram ou se formam no próprio reator, ou substâncias
que passam para o interior das células interferindo nos processos metabólicos. Os modelos
cinéticos de inibição considerados no ADM1 são:
(a) aqueles reversı́veis com o uso extensivo do modelo de inibição não-competitiva;
(b) impacto direto do inibidor no crescimento e decaimento de biomassa;
(c) duas fórmulas empı́ricas usadas para inibição pelo valor de pH do meio;
(d) consumo competitivo, considerado como complemento apesar de não ser modelo de
inibição; e
(e) modelo cinético de Monod para substrato secundário necessário para descrever a
diminuição do crescimento celular pela limitada concentração de nitrogênio (incluı́do
como complemento do modelo e não como modelo de inibição).
2.4.2 Processos fı́sico-quı́micos
O ADM1 considera dois processos não biológicos que comumente ocorrem em reatores
anaeróbios: processo lı́quido-lı́quido (associações e dissociação iônica) e o processo lı́quido-
gás (transferência da fase gasosa para lı́quida). O processo sólido-lı́quido (precipitação
e solubilização) não foi incluı́do no ADM1 devido à complexidade de descrição, apesar

da importância em sistemas com formação de precipitados de cálcio (Ca2+ ) e magnésio
(Mg2 +).
2.4.2.1 Processo lı́quido-lı́quido
Os processos de associação e dissociação iônica são frequentemente referidos como pro-

cessos de equilı́brio, e podem ser descritos por meios de equações algébricas para calcular
as concentrações de ı́ons de hidrogênio, amônia livre, AGV e dióxido de carbono. Os três
principais componentes do processo fı́sico quı́mico de stripping dos compostos gasoso são:
hidrogênio, metano e gás carbônico. Esses componentes representam a produção de biogás.
No ADM1 os diversos equilı́brios entre as várias espécies quı́micas, incluindo o equilı́brio
ácido-base, são considerados como processo lı́quido-lı́quido. A modelagem do balanço de
carga é expressa como a diferença entre as somatórias das concentrações de cátions (SC + )
e ânions (S−
A ):
X X
SC + − S A− . (2.7)
A concentração equivalente de cada ı́on é a sua valência multiplicada pela concentração

molar. Os coeficientes dos processos do balanço de carga iônica, de carbono inorgânico e
nitrogênio inorgânico são representados pelas Equações 2.8, 2.9 e 2.10, respectivamente.
Para cada uma destas espécies quı́micas, as equações de balanço de cargas iônica são
expressas como:
SAc− SP r− SBu− SV a−
SCat+ + SN H4+ + SH + − SHCO3− − − − − − SOH − − SAn− = 0 , (2.8)
64 112 160 208
SIC − SCO2 − SHCO3− = 0 , (2.9)
SIN − SN H3 − SHCO4− = 0 . (2.10)
2.4.2.2 Processos lı́quido-gás
Os três gases considerados significativos na digestão anaeróbia, sejam como inter-

mediários ou exercendo fortes efeitos nos processos biológicos foram: hidrogênio, metano e
dióxido de carbono. A Lei de Henry foi utilizada na modelagem dos estados gasosos para
descrever a relação de equilı́brio entre as concentrações nas fases lı́quida e gasosa (Equação
2.11) e deu origem às equações que traduzem o fluxo de gás entre o volume livre do reator
e o lı́quido.
KH pgas,i,ss − Sliq,i,ss = 0 , (2.11)
onde, Sliq,i,ss é a concentração do componente i (M) na fase lı́quida no estado estacionário,

pgas,i,ss é a pressão parcial do componente i na fase gasosa (bar) e KH é a constante da lei
de Henry (M.bar−1 )
2.4.3 Processos excluı́dos do ADM1
Durante a elaboração do ADM1, alguns processos que podem ocorrer em um reator

anaeróbio foram excluı́dos, na maioria dos casos devido à limitada informação disponı́vel
na literatura àquela época (Batstone et al., 2002a). Dentre esses processos destacam-se:
degradação de produtos alternativos da glicose, tais como, o lactato e o etanol; redução
biológica de sulfato e inibição do processo pelo H2 S; inibição por ácidos graxos de cadeia
longa; redução biológica de nitrato; oxidação do acetato em duas vias distintas, onde
o acetato é oxidado a CO2 e H2 por um micro-organismo e o H2 é subsequentemente
utilizado por um segundo organismo para reduzir CO2 a CH4 ; consumo competitivo de
H2 e de CO2 entre as Archaeas metanogênica hidrogênotróficas e os micro-organismos
homoacetogênicos; e precipitação de sólidos devido à alcalinidade elevada ou a outras
reações de precipitação quı́mica.
2.4.4 Implementação do ADM1 em um reator CSTR
O modelo ADM1 descreve o processo de digestão anaeróbia em um reator de mistura

completa com fluxo contı́nuo. O modelo pode ser implementado de duas diferentes for-
mas: sistema de equações diferenciais ordinárias (EDOs) e sistema de equações algébrico-
diferenciais (EADs). No presente trabalho, o modelo foi implementado como EDOs.
2.4.4.1 Modelagem da fase lı́quida
Na fase lı́quida, as equações dos estados são balanços de massas, contendo termos
cinéticos que caracterizam as velocidades especı́ficas de conversão. As Equações 2.12 e
2.13 representam o cálculo das variáveis em estado solúvel e particulado, respectivamente.
19
dSliq,i q X
= (Sin,i − Sliq,i ) + ρj νi,j i = 1, ..., 12 , i = 25 − 26 , (2.12)
dt Vliq j=1
19
dXliq,i q X
= (Xin,i − Xliq,i ) + ρj νi,j i = 13, ..., 24 , (2.13)
dt Vliq j=1
onde, Sliq,i é a concentração do material solúvel na fase lı́quida, Xliq,i é a concentração

do material particulado e biomassa fase lı́quida, Vliq corresponde ao volume lı́quido do
reator, q corresponde a vazão afluente e efluente do reator, Sin,i é a concentração afluente
do material solúvel, Xin,i é a concentração afluente do material particulado e biomassa.
A cada ρj corresponde a taxa especı́fica cinética para o processo j, multiplicado pelo
coeficiente estequiométrico (νi,j ).
2.4.4.2 Modelagem da fase gasosa
A equação dinâmica para cada um dos subprodutos gasosos considerados no ADM1, é

dada por:
dSgas,i Sgas,i qgas Vliq

= + ρT,i i = CH4 , CO2 e H2 , (2.14)
dt Vgas Vgas
onde qgas é a vazão de gás (m3 .d−1 ); Vgas é o volume de gás no reator; Sgas,i (mol.L−1 ) é
a concentração de gás na fase gasosa do componente i e ρT,i é a taxa de transferência de
massa do gás i, dada por:
ρT,i = kL a(Sliq,i − KH pgas,i ) i = CH4 , CO2 e H2 , (2.15)
onde kL a é o coeficiente de transferência de massa global multiplicada pela área de trans-

ferência especifica. A pressão parcial dos gases é obtida pelas Equações 2.16, 2.17 e 2.18.
As Equações 2.16 e 2.17 estão sujeitas a correções de KH do hidrogênio e metano pelos
fatores 16 e 64, respectivamente, com a finalidade de contabilizar os resultados em termo
de DQO.
RT
pgas,H2 = Sgas,H2 , (2.16)
16
RT
pgas,CH4 = Sgas,CH4 , (2.17)
64
pgas,CO2 = Sgas,CO2 RT . (2.18)
Substituindo a pressão de vapor a 298K por 0,0313 bar e ∆H◦ vap por 43800 J.mol−1 ,
obtêm-se a seguinte correção:

1 1
pgas,H2 O = 0, 00313 exp 5290 − . (2.19)
298 T
A pressão parcial total dos gases é estimada pela Equação 2.20.
Pgas = pgas,H2 + pgas,CH4 + pgas,CO2 + pgas,H2 O . (2.20)
A vazão de biogás é dada pela transferência total para a fase gasosa, corrigida pelo
vapor de água é dada por:
RT ρ
T,H2 ρT,CH4
qgas = Vliq + + ρT,CO2 . (2.21)
Pgas − pgas,H2 O 16 64
2.4.4.3 Estimativa do pH
O valor do pH no ADM1 é estimado a partir da seguinte equação:
pH = −log10 [SH + ] , (2.22)
onde SH + é a concentração de ı́ons de hidrogênio obtida através da Equação 2.8. Conside-

rando que SOH − representa a concentração de ı́ons hidroxilas, a implementação na forma
algébrica implı́cita no modelo ADM1 é dada por:
KW
SOH − = , (2.23)
SH +
assumindo que:
SAc− SP r − SBu− SV a−
Θ = SCat+ + SN H4+ − SHCO3− − − − − − SAn− . (2.24)
64 112 160 208
A Equação 2.8 que representa o balanço de carga pode ser resolvida como:
√
Θ2 + Θ2 + 4KW
SH + = . (2.25)
2
2.5 Extensões do modelo ADM1
Extensões foram desenvolvidas para incorporar alguns processos ausentes no modelo

ADM1 para determinados tipos de águas residuárias. A Tabela 2.6 mostra as referências
publicadas entre os anos de 2003 e 2015, que descrevem as principais extensões e aplicações
do modelo ADM1.
Seção 2.5. Extensões do modelo ADM1 61
Tabela 2.6 - Extensões e aplicações do ADM1.

Extensão/adaptação do ADM1 Aplicação Referência
Precipitação de CaCO3 Vários * Batstone e Keller (2003)
Acetogênese de isovalerato Substrato rico em Batstone et al. (2003)
proteı́na
Redução de sulfato a H2 S utilizando propionato, acetato, bu- Substrato rico em sul- Fedorovich et al. (2003)
tirato e hidrogênio como doador de elétrons fato
Inclusão do decaimento de inertes nos produtos Vários* Wett et al. (2006)
Compostos aromáticos metoxilados, o metil mercaptana, con- Substrato rico em sul- Parker e Wu (2006)
versão do dimetilssulfeto em sulfeto de hidrogênio precipitação fato
de sulfeto metálico.
A oxidação do nitrato, nitrito, óxido nı́trico, óxido nitroso e Vários * Tugtas et al. (2006)
nitrogênio por bactérias degradadoras de propionato e buti-
rato ou valerato.
Influência do pH e hidrogênio na estequiometria da fer- Vários * Rodrı́guez et al. (2006)
mentação de carboidratos
Lactato e etanol como intermediários do processo de digestão Produção de biohi- Peiris et al. (2006)
anaeróbia drogênio
Degradação de di-etilhexilftalato (baseado na cinética de Lodo Fountoulakis et al. (2006)
sorção-dessorção)
Produção e reserva de polı́meros pelos micro-organismos Digestão com rápidas Shimada et al. (2007)
condições de mudança
Variabilidade na estequiometria de acordo com modelo pro- Degradação de lac- Penumathsa et al. (2008)
posto por Rodrı́guez et al. (2006). tato Produção de
biohidrogênio
Inibição das Archaeas metanogênicas acetoclástica por ácidos Vários* Fezzani e Cheikh (2008)
graxos voláteis.
Hidrólise e consumo de cianeto, inibição de cianeto sobre Resı́duos contendo cia- Zaher et al. (2009)
Archaeas metanogênicas acetoclásticas e classificação desses netos. Pode ser genera-
micro-organismos resistentes e não resistentes a cianeto. lizado a outros tóxicos
Inclusão do modelo de Contois para os processos de desin- Lodo Ramirez et al. (2009)
tegração e hidrólise substituindo a cinética de primeira or-
dem. A função de Hill é utilizada para modelar a inibição
por amônia sobre as Archaeas metanogênicas acetoclásticas
substituindo uma função de inibição não-competitivo.
Inserção dos processos de degradação anaeróbia de compostos Substrato rico em Fezzani e Cheikh (2009)
fenólicos e homólogos. fenóis.
Cinética de Haldane para consumo de LCFA. Inibição das Substrato rico em Palatsi et al. (2010)
Archaeas metanogênicas hidrogênotróficas e bactérias ace- lipı́dio
togênica por LCFA (não competitiva).
Desintegração, hidrólise e acidogênese como única etapa do Digestão seca Bollon et al. (2011)
processo de digestão anaeróbia.
Modelo de Contois para hidrólise. Resı́duos particulados Mairet et al. (2011)
Cinética da hidrólise baseadas em superfı́cie para a desinte- Resı́duos particulados Esposito et al. (2011)
gração e conversão de carboidratos, proteı́na e lipı́dios.
Sódio como uma variável extra. Inibição não competitiva de Substrato rico em Hierholtzer e Akunna (2012)
sódio sobre as Archaeas metanogênicas acetoclásticas. sódio.
Incluiu a fermentação de ácido lático contendo a degradação e Águas residuárias con- Hinken et al. (2014)
formação de lactato e uma função de inibição não-competitiva. centradas.
Predição de compostos de enxofre nas fases liquida e gasosa. Águas residuárias con- Barrera et al. (2015)
centradas.
Incluiu o consumo de AGVs pelos organismos acumuladores Águas residuárias con- Wang et al. (2015)
de polifosfato, o efeito do Polihidroxialcanoatos na taxa de centradas.
desintegração e a precipitação do fósforo
* Vários: tratamento de águas residuárias, resı́duos sólidos, lodo, etc
2.5.1 Extensão do ADM1 para processo de desnitrificação
Tugtas et al. (2006) desenvolveram uma extensão incluindo quatro etapas do processo
de desnitrificação (NO− −
3 , NO2 , NO, N2 O e N2 ) no modelo ADM1. Os resultados das si-
mulações foram comparados com dados obtidos a partir de uma série de experimentos con-
duzidos em um reator operado em bateladas utilizando cultura mista de micro-organismos
(fermentativos e metanogênicos). O reator foi alimentado com uma mistura de dextrina e
peptona e alterada com várias concentrações de nitrato.
O processo de redução desassimilatória do nitrato a ı́on amônio não foi considerado
na extensão por causa das condições experimentais (ausência de sulfeto e baixa relação
DQO/NO−
3 ). Durante a realização do experimento, os autores observaram redução de
nitrato na ausência de N-óxido. Portanto, nessa extensão do ADM1 foi assumido que
a desnitrificação foi realizada por um grupo de micro-organismos capazes de utilizar do-
adores de elétrons na ausência desses aceptores (chamado pelo autor de desnitrificantes
fermentativas). Os substratos utilizados (doadores de elétrons) pelos micro-organismos
desnitrificantes na presença de N-óxido foram butirato, propionato, acetato e hidrogênio.
Na ausência de N-óxido, os micro-organismos utilizaram butirato, valerato e propionato.
A utilização de açúcares, aminoácidos, ácidos graxos de cadeia longa pelas desnitrificantes
foi negligenciado por causa da cinética mais lenta da utilização de glicose e ácido glutâmico
por desnitrificantes. O efeito inibitório de N-óxido sobre as Archaeas metanogênicas foi si-
mulado utilizando uma função de inibição não-competitiva. Entretanto, o efeito inibitório
de N-óxido nos micro-organismos fermentativos e desnitrificantes foi negligenciado devido
a pouca informação disponı́vel na literatura.
A composição de alimentação do ADM1 foi ajustada para refletir as condições dos subs-
tratos utilizados no experimento realizado. Portanto, foram realizadas duas modificações,
a saber: o fracionamento da matéria orgânica e a distinção entre a constante cinética da
hidrólise de decaimento de biomassa e substrato afluente. Após a etapa da desintegração,
assumiu-se que o fracionamento do decaimento da biomassa resulta em 80% de matéria
orgânica biodegradável (10,4% carboidrato, 66,4% proteı́na, e 3,2% lipı́deo) e 20% de ma-
terial inerte (10% solúvel e 10% particulado). A hidrólise do substrato e decaimento de
biomassa foram considerados dois processos separados: a) decaimento de biomassa, com
constante de desintegração igual a 2 d−1 e constante de hidrólise da degradação de car-
boidrato, proteı́na e lipı́deo iguais a 0,15, 0,5, 0,15 d−1 , respectivamente; b) alimentação,
com constante de hidrólise da degradação de carboidrato e proteı́na iguais a 2 d−1 . A
etapa de desintegração foi excluı́da da alimentação, devido a caracterı́stica do substrato
Seção 2.6. Metodologias e estratégias para aplicação do modelo ADM1 63
utilizado nesse estudo (dextrina e peptona). A velocidade máxima especı́fica de consumo

do propionato (km,pro ) e acetato (km,ac ) foi estimada a partir de princı́pios bioenergéticos
e a constante de meia saturação de propionato (KS,pro ) foi obtida a partir do ajuste do
modelo aos dados experimentais.
A cinética da desnitrificação com seus respectivos doadores de elétrons (mistura de
dextrina/peptona, glicose, propionato, acetato, e mistura de H2 /CO2 ) foram obtidas expe-
rimentalmente (Tugtas e Pavlostathis, 2007a). A velocidade especı́fica de consumo dos doa-
dores de elétrons pelos micro-organismos desnitrificantes (km,nox ) não pode ser determinado
experimentalmente devido a utilização simultânea desses compostos pelas desnitrificantes e
fermentativos. Portanto, os valores de km,nox foram correlacionados estequiometricamente
com a taxa de utilização de nitrato. Um coeficiente (f , kg substrato DQO/ kg N), que
converte a taxa de utilização de N-óxido em velocidade especı́fica de consumo dos doa-
dores de elétrons pelos micro-organismos desnitrificantes, foi obtida pela razão molar de
doadores de elétrons para N-óxido.
Os valores das constantes de meia saturação para utilização dos doadores de elétrons e
das constantes de meia saturação para utilização dos aceptores de elétrons reportados na
literatura Zumft (1997), foram utilizados como estimativa inicial para ajuste do modelo
aos dados experimentais. O valor da constante de meia saturação estimada para utilização
de nitrato estava compreendido entre 78 e 171 mg N/L. A literatura reporta uma faixa
de valores inferiores, entre 4 - 53 mg N/L (Zumft, 1997). Os autores atribuı́ram esses
resultados a fatores relacionados ao experimento realizado. Portanto o valor de KS,no3
assumido na extensão do modelo foi 4,2 mg N/L. Estimativas dos valores assumidos para
outros parâmetros cinéticos estão descritos com mais detalhes em Tugtas et al. (2006).
Apesar do avanço na proposição da extensão do ADM1, os autores não consideraram
variáveis importantes para avaliar a ocorrência dos processos simultâneos de fermentação,
metanogênese e desnitrificação como pH, e o efeito do TDH e da razão DQO/NO−
3 na
RSCN em reatores anaeróbio-anóxico.
2.6 Metodologias e estratégias para aplicação do modelo ADM1
Muitas crı́ticas feitas ao ADM1 estão baseadas na complexidade ou simplicidade do

modelo Bernard et al. (2006); Rodrı́guez et al. (2008) e necessidade da caracterização
detalhada do afluente como variáveis de entrada (monossacarı́deos, aminoácidos, ácidos
graxos de cadeia longa, etc.) Kleerebezem e van Loosdrecht (2006); Girault et al. (2012).
Nas duas próximas sessões são apresentados, alguns métodos para simplificação e caracte-
rização do afluente.
2.6.1 Simplificações do ADM1
Uma das principais limitações na aplicação do ADM1 é a necessidade de determinação

experimental de uma grande quantidade de parâmetros cinéticos e variáveis de estado.
Entretanto, esforços têm sido realizados nos últimos anos para simplificação estrutural do
ADM1 Bernard et al. (2006); Rodrı́guez et al. (2008); Garca-Diéguez et al. (2013).
Existem diferentes métodos que podem ser aplicados para a redução de modelos, que,
por sua vez, são divididos, principalmente, em métodos de projeção, tais como os métodos
baseados na decomposição em valores singulares ou decomposição ortogonal adequada,
e métodos de não-projeção, tal como perturbação singular. Alguns trabalhos têm sido
desenvolvidos sobre a ampliação ou aplicação desses métodos para simplificação de modelos
de tratamento de águas residuárias Steffens et al. (1997); Chachuat et al. (2003); Bernard
e Bastin (2005a,b); Barrou et al. (2008).
Bernard et al. (2006) desenvolveram uma metodologia para determinar a estrutura de
uma matriz K pseudo-estacionária de um modelo baseado em equações de balanço de
massa. A estrutura dessa matriz consiste na estimativa do número de grupos de micro-
organismos independentes que representem a transferência de massa que ocorre dentro do
reator, fornecendo assim a dimensão da matriz K. O método foi aplicado a dois modelos
proposto por Bernard et al. (2001) que consideram um e dois grupos de micro-organismos,
respectivamente, e no modelo ADM1 que considera sete grupos de micro-organismos. Fo-
ram utilizados dados experimentais de um reator anaeróbio de leito fixo operado em escala
piloto. Os resultados mostraram que o modelo com um e dois grupos de micro-organismo
explicam 83,2% e 97,8% da variância total dos dados, respectivamente. Essa análise indi-
cou uma boa concordância entre os dados experimentais e o modelo simulado com apenas
um grupo de micro-organismos envolvido no processo. O modelo ADM1 apresentou resul-
tados similares àqueles obtidos experimentalmente. Bernard et al. (2006) ressaltaram que
a utilização da metodologia proposta para simplificar modelos complexos é uma ferramenta
poderosa para reproduzir algumas variáveis do processo, reduzindo o número de longos e
dispendiosos experimentos a serem realizados.
Rodrı́guez et al. (2008) utilizaram a técnica de análise de componentes principais (ACP)
para avaliar o número mı́nimo de reações que deveriam ser consideradas na estrutura da
extensão do modelo ADM1, incluindo o processo de degradação de etanol. A ACP foi
aplicada aos dados experimentais, de modo a avaliar o número de reações necessárias para
Seção 2.6. Metodologias e estratégias para aplicação do modelo ADM1 65
descrever a variabilidade dos dados em termos de sete variáveis medidas: concentrações

de etanol, butirato, propionato e acetato e dos gases hidrogênio, dióxido de carbono e
metano. Esses dados foram utilizados para ajuste do modelo aos dados experimentais. Os
resultados obtidos sugerem que apenas quatro grupos de micro-organismos são necessários
para descrever o processo com 85,46% da variabilidade dos dados experimentais.
Como mesmo objetivo, Garca-Diéguez et al. (2013) aplicaram ACP para reduzir o
número de reações do modelo ADM1 com base na metodologia desenvolvida por Bernard e
Bastin (2005a), Bernard e Bastin (2005b), Bernard et al. (2006) e Rodrı́guez et al. (2008). A
Tabela 2.7 mostra o número de parâmetros, variáveis de estado, tipos de micro-organismos,
reações e saı́das do modelo ADM1 em relação ao modelo.
Tabela 2.7 - Comparação da complexidade entre o ADM1 e o modelo simplificado (Garca-Diéguez et al.,
2013)
ADM1 Modelo reduzido
Parâmetros 89 13
Variáveis de estado 28 7
Tipos de micro-organismos 8 2
Número de reações 21 2
Saı́das 32 8
Razão da complexidade 7/89 = 0,08 7/13=0,54
A razão da complexidade é definida como a razão entre o número de medições dis-

ponı́veis e o número de parâmetros. Uma alta razão de complexidade indica melhor sim-
plificação do modelo. Garca-Diéguez et al. (2013) constatou através dos resultados que
apenas dois grupos de micro-organismos (acidogênico e metanogênico) explicam 90% da
variância total dos dados.
2.6.2 Métodos para caracterização do afluente
A caracterização do afluente nas frações exigidas pelo ADM1 é baseada na identificação

de numerosas concentrações de componentes especı́ficos. No ADM1, a DQO afluente é di-
vidida em 13 variáveis de estado incluindo, onze frações biodegradáveis: compostos (Xc ),
polı́meros - carboidratos (Xch ), proteı́nas (Xpr ), e lipı́deos (Xli ), monômeros - monossa-
carı́deo (Ssu ), aminoácidos (Saa ) e ácidos graxos de cadeia longa (Sfa ), e ácidos graxos
voláteis - butirato (Sbu ), valerato (Sva ), propionato (Spro ) e acetato (Sac ); e duas frações
inertes: solúvel (Si ) e particulada (Xi ). A identificação das concentrações das frações que
compõem o substrato requer técnicas analı́ticas especı́ficas e, em muitos casos, somente
medições da demanda bioquı́mica de oxigênio (DBO), DQO, carbono orgânico total, ni-
trogênio orgânico e amônia estão disponı́veis. A necessidade de caracterização detalhada

do substrato é uma das principais limitações apontada por diversos pesquisadores (Klee-
rebezem e van Loosdrecht, 2006; Batstone et al., 2009; Girault et al., 2012). Para superar
tais limitações vários métodos têm sido desenvolvidos (Copp et al., 2003; Vanrolleghem
et al., 2005; Kleerebezem e van Loosdrecht, 2006; Zaher et al., 2007; Lubken et al., 2007;
Yasui et al., 2008; Wichern et al., 2009; Batstone et al., 2009; Nopens et al., 2009; Zaher
et al., 2009; Girault et al., 2012). Alguns desses métodos são aqui descritos de maneira
resumida:
• Análise fı́sico-quı́mica (Lubken et al., 2007; Wichern et al., 2009)
Nesse método, o fracionamento bioquı́mico (análise de proteı́nas, carboidratos e lipı́deos)

e a análise de ácidos graxos voláteis e extração de fibra são usados para dividir a DQO total
em cada variável de entrada do modelo. Esse tipo de método utiliza conceitos bastante
simples, mas a conversão das frações analı́ticas em unidades de DQO é complexa. Além
disso, testes adicionais e ajuste do modelo aos dados experimentais são necessários para
estimar frações inertes e a cinética da hidrólise.
• Análise elementar (Kleerebezem e van Loosdrecht, 2006).
Esse é um método para caracterização e fracionamento a partir de um número limi-

tado de medidas disponı́veis (DQO, carbono orgânico total (COT), Nitrogênio orgânico e
Alcalinidade). Nesse método, a composição elementar dos substratos (C, H, O, N e P) é
usada para obter o fracionamento bioquı́mico do substrato de acordo com as exigências
do ADM1. Além da caracterização do afluente, o método permitiu a estimativa do pH a
partir da degradação do substrato pelas Archaeas metanogênicas.
• Análise fı́sico-quı́mica combinada com a calibração on line (Batstone et al., 2009).
Esse método pode ser aplicado para diversos tipos de substratos e de reatores opera-
dos em escala real. No entanto, a determinação conjunta da caracterização do afluente e
parâmetros cinéticos pode ser complexa, devido às questões de identificabilidade, aumen-
tando as incertezas associada aos resultados (Batstone et al., 2009).
• Conversão da saı́da do ASM nas variáveis de entrada do modelo ADM1 (Copp et al.,
2003; Vanrolleghem et al., 2005; Zaher et al., 2007; Nopens et al., 2009)
Seção 2.7. Considerações finais 67
Esse método é recomendado para modelagem e simulação de uma estação de tratamento

de águas residuárias. A caracterização e fracionamento do substrato afluente podem ser
obtidos a partir dos valores de saı́da do Activated Sludge Model (ASM) Henze et al. (2000),
atualmente, considerado um modelo padrão do processo de lodo ativado. A interface
do modelo foi desenvolvida para converter as variáveis de saı́da dos modelos ASM nas
entradas do ADM1. Essa abordagem foi utilizada no trabalho de Rousseau et al. (2008).
No entanto, dependendo das circunstâncias, uma discrepância entre biodegradabilidade
aeróbia e anaeróbia pode levar a inconsistências nos resultados da modelagem (Buendı́a
et al., 2008).
• Testes de biodegradabilidade anaeróbia (Yasui et al., 2008; Zaher et al., 2009; Girault
et al., 2012)
Os testes de biodegradabilidade anaeróbia consistem em identificar as frações de DQO

e parâmetros cinéticos associados à sua degradação por meio da interpretação da curva da
taxa de produção de metano. Essa curva é obtida durante a degradação anaeróbia do subs-
trato, geralmente em experimentos conduzidos em batelada. O método foi desenvolvido
com base no procedimento de respirometria aeróbia, no qual o fracionamento da matéria
orgânica afluente é obtido para aplicação nos modelos ASM.
Os testes permitem obter simultaneamente as variáveis de estado de entrada (fraciona-
mento da matéria orgânica) e parâmetros cinéticos da hidrólise, por exemplo. Esse método
foi aplicado por Yasui et al. (2008); Zaher et al. (2009) para investigar a degradação de
substratos especı́ficos (lodo primário). Com base nesse método, Girault et al. (2012) de-
senvolveram um procedimento combinando teste de degradação (respirometria anaeróbia)
e interpretação numérica dos resultados obtidos (otimização das variáveis de entrada do
ADM1).
Quanto a capacidade para identificação das frações de DQO e parâmetros cinéticos
compatı́veis com o ADM1 os testes de biodegradabilidade anaeróbia representam o melhor
método para pesquisa, projeto, operação e otimização de processos anaeróbios e é aplicável
a uma ampla gama de substratos (Girault et al., 2012).
2.7 Considerações finais
No presente capı́tulo foi apresentada e discutida as bases conceituais dos processos

biológicos de digestão anaeróbia, processo convencional de remoção de nitrogênio e uma
breve revisão da literatura sobre a RSCN em sistema integrado anaeróbio-anóxico e sobre
os modelos matemáticos elaborados para estudar a digestão anaeróbia até a proposição

do ADM1, incluindo: descrição do ADM1, suas extensões, metodologias e estratégias sua
aplicação.
Apesar de ter sido aplicado com sucesso no tratamento de águas residuárias, o processo
anaeróbio não remove completamente a matéria orgânica, havendo também a necessidade
da remoção de nutrientes e patógenos. O sistema A2 O é o método convencional aplicado
para remoção de matéria orgânica e nutriente. No entanto, o sistema A2 O possui algumas
desvantagens, como a necessidade de uma fonte de carbono externo para realização do
processo de desnitrificação, aumentando o custo da operação. Um sistema alternativo
para remover simultaneamente a matéria orgânica e nitrogênio em um único estágio, tem
sido aplicado no tratamento de águas residuárias concentradas e diluı́das com eficiência na
remoção de matéria orgânica e nitrogênio variando de 80 - 99%.
Modelos matemáticos podem ser utilizados como uma ferramenta para previsão, con-
trole e otimização desses processos simultâneo. Nesse sentido, o ADM1 foi proposto como
um modelo generalizado para diversos tipos de processos, considerando reações bioquı́micas
e fı́sico-quı́micas decorrentes da digestão anaeróbia de compostos orgânicos complexos até
produtos finais (CH4 e CO2 ). No entanto, esse modelo possui algumas limitações quanto
sua aplicação. Nesse contexto, metodologias e estratégias têm sido desenvolvidas para
aumentar a capacidade de predição do ADM1.
Buscando elucidar aspectos mais complexos da digestão anaeróbia, o ADM1 é con-
siderado um modelo simples e de alcance limitado. Isso pode ser superado pela adição
de extensões, que abordam os requisitos para aplicação. Por outro lado, a aplicação e
utilização prática do ADM1 é considerada complexo. Nesse sentido, métodos para simpli-
ficação e caracterização detalhada do afluente do ADM1 têm sido reportados na literatura,
a saber, métodos de projeção e não projeção, análise fı́sico-quı́mica, análise elementar,
análise fı́sico-quı́mica com a calibração on line, conversão de saı́da do ASM nas variáveis
de estado do ADM1 e testes de biodegradabilidade anaeróbia.
Apesar do avanço na proposição de métodos para melhorar a capacidade de predição do
ADM1, ainda existe à necessidade de desenvolvimento de modelos matemáticos para avaliar
a remoção simultânea de carbono e nitrogênio em reatores anaeróbio-anóxico tratando
águas residuárias diluı́das e concentradas. Além disso, avaliação do efeito de variáveis
importantes para o processo como o pH, TDH e razão DQO/NO−
3 na RSCN é incipiente
e pouco explorada.
Capı́tulo 3
Application of Anaerobic Digestion Model No. 1 for
simulating anaerobic mesophilic sludge digestion
Artigo publicado na revista

Waste Management (0956-053X),
Fator de impacto (JCR 2014): 3.522.
Abstract
Improving anaerobic digestion of sewage sludge by monitoring common indicators such as

volatile fatty acids (VFAs), gas composition and pH is a suitable solution for better sludge
management. Modeling is an important tool to assess and to predict process performance.
The present chapter focuses on the application of the Anaerobic Digestion Model No.1
(ADM1) to simulate the dynamic behaviour of a reactor fed with sewage sludge under
mesophilic conditions. Parametric sensitivity analysis is used to select the most sensitive
ADM1 parameters for estimation using a numerical procedure while other parameters are
applied without any modification to the original values presented in the ADM1 report.
The results indicate that the ADM1 model after parameter estimation was able to predict
the experimental results of effluent acetate, propionate, composites and biogas flows and
pH with reasonable accuracy. The simulation of the effect of organic shock loading clearly
showed that an organic shock loading rate above of 35 kg/m3 day affects the performance of
the reactor. The results demonstrate that simulations can be helpful to support decisions
on predicting the anaerobic digestion process of sewage sludge.
keywords
Anaerobic Digestion Model No. 1; Simulation; Sewage sludge; Hydraulic retention time
70 Capı́tulo 3. Application of Anaerobic Digestion Model No. 1 for simulating anaerobic mesophilic sludge digestion
3.1 Introduction
Sludge is generated as a residue from the treatment of municipal wastewater. It contains

mainly water, organic matter, nutrients, living organisms and some non-biodegradable
particles (Wawrzynczyk, 2007). Anaerobic digestion has been proven to be one of the most
efficient process technologies for stabilizing sludge and furthermore it produces methane as
a form of fuel which may reduce treatment cost (Wang et al., 2013). A mesophilic regime,
25 to 40 ◦ C, can be more attractive than a thermophilic temperature regime, mainly
because of the lower energy requirements. Nevertheless, the heat required by thermophilic
processes can be balanced with faster kinetics and increased methane generation, resulting
in a similar (or higher) net energy outcome.
Anaerobic digestion of sludge may be highly affected by the specific characteristics
of the feed substrate, operational parameters such as hydraulic retention time (HRT)
and organic loading rate, and environmental factors such as pH, presence of inhibitory
or toxic substances, temperature, reactor configuration etc. Volatile fatty acids (VFAs),
gas composition and pH are useful indicators for monitoring gradual changes in sludge
treatment processes (Wang et al., 1997).
Modeling is an important tool for assessing aspects of bioprocesses, process design,
determination of optimal operating conditions and controlling the anaerobic digestion pro-
cesses. They improve the understanding of the system, the formulation and validation of
hypothesis and are able to predict the system performance (Donoso-Bravo et al., 2011).
Fundamental models describing anaerobic digestion process have been developed in recent
decades (Pavlostathis e Gossett, 1986; Siegrist et al., 1993; Kalyuzhnyi, 1997a; Angelidaki
et al., 1999; Siegrist et al., 2002; Dalmau et al., 2010; Gavala et al., 2003; Mu et al., 2008;
Tartakovsky et al., 2008; Lee et al., 2009; Mairet et al., 2011; Muha et al., 2012).
A specialist group working on anaerobic digestion for the International Water Associ-
ation have been trying to condense all this knowledge into a model known as Anaerobic
Digestion Model No.1 (ADM1) which has been successfully applied to simulate the anae-
robic digestion of sewage sludge (Batstone et al., 2002a; Blumensaat e Keller, 2005; Galı́
et al., 2009; Mairet et al., 2011). The aim of this chapter is to apply the ADM1 model to
simulate the dynamic behaviour of a reactor fed with sludge and simulate the application
of organic shock loads under mesophilic conditions.
Seção 3.2. Materials and methods 71
3.2 Materials and methods
3.2.1 Experimental data
The investigation was carried out using a set of data derived from experiments con-
ducted by Siegrist et al. (2002). Two anaerobic rectors in lab-scale (28 L liquid volume,
5 L headspace) were maintained at 35◦ C and continuously fed with a mixture of primary,
secondary, and tertiary sludge from a municipal treatment plant mixed with acetate and
propionate. The reactor was operated with a hydraulic and solids retention time of 17.5
days. The hydraulic and solids retention time was reduced to 8.7 on day 16 and then to
5.3 on day 37. Table 3.1 shows the composition of the wastewater sludge described in the
study of Siegrist et al. (2002).
Table 3.1 - Composition of the wastewater sludge (Siegrist et al., 2002)

Parameter Values Unit
Total COD 40 kg m−3
Soluble COD 4.5 kg m−3
Acetate 1.5 kg m−3
Propionate 0.7 kg m−3
NH4 -N 0.4 kg m−3
TKN 1.6 kg m−3
pH 6.8 -
3.2.2 Input characterization: COD fractioning
Fractionation of organic matter measured as total COD of the sewage sludge (46.7 kg
m−3 ) was divided into 40% inert particulate COD, 30% particulate degradable COD, and
30% hydrolysis products which were divided into 9% amino acids, 6% sugars, 13.5% LCFA,
1.5% inert soluble COD. The inputs applied to the variables of the implemented model are
described in Table 3.2.
3.2.3 Implementation of ADM1 in Matlab/Simulink
ADM1 has been implemented in many software packages WEST, GPS-X, SIMBA and
Aquasim (Gernaey et al., 2004). In the present research, the set of ordinary differential
equations of ADM1 was coded and implemented using Matlab/Simulinkr and integrated
with the ODE15s solvers to solve stiff ODE systems. ADM1 is a highly complex model
and its implementation requires simultaneously solving 32 differential equations in the
liquid phase and 3 differential equations in the gas phase. There are 24 equations based
Table 3.2 - Model inputs used for simulations

Variable Component Value Unit
Ssu Monosaccharides 2.8 kg COD m−3
Saa Aminoacids 4.2 kg COD m−3
Sfa L.C. fatty acids 6.3 kg COD m−3
Sic Inorganic carbon 0.005 kmole C m−3
Sin Inorganic nitrogen 0.003571 kmole N m−3
Si Soluble inerts 0.7 kg COD m−3
Scat Cations 0.04 kmole m−3
San Anions 0.003571 kmole m−3
Xc Composites 10 kg COD m−3
Xi Particulate inerts 18 kg COD m−3
on biochemical processes in the liquid phase, two for cation and anion states and 6 for
acid-base pairs.
The Matlab/Simulink was chosen because of its flexibility for further structural modifi-
cations (ADM1 implement extensions) to enable integration of the model with other blocks
containing other elements (e.g., controllers) and to monitor the evolution of each variable in
the course of a simulation in real time throughout a output block (display). Furthermore,
when a simulation is complete, its output matrix can be exported to a spreadsheet.
3.2.4 Sensitivity analysis and parameter estimation
A parametric sensitivity analysis was performed to define the most sensitive ADM1
parameters during the production and consumption of acetate, propionate and biogas flow.
The parameters evaluated were hydrolysis constant rates for disintegration of composites,
carbohydrates, proteins and lipids (kdis , khyd,ch , khyd,pr and khyd,li , respectively), Monod
maximum specific substrate uptake rate (km ), half-saturation constant (KS ) and yield of
biomass production (Y ). The km , KS and Y values underwent a sensitivity analysis for
each of the seven different microbial communities. Although Jeong et al. (2005) mentioned
that yield of product on substrate (f ) values showed high sensitivities to components of
the ADM1 analyzed, in this study these parameters were negligible due to their small
variations. Lee et al. (2009) also reported that with regard to sensitivity, f is negligible.
According to Jeong et al. (2005) the methane concentration was set as the focused variable
to measure sensitivity methane production and the sensitivity index (SI) is defined as
follows:
P
|CST D (t) − CSEN S (t)|
SI = , (3.1)
N
Seção 3.2. Materials and methods 73
where CST D (t) and CSEN S (t) are the simulation results with the suggested parameter values
and the parameters with a relative change of target parameter (m3 .d−1 ), respectively, for
each given time (t), and N is the number of data. The ranges tested were defined as those
suggested by Batstone et al. (2002a), varying within 30, 100 and 300% depending on the
parameter. Whereas the original parameter has a value of 100%, the generated values
range between: 70 and 130% in condition 1 (varying within 30%), 0 and 200% in condition
2 (varying within 100%), and -200 to 400% in condition 3 (varying within 300%). Since
the values of 0 and -200% in conditions 2 and 3 are not physically acceptable, we used a
minimum value in these cases, 10%.
The ADM1 parameters selected by the sensitivity analysis were estimated by minimi-
zing the difference between the measurements and model outputs. Measurable process
variables included flow biogas, pH, propionate, and acetate concentrations. Parameters
were optimized only with measured experiment data of the methane concentration. The
following objective function (Fobj ) was used for parameter estimation:
n
X
Fobj (Θ) = min (yexp (t) − ysim (t, Θ))2 , (3.2)
t=1
yexp are the collected measurements, ysim are the model predicted outputs, Θ represents
the parameters to be determined and n is the number of measurements. Minimizing the
objective function is an important issue for prediction purposes or process stability (Bats-
tone et al., 2003). Optimization was made with the fminsearch function in the Matlabr
toolset.
Figure 3.1 describes the method used to fit the ADM1 model to the set of experimental
results. In the first step, the sensitivity analysis was performed using the set of initial values
of all the model parameters. Then simulations were undertaken to fit the model outputs
to the experimental results data by changing ADM1 parameters selected by the sensitivity
analysis until the best parameters values are found. The other parameters showed low
sensitivity in the model outputs and therefore they were used without any modification.
3.2.5 Effect of organic shock loads on the stability of the anaerobic reactor
Model simulations were used to investigate the effect of shock loads on the composition
of biogas, pH and VFA production under continuous conditions over 100 days. The liquid
and gas volume of the reactor were the same as those used in the parameter estimation. The
composition of the simulated influent was same as that used in the parameter estimation,
see Table 3.2 with a flow rate of 1.5 m3 day−1 . The temperature and HRT of the system
Initial values of
Sensitivity analysis
ADM1 parameters
Initial estimation
of sensitive parameter set
Imput data
ADM1 (Flow,
concentration, etc.)
Integration of ODE 15s

ADM1 equations solver
Experimental output
Comparaison with lab-
data (output
scale experiments
concentration)
yes
Optimised
Optimal fit
parameters set
no
Adjusment of
parameters set
Figure 3.1: Estimation procedure to identify most sensitive parameters.
were assumed to be 35◦ C and 18 days, respectively.
3.3 Results and discussion
3.3.1 Sensitivity analysis and parameter estimation
The results of the sensitive analysis are shown in Figure 3.2. The sensitivities of the
hydrolysis constants, kdis , Y , and KS displayed significant range variation and asymmetric
characteristics, as shown in Figure 3.2a, c and d.
Higher sensitivity values were also observed for km,ac , km,h2 , km,su and km,aa at ranges
below (Figure 3.2b). Low values of km,ac caused accumulation of acetic acid, lowering the
pH and inhibiting methanogenesis, which resulted in greater sensitivity. Higher hydrogen
concentrations (km,h2 ) are quite important as they may inhibit LCFA and organic acid
degradation. The same is true for km,f a which showed moderate sensitivity. KS and Y
presented very low sensitivities (Figure 3.2a, c and d) with values below 1.5 m3 d−1 . Jeong
et al. (2005) also reported that, with regard to sensitivity, Y and KS are not so important,
except for KS,ac sensitivity to acetate and methane concentrations. Based on the sensitivity
Seção 3.3. Results and discussion 75
( a ) 1 .2 (b ) 3 0
k k m ,h 2
d is
k k m ,s u
1 .0 h y d ,c h 2 5
k k m ,a a
h y d ,p r
k k m ,p ro
0 .8 h y d ,li 2 0
] k m ,c 4
]
-1
-1
m ,a c
.d
0 .6 1 5
.d
3
3
m ,fa
S I [m
S I [m
0 .4 1 0
0 .2 5
0 .0 0
6 0 8 0 1 0 0 1 2 0 1 4 0 0 1 0 0 2 0 0 3 0 0 4 0 0
R a n g e [% ] R a n g e [% ]
( c ) 1 .2 ( d ) 1 .2
K S ,c 4 Y h 2
K S ,s u Y
1 .0 1 .0 p ro
K S ,a a Y c 4
K S ,h 2 Y fa
0 .8 K 0 .8
S ,p ro Y s u
]
-1
]
K S ,a c -1 Y
.d
a a
0 .6 K .d 0 .6
3
Y
3
S ,fa a c
S I [m
S I [m
0 .4 0 .4
0 .2 0 .2
0 .0 0 .0
0 1 0 0 2 0 0 3 0 0 4 0 0 6 0 8 0 1 0 0 1 2 0 1 4 0
R a n g e [% ] R a n g e [% ]
Figure 3.2: Sensitivity indices (SI) calculated for different ranges of (a) hydrolysis constant rates; (b) Mo-
nod maximum specific uptake rates; (c) half-saturation constants; and (d) yields of biomass on substrate.
results, we decided to use km,ac , km,h2 , km,su and km,aa for parameter estimation.
Table 3.3 shows the initial values of the kinetic parameters and their respective optimi-
zed values. Maximum specific substrate uptake rate of acetate (km,ac ), hydrogen (km,h2 ),
monosaccharides (km,su ) and amino acids (km,aa ), were considered as having the greatest
impact on experimental data and they were estimated in comparison with the experimental
data.
Table 3.3 - Initial and estimated values of kinetic parameters

Kinetic parameters Initial value Estimated value Units
km,ac 8.00 13.80 day−1
km,h2 35.00 26.01 day−1
km,su 30.00 20.22 day−1
km,aa 50.00 41.12 day−1
Figure 3.3 shows a comparison of simulation and experimental results for anaerobic di-
gestion of acetate and propionate. We can conclude that the concentrations of acetate and
propionate were well predicted by the ADM1 model. Moreover, the simulation results with
optimized parameters showed good agreement with the experimental data. Furthermore,
acetate and propionate failure at a HRT of 5.3 days was well justified by the ADM1 model
but it was predicted with some deviations between the simulation results without optimi-
zation and measurements. These discrepancies between the simulations and measurements
observed may be due to the correlation of some sensitive parameters such as hydrolysis
constants with feed concentration and HRT as mentioned by Gavala et al. (2003) and to
some potential weaknesses in the ADM1 model as addressed by many researchers such as
Parker (2005); Kleerebezem e van Loosdrecht (2006); Batstone (2006c).
(a ) 1 .0
H R T = 1 7 .5 d 8 .7 d 5 .3 d 4 d
0 .8
]
3
A c e ta te [k g C O D .m
0 .6
0 .4
0 .2
0 .0
(b ) 1 .0
W ith o p tim iz a tio n
W ith o u t o p tim iz a tio n
0 .8 E x p e r im e n ta l
]
3
P r o p io n a te [ k g C O D .m
0 .6
0 .4
0 .2
0 .0
0 1 0 2 0 3 0 4 0 5 0 6 0 7 0
T im e [d a y s ]
Figure 3.3: Comparison between simulations and experimental data with and without optimization: (a)
acetate and (b) propionate.
The model calibration results for effluent biogas flow and composition, pH, using the
optimized parameter set are presented in Figure 3.4. Effluent pH value, biogas flow and
methane, carbon-dioxide and hydrogen composition were predicted with high accuracy for
all HRTs.
Seção 3.3. Results and discussion 77
(a ) 6 0
H R T = 1 7 .5 d 8 .7 d 5 .3 d 4 d
5 0
S im u la te d
E x p e r im e n ta l
4 0
]
-1
.d
3 3 0
B io g a s [m
2 0
1 0
0
(b ) 8 0
6 0 C H S im u la te d
4
C O 2
S im u la te d
H 2
S im u la te d
C H E x p e r im e n ta l
[% ]
4
4 0
C O 2
H 2
2 0
(c ) 0
8 .0 S im u la te d
7 .5
E fflu e n t p H
7 .0
6 .5
0 1 0 2 0 3 0 4 0 5 0 6 0 7 0
T im e [d a y s ]
Figure 3.4: Comparison between simulations and experimental data after parameter estimation: (a)
biogas production, (b) biogas composition and (c) pH
As can be seen in Figures 3.3 and 3.4, from day 11 there was an increase in the
concentrations of acetate, propionate and biogas flow. This was due to a reduction in the
HRT. In this case, the reduction in HRT did not cause process inhibition and the methane
concentration in the biogas was not significantly affected. The pH value remains within the
range (6.5 - 8.5) recommended in the literature for anaerobic bacteria (the methanogens)
(Kalyuzhnyi, 1997a).
The performance of the microbial community growing in an anaerobic digestion system
is determined primarily by biological solids retention time (SRT). For a CSTR system
without recycling, SRT equals HRT, as is the case in this study. The reactions of the
anaerobic digestion process are directly related to the SRT. There is a minimum SRT for
each reaction. If the SRT is less than the minimum SRT, the microorganisms can not
grow rapidly enough and the digestion process will fail eventually. Metcalf e Eddy (2007)
reported the typical minimum value and design of SRT as 4 and 10 days, respectively,
when the reactor is operated in mesophilic conditions. After the simulation, we therefore
concluded that the optimum HRT for such a system is in the range of 6 - 9 days.
3.3.2 Effect of organic shock loads on the stability of the anaerobic reactor
The results of model simulation for a organic shock loading are shown in Figure 3.5.
The organic loading rate increased from 5 to 35 kg/m3 day on day 30, then to 70 kg/m3
day on day 60. The organic loading rate was chosen to be above the value range minimum
suggested in the literature (2 - 32 kg/m3 day) (Metcalf e Eddy, 2007).
Model simulations revealed that the accumulation of carbon dioxide results in an incre-
ase in the dissolved carbonic acid concentration, thereby increasing the production of H+ .
The accumulation of hydrogen results in a greater production of propionate and butyrate
whose degradation is inhibited by the presence of a high concentration of acetate (Figure
3.5c). Although ADM1 reported a low value of KS for hydrogenotrophic methanogenic
Archaea compared to other organisms of the microbial consortium, the results in Figure
3.5b show that these microorganisms reduce carbon dioxide below the maximum working
capacity. According to Giraldo-Gomez et al. (1992) this occurs because the concentration
of dissolved hydrogen is available and generally low due to mass transfer limitation.
Inhibition of the syntrophic bacteria by the accumulation of the products formed by
fermentative bacteria will lead to a continued accumulation of VFA which contributes to
the consumption of alkalinity and a pH decrease (Figure 3.5a). The pH decrease affects
the growth of methanogenic microorganisms and consequently reduces the composition of
Seção 3.4. Conclusions 79
(a ) 8
p H
5
3
(b ) 1 0 0
C H 4
8 0 C O 2
H 2
6 0
[% ]
4 0
2 0
(c ) 0
1 2 0
B u ty ra te
]
P r o p io n a te
-3
V F A s [k g C O D .m
9 0 A c e ta te
6 0
3 0
0
0 2 0 4 0 6 0 8 0 1 0 0
T im e [d a y s ]
Figure 3.5: Model simulations for a continuous flow system with shock loading of 5 kg/m3 day experiencing
a step increase in organic loading rate to 35 kg/m3 day on day 30 and to 70 kg/m3 day on day 60. (a):
pH and (b) Biogas composite (c) VFAs (18 days HRT).
methane in the biogas, as shown in Figure 3.5b.
3.4 Conclusions
The ADM1 model is a powerful tool for predicting the dynamic behaviour of anaerobic
digesters treating sewage sludge under mesophilic conditions. This chapter evaluates the
application of the Anaerobic Digestion Model No. 1 (ADM1) developed by the IWA task
group for mathematical modelling of anaerobic processes.
A sensitivity analysis showed that among all the kinetic parameters, the Monod maxi-
mum specific substrate uptake rate affects biogas production significantly, which was re-
lated to all the processes used in the model. Important parameters including km,ac , km,h2 ,
km,su and km,aa were estimated using iterative methods, which optimized the parameters
with experimental results. With regard to the model output for acetate and propionate
concentrations, the simulation results showed good agreement with the experimental data
after parameter optimization. Despite in practice for CSTR sludge digestion, values for
the SRTs range from 10 to 20 days are reasonable considering that the optimum HRT for
such a system is in the range of 6-9 days where there was a greater production of methane
in the reactor.
The simulation results obtained to evaluate the effect of organic shock loads on the
anaerobic reactor using the ADM1 clearly show that an organic loading rate above 35
kg/m3 day causes process instability and even failure when fermentation processes exceed
methanogenesis, leading to the accumulation of VFAs, falling pH, and a reduction in
methane production.
Capı́tulo 4
Modeling simultaneous carbon and nitrogen removal in
an anaerobic-anoxic reactor treating domestic
wastewater
Artigo submetido (em revisão) para publicação na revista
Journal of Environmental Management (0301-4797),
Abstract
This chapter presents a mathematical model based on the Anaerobic Digestion Model No.
1 (ADM1) to simulate the effects of nitrate concentration and hydraulic retention time
(HRT) on the simultaneous carbon and nitrogen removal (SCNR) in anaerobic-anoxic re-
actor treating domestic wastewater. The model was calibrated using previously published
experimental data obtained from anaerobic batch tests for different COD/NO−
3 ratios. Mo-
del simulations were performed to predict the SCNR in a continuously stirred tank reactor
(CSTR) operating under mesophilic conditions (35 ◦ C). Six different scenarios were evalu-
ated to investigate the performance of the SCNR based on typical influent characteristics
of domestic wastewater. The variables analyzed were chemical oxygen demand (COD) re-
moval, nitrate concentration, methane production, nitrogen gas, volatile fatty acids (VFA)
concentration, pH and percentage of COD used by the denitrifying and methanogenic mi-
croorganisms. The HRT was decreased stepwise from 15 to 4 hours. The results indicate
that Scenario (S5) with a COD/NO−
3 ratio equal to 10 and an HRT equal to 15 hours
ensures the occurrence of the stable SCNR. Furthermore, the accumulation of denitrifica-
tion intermediates and a significant reduction in the biogas production when the organic
matter is limited was verified.
keywords
Simulation; Denitrification; Anaerobic Digestion Model No. 1; Domestic wastewater
Capı́tulo 4. Modeling simultaneous carbon and nitrogen removal in an anaerobic-anoxic reactor treating domestic
82 wastewater
4.1 Introduction
Anaerobic processes have been successfully applied to highly concentrated soluble was-
tewater treatments. This process is also feasible for the treatment of low-strength was-
tewaters, such as domestic wastewater, particularly under tropical conditions in which the
supply of heat is not required. Experience has been gained from full scale application of
anaerobic systems in Brazil, Indonesia, India and Colombia (Khan et al., 2015).
Compared to conventional aerobic treatment systems, the anaerobic treatment process
can serve as a viable and cost-effective alternative due to its relatively low operational cost,
operational simplicity, low production of excess sludge, production of methane gas, with
can be collected and used as an energy source, and low energy consumption (van Haandel
e Lettinga, 1994). However, the residual concentration of certain pollutants, primarily
organic matter and nutrients, in its effluent typically exceeds the maximum permissible
level prescribed by the effluent discharge standards of most countries. Little or no nutrient
removal may be expected in an anaerobic system treating domestic wastewater (Foresti
et al., 2006). The reason for low nutrient removal is that during the anaerobic process,
organic nitrogen and phosphorous are hydrolyzed to ammonium nitrogen and phosphate,
which are not removed from the system; consequently, the nutrient concentration increases
in the liquid phase. Thus, post treatment of the anaerobic effluent is necessary to further
polish the remaining organic pollutants and nutrient removal (Moawad et al., 2009).
The conventional biological treatment method used to remove the organic matter and
nutrients consists of a three-stage process, anaerobic-anoxic-aerobic (A2 O). For the nitro-
gen removal, the treatment is accomplished in two separate stages, i.e., the oxidation of
ammonia to nitrite, and then to nitrate in the first aerobic nitrification stage followed by
the second anoxic denitrification stage related to the sequential reductive reactions from
nitrate (NO− −
3 ) to nitrite (NO2 ), nitric oxide (NO), nitrous oxide (N2 O) and finally to ni-
trogen gas (N2 ). The practical application of this technological arrangement often poses
extra costs due to the need of an external carbon source because the amount of organic
matter in the effluent of the anaerobic systems is normally lower than that needed for
denitrification and its organic fraction is not readily biodegradable.
Therefore, a possible alternative is the integrate anaerobic digestion and the denitrifi-
cation (AD/DN) processes in the first stage to create an anaerobic pre-treatment system
for simultaneous carbon and nitrogen removal (SCNR) purposes. The first stage is then
followed by an second stage (aerobic) for nitrification, and a subsequent partial recircula-
tion of the nitrified aerobic effluent remains in the anaerobic system (van Lier, 2008). This
Seção 4.1. Introduction 83
configuration could lead to an SCNR with a reduction in the treatment costs, removal
of toxic compounds, lower energy consumption, and reduced sludge production (Bernet
et al., 2000; Del Pozo e Diez, 2005; Chuan et al., 2009). Nevertheless, previous studies
indicate that the presence of nitrite and nitrous oxides, generated as intermediates, have a
strong impact and a deleterious effect on the fermentative/methanogenic consortia (Kluber
e Conrad, 1998b,a; Tugtas e Pavlostathis, 2007b,a). These oxides are generated as inter-
mediate compounds during the denitrification pathway of the anaerobic organisms with
denitrifying (facultative anaerobes) capabilities existent in the anaerobic sludge, rather
than the presence of the nitrate itself. Nevertheless, the effects this production appeared
to be reversible, and the methanogenic activity was recovered after the nitrogenous com-
pounds were fully reduced to dinitrogen gas by the denitrifying microorganisms (Kluber e
Conrad, 1998b; Tugtas et al., 2006).
Several studies have reported the use of integrating AD/DN processes in the anaerobic
pre-treatment system to the SCNR treating high-strength wastewaters (Bernet et al., 2001;
Del Pozo e Diez, 2005; Baloch et al., 2006; Huang et al., 2007; Chuan et al., 2009; Ghaniyari-
Benis et al., 2010; Huiliñir et al., 2011) and only a few studies have focused on low-strength
wastewaters (Tai et al., 2006; Yingyu et al., 2008, 2009; Banihani et al., 2009; Kassab,
2009). Within this context, research on the integrate AD/DN processes in the anaerobic
pre-treatment system treating domestic wastewater to the SCNR is justified to consolidate
the application of the two-stage process (sequential anaerobic-aerobic) primarily in warm-
climate countries, with several treatment systems operating in full scale.
Models describing the anaerobic degradation of organic matter have been proposed
in the literature, such as the ones reported by Angelidaki et al. (1999), Siegrist et al.
(2002), Tartakovsky et al. (2008) and Muha et al. (2012). They present different levels
of complexity and are based on various assumptions and simplifications. The Anaerobic
Digestion Model No. 1 (ADM1), which has been developed by the IWA Task Group for
Mathematical Modeling of Anaerobic Digestion Processes, is one of the most sophisticated
and complex anaerobic digestion models, involving nineteen biochemical processes and two
types of physicochemical processes. However, the denitrification process is not currently
a part of the model, even though a few suggestions have been made by Batstone et al.
(2002a) to take them into consideration.
Tugtas et al. (2006) developed an extension including four stages of nitrate reduction to
nitrogen gas for the ADM1 model. The model was calibrated against data from a series of
experiments conducted in a batch reactor operating with mixed cultures of microorganisms
(fermentative and methanogenic). The simulation results from the extended ADM1 had a
84 wastewater
good agreement with the measurements. Other models describing the integrate AD/DN
processes also been proposed in the literature (Chaudhry e Beg, 1997; Garibay-Orijel et al.,
2006; Huiliñir et al., 2008, 2009, 2011; Huiliñir e Montalvo, 2013). Although there are
several studies on the performance and kinetic evaluation of the AD/DN processes, there
is a lack of applications in the SCNR for the anaerobic pre-treatment system treating
domestic wastewater.
Therefore, this chapter aims to adapt an extension of the ADM1 which considers the
denitrification process to evaluate the SCNR in an anaerobic pre-treatment system treat-
ment domestic wastewater. To achieve this objective, the effect of nitrate concentration
on the SCNR in a anaerobic-anoxic reactor was simulated for various domestic wastewater
composition. Furthermore, the influence of the hydraulic retention time (HRT) on the
SCNR was investigated.
4.2 Methodology
4.2.1 ADM1 extension modifications and implementation
The mathematical model proposed here presents ADM1 extension with denitrification
to describe the SCNR in the anaerobic pre-treatment system treating domestic wastewater.
The proposed model only considered a two-step denitrification process (NO− −
3 → NO2 →
N2 ). This assumption is consistent with several nitrification and denitrification models,

including those proposed by Viraj e Bruce (2000); Henze et al. (2000); Huiliñir et al.
(2009). The denitrifier biomass (facultative anaerobes) was introduced as a fraction of the
fermentative consortia, and it was capable of fermenting valerate, butyrate and propionate
in the absence of N-oxides (nitrate and nitrite) (Tugtas et al., 2006). In the presence of the
N-oxides, it was assumed that butyrate, valerate, propionate, and acetate were the electron
donors used by the denitrifiers (Kluber e Conrad, 1998a). In accordance with Aquino e
Stuckey (2008), the hydrogen (H2 ) consumption was omitted from the competition of
the denitrifiers with the hydrogenotrophic methanogens due to their slow consumption.
Furthermore, the consumption of sugars, amino acids, and long-chain fatty acids (LCFAs)
by the denitrifiers was neglected because of their slower consumption kinetics (Tugtas
et al., 2006).
A dissimilatory nitrate reduction to ammonia (DNRA) was not included in the model
because low COD/NO− −
3 ratios (< 50 g COD/g NO3 ) were used in the simulation (Akunna
et al., 1992). Additionally, although sulfate is often found in domestic wastewater with
concentrations in the range of 50-100 mg.L−1 , the biological sulfate reduction process was
Seção 4.2. Methodology 85
not considered (Batstone, 2006c). The prediction of the pH charge balance proposed by
Batstone et al. (2002a) was modified by adding a concentration of the NO− −
3 and NO2 ions.
The denitrifying and the anaerobic microorganisms were assumed to reach optimal growth
at a nearly neutral pH (Campos e Flotats, 2003; Estuardo et al., 2008).
Lastly, the hydrolysis of the influent substrate and the decay of the biomass were
considered as two separate processes (Tugtas et al., 2006). The hydrolysis rate for the
carbohydrates, proteins and lipids of the influent substrate and the decay of the biomass
were selected based on the classical literature data, as indicated in Table 4.1 (O’Rourke,
1968; Christ et al., 2000; Batstone et al., 2002a). Because the influent substrate is readily
biodegradable and rapidly hydrolysable, only a disintegration step was used to represent
the lysis of the biomass.
Table 4.1 - Disintegration and hydrolysis kinetic constants

Parameter Units Value Reference
Influent substrate
khyd,ch,f eed d−1 0.21 O’Rourke (1968)
khyd,pr,f eed d−1 0.10 O’Rourke (1968)
khyd,li,f eed d−1 0.17 O’Rourke (1968)
Biomass decay
kdis d−1 0.5 Batstone et al. (2002a)
khyd,ch d−1 0.025 Christ et al. (2000)
khyd,pr d−1 0.015 Christ et al. (2000)
khyd,li d−1 0.005 Christ et al. (2000)
Figure 4.1 presents the proposed conceptual model for the simultaneous AD/DN pro-
cesses of the domestic wastewater. It includes the following biochemical processes: (1)
hydrolysis of carbohydrates; (2) hydrolysis of proteins; (3) hydrolysis of lipids; (4) aci-
dogenesis from sugars; (5) acidogenesis from amino acids; (6) acetogenesis from LCFA;
(7) propionate from acetogenesis; (8) butyrate and valerate acetogenesis; (9) acetoclastic
methanogenesis; (10) hydrogenotrophic methanogenesis; (11) nitrate reduction; (12) nitrite
reduction; and (13) death and disintegration of the biomass into carbohydrates, proteins,
lipids, and inert. The biological kinetic rate expressions and the stoichiometric parameters
of the extended ADM1 model related to the AD/DN process are reported in the Petersen
matrix form as indicated in Table 4.2.
The degradation equations of the soluble substrate (valerate, butyrate, propionate and
acetate, i=4-7) in the ADM1, including the denitrification processes with the proposed
modification is given as follows:
Table 4.2 - Denitrification process included in the modified Anaerobic Digestion Model no1
Component → i 4 5 6 7 25 26 27 30 Process rate (ρj,n )
j Process Sva Sbu Spro Sac Sno3 Sno2 Sn2 Xf,nox (kg COD m−3 .d−1 )
↓
Sva Sno3
8-1 Uptake − -1 - (1-Yc4 )/1.14 (1-Yc4 )/1.14 Yc4 kno3,va K Xf,nox k
NO3 no3,va +Sva no3 +Sno3
of vale-
rate
Sva Sno2
8-2 − -1 - (1-Yc4 )/0.57 (1-Yc4 )/0.57 Yc4 kno2,va K Xf,nox k
NO2 no2,va +Sva no2 +Sno2
Sbu Sno3
9-1 Uptake − -1 - (1-Yc4 )/1.14 (1-Yc4 )/1.14 Yc4 kno3,bu K Xf,nox k
NO3 no3,bu +Sva no3 +Sno3
of
wastewater
buty-
rate
Sbu Sno2
9-2 − -1 - (1-Yc4 )/0.57 (1-Yc4 )/0.57 Yc4 kno2,bu K Xf,nox k
NO2 no2,bu +Sva no2 +Sno2
Spro Sno3
10-1 Uptake − -1 - (1-Ypro )/1.14 (1-Ypro )/1.14 Ypro kno3,pro K Xf,nox k
NO3 no3,pro +Spro no3 +Sno3
of pro-
pionate
Spro Sno2
10-2 − -1 - (1-Ypro )/0.57 (1-Ypro )/0.57 Ypro kno2,pro K Xf,nox k
NO2 no2,pro +Spro no2 +Sno2
Sac Sno3
11-1 Uptake − -1 - (1-Yac )/1.14 (1-Yac )/1.14 Yac kno3,ac K Xf,nox k
NO3 no3,ac +Sac no3 +Sno3
of
acetate
Sac Sno2
11-2 − -1 - (1-Yac )/0.57 (1-Yac )/0.57 Yac kno2,ac K Xf,nox k
NO2 no2,ac +Sac no2 +Sno2
20 Decay -1 kdec,Xf,nox Xf,nox
86
Inactive biomass
Domestic wastewater Disintegration
NO3- -N Xch Xpr Xli Inert Xch Xpr Xli Inert

13
Hydrolysis
1 2 3
11
Amino Acids Long Chain
Denitrificatrion
Sugars (Ssu) (Saa) Fatty Acids (Sfa)
4 5 6
Acidogenesis Death
Propionate Valerate (Sva)

NO2- -N (Spro) Butyrate (Sbu)
8
Acetogenesis 7
Acetate (Sac) Hydrogen (Sh2)

12
Methanogenesis
N2 9 10
Methane (Sch4)
Figure 4.1: Conceptual model of the simultaneous AD/DN processes of domestic wastewater.
dSliq,i q X X
= (Sin,i − Sliq,i ) − ρj νi,j − (ρj + ρj,nox ) νi,j − ρ11 ν7,11 Inox − ρj,n (4.1)
dt Vliq j=5−7 j=8−10
In Equation 4.1, the first term on right side is the substrate mass balance for a CSTR.
The second term represents the process equations for the uptake of sugars, amino acids
and LCFA by fermentation, as obtained from the ADM1 matrix. The third term is the
fermentation (acetogenesis) of the soluble compounds i in the absence of the N-oxides
by the denitrifiers. The fourth term represents the methanogesis obtained by the ADM1
matrix multiplied by an N-oxide inhibition factor (Inox ) to account for the inhibitory
effect of the N-oxide on the acetoclastic methanogens. The inhibitory effect was simulated
using non-competitive inhibition functions. The fifth term represents the uptake of the
soluble substrates by the denitrifiers in the presence of nitrate and nitrite. The general
rate equation for the utilization of butyrate, valerate and propionate using facultative
anaerobes in the absence of the N-oxides can be expressed as follows:

Si KS,no3 KS,no2 Si
ρj,nox = km,j × Xf,nox , (4.2)
KS + Si KS,no3 + Sno3 KS,no2 + Sno2 Sva + Sbu + Spro
where km,j is the rate constant for the fermentation of process j (kg substrate COD/kg
biomass-COD.d); and Xf,nox is the facultative anaerobes population (kg biomass-COD/m3 ).
The general rate equation for the use of the N-oxides (nitrate, nitrite and nitrogen gas, i
= 25-27) can be defined as follows:
88 wastewater
dSi,nox q
= (Sin,i,nox − Sliq,i,nox ) + ρj,n , (4.3)
dt Vliq
where Si,nox is the concentration of the N-oxide i (kg N/m3 ); and ρi,nox is the rate of pro-
duction/consumption of the N-compound i as a result of the denitrification (kg N/m3 .d).
The biomass retention in the reactor has been modeled in a simplified manner as suggested
in the ADM1 report by Batstone et al. (2002a), which includes a term of the residence time
of the solids (tres,X ) in the biomass mass balance equation to account for the difference
between the hydraulic retention time (HRT) and the solid retention time (SRT). Equation
4.4 can represent the mass balance used by the ADM1 model to describe the dynamic
behavior of the particulate in the liquid phase as follows:
dXliq,i qXin,i Xliq,i X

= − + ρj νi,j . (4.4)
dt Vliq tres,X+ Vliq j=1−20
q
To predict the pH temporal profile and the associated potential inhibition in the acid-
base equilibrium, the charge balances of the ADM1 model can be modified by adding the
SNO−3 and SNO−2 components (Huiliñir e Montalvo, 2013), in Equation 4.5 as follows:
SAc− SPr− SBu− SVa−

SCat+ + SNH+ + SH+ − SHCO− − − − − − S− −
NO3 − SNO2 − SOH− − SAn− = 0 , (4.5)
4 3 64 112 160 208
where, SAc− , SPr− , SBu− , SVa− , SNH+4 , SHCO−3 , SNO−3 and SNO−2 are the concentrations of the
ionized forms of the buffer components.
The set of governing equations is solved simultaneously to simulate the dynamics and
the biological kinetics in the reactors. The details of the model governing equations, input
parameters, and underlying assumptions are described in Batstone et al. (2002a) and Tug-
tas et al. (2006). The ADM1 extension using denitrification for the domestic wastewater
was coded by the authors using MATLAB/Simulink 2008b following the original ADM1
(Batstone et al., 2002a). The model was implemented as a set of ordinary differential
equations using ODE15s as the numerical solver.
4.2.2 Batch model calibration
4.2.2.1 Experimental setup
The model calibration was performance using a set of data derived from batch expe-
riments conducted by (Kassab, 2009). His experiments consisted of inoculating a reactor
with 10 L of flocculent anaerobic sludge, originating from a continuous pilot upflow ana-
erobic sludge blanket (UASB) reactor. A synthetic wastewater medium of approximately
1500 mg COD.L−1 , which simulate the organic matter strength of the domestic sewage,
was fed to the UASB reactor.
During the batch experiments, the sludge withdrawn from the bottom of the UASB
reactor corresponding to a concentration of 500 mg VSS.L−1 was added to twelve 250 ml
capacity serum bottles. Sodium acetate corresponding to an initial concentration of 1500
mg COD.L−1 was added as the substrate. Different amounts of potassium nitrate were
added to four of these bottles to achieve COD/NO−
3 ratios of 150, and 40 in duplicates.
The remaining two bottles did not receive any potassium nitrate and served as blanks. The
total volume of 250 ml was achieved by adding distilled water. The anaerobic conditions
were established by flushing each bottle with pure helium for five minutes. Incubation
was performed at 32 o C with agitation. The total pressure of the cumulative biogas inside
each bottle was measured using a digital manometer (SPER Scientific). The cumulative
biogas composition was measured using a Philips Pye Unicam pu 4500 gas chromatograph
operated under the previously mentioned conditions.
4.2.2.2 Model calibration procedure
The proposed mathematical model was calibrated to minimize the sum of squared er-
rors between the experimental and the simulated data. The calibration procedure was
based on the results conducted using the data from batch assay conditions consecutively
performed with 1500 mg COD.L−1 with 0, 10, and 37.5 mg NO− −1
3 .L . During the calibra-
tion procedure, the denitrifying and the methanogenic biomass concentration represented
by Equation 4.6 as the objective function (Fobj ) can be used for the parameter estimation
as follows:
n
X
Fobj (Θ) = min (yexp (t) − ysim (t, Θ))2 (4.6)
t=1
where yexp are the collected measurements; ysim are the model predicted outputs, Θ re-
presents the biomass concentration to be determined; and n is the number of measure-
ments. The optimization was performed using the fminsearch optimization function in the
MATLAB
R toolbox, which is based on the Nelder-Mead simplex algorithm. The stoichio-
metric and kinetic parameters were the same as described here and in the original extended
model (Tugtas et al., 2006).
90 wastewater
4.2.3 Model simulation in a continuous reactor
The mathematical model proposed in this study was applied to simulate the AD/DN
process in a completely mixed reactor (CSTR) operating under mesophilic conditions (35
◦
C) with tres,X equal to 40 days. The flow rate and the volume liquid were equal to 4,000
m3 .d−1 and 1,000 m3 , respectively, based an HRT value of 6 hours. The assumed gas
headspace of the volume reactor was equal to 100 m3 .
4.2.3.1 Model inputs and initial conditions
The input fractionation and characterization, i.e., the definition of the composition of
the influent based on the model input variables is a critical step in modeling anaerobic
digestion (Kleerebezem e van Loosdrecht, 2006). Thus, for this reason, this study used a
few simplifications based on the general information of domestic wastewater, as indicated
in Figure 4.2 (Van Haandel e Van Der Lubbev, 2007).
CODT
88 % 12 %
CODb CODi
68 % 20 % 5% 7%
CODbp CODbs CODip CODis
Xch 60 % Ssu 60 % Xi 100 % Si 100 %
X pr 30 % Saa 40 %
X li 10 %
Figure 4.2: Schematic representation of the simplified COD fractionation. Percentages correspond to
COD fractions immediately before each compound.
The influent total COD concentration can be divided into a biodegradable (CODb ) and
a non-biodegradable (CODi ) fraction, both subdivided into soluble and particulate. The
subdivision considers the physical size of the organic material.
The soluble biodegradable COD (CODbs ) was considered to be composed of sugars
and amino acids (Ssu and Saa ) and the other biodegradable soluble inputs were assumed
to be zero (Sva , Sbu , Spro , Sac , Sh2 and Sch4 ). The soluble inert COD (CODis ) was allocated
directly to Si . The particulate biodegradable COD (CODbp ) was divided among the car-
bohydrates, proteins and lipids (Xch , Xpr , Xli ), and the particulate inert (Xi ) was assumed
to be equal to the particulate inert COD (CODip ). Ammonia (Sin ) and bicarbonate (Sic )
were obtained from the data provided in Table 4.3. The anion concentration (San ) was
assumed to be equal to Sin according to Rosen e Jeppsson (2002), and the cation concen-
tration (Scat ) was adjusted in each case based on each pH assumed in the simulation. The
organic matter proportions between the carbohydrates, proteins and lipids were obtained
by the characterization of the domestic wastewater, resulting in 60%, 30% and 10% of the
corresponding particulate biodegradable COD, respectively, as reported by Orozco (2005)
and Metcalf e Eddy (2007).
The initial biomass concentration of the seven microbial populations contained in the
ADM1 (Xsu , Xaa , Xfa , Xc4 , Xpro , Xac and Xh2 ) can be defined as follows:
Yi
Xi = Xtotal , i = 17, ..., 23 , (4.7)
Ynox
where, Yi is the observed yield coefficient of each microbial group; Ynox is the overall
observed yield coefficient; and Xtotal is the total biomass concentration in the reactor
measured as the volatile suspended solids. The Yi and Ynox values were obtained from the
yield coefficients and decay rate (0.02 d−1 ), respectively, as suggested by Batstone et al.
(2002a) and Tugtas et al. (2006). The denitrifiers were assumed to be 40% of the butyrate,
valerate and propionate degraders in the baseline model.
4.2.3.2 Definition of scenarios
From the assumptions made for the input fractionation and characterization, six dif-
ferent scenarios (S1-S6) were evaluated to investigate the effects of nitrate concentration
on the SCNR. The proposed scenarios were based on the typical influent characteristic
of domestic wastewater described by Metcalf e Eddy (2007): i) Scenarios S1, S2 and S3
correspond to weak, medium and strong domestic wastewater, respectively, with a nitrate
concentration of 50 mg.L−1 ; ii) Scenario S4 considers strong domestic wastewater without
the presence of nitrate (control); and iii) Scenarios S5 and S6 correspond to strong do-
mestic wastewater with nitrate concentrations of 100 mg.L−1 and 20 mg.L−1 , respectively.
These scenarios are summarized in Table 4.3.
The model simulations were used to investigate the effects of the change in the HRT on
the SCNR under continuous conditions over 150 days. The composition of the simulated
influent was the same as considered in Scenarios S3, S4, S5 and S6.
92 wastewater
Table 4.3 - Different scenarios to simulate the effects of nitrate concentration on the SCNR in anaerobic-
anoxic reactor for weak domestic wastewater (S1), medium domestic wastewater (S2), and strong domestic
wastewater (S3 to S6)
Parameter/Compounds Symbol Units S1 S2 S3 S4 S5 S6
Total COD CODT kg m−3 0.27 0.50 1.00
Particulate COD CODbp kg m−3 0.18 0.34 0.68
Soluble COD CODbs kg m−3 0.054 0.10 0.20
Volatile suspended solids VSS kg m−3 0.095 0.16 0.315
Total alkalinity Salk kg CaCO3 m−3 0.050 0.100 0.200
Ammonium nitrogen N-NH4 kg N m−3 0.012 0.025 0.045
Carbohydrates Xch kg m−3 0.110 0.204 0.408
Proteins Xpr kg m−3 0.055 0.102 0.204
Lipids Xli kg m−3 0.018 0.034 0.068
Particulate inerts Xi kg m−3 0.014 0.025 0.050
Monosaccharides Ssu kg m−3 0.032 0.060 0.120
Amino acid Saa kg m−3 0.022 0.040 0.080
Soluble inerts Ssi kg m−3 0.019 0.035 0.070
Inorganic carbon Sic kmol C m−3 0.0033 0.0067 0.010
Inorganic nitrogen Sin kmol N m−3 0.0009 0.0018 0.0036
Cations Scat kmol m−3 0.0383 0.0417 0.0450
Anions San kmol m−3 0.001 0.002 0.004
COD/NO− 3 ratio 5.4 10 20 0 10 50
Nitrate concentration Sno3 kg m−3 0.05 0.05 0.05 0 0.1 0.02
4.3 Results and Discussion
4.3.1 Batch reactor
4.3.1.1 Model calibration
The results of the model calibration are illustrated in Figure 4.9a. The experimental
data were obtained through a series of batch assays allowing the inhibitory effects of
nitrate on the methane production to be assessed; strict methanogenic sludge was used
as the inoculum at an initial nitrate concentration of 0 (control), 10 and 37.5 mg N/L.
Generally, the simulation results showed a good agreement with the experimental data.
Figure 4.9a indicates that the early addition of nitrate resulted in the immediate sup-
pression of the methanogenesis in the 10 and 37.5 mg N/L solutions. The recovery of
methanogenesis was concomitant to the complete reduction in the nitrate and the nitrite.
The observed inhibition of the methanogenesis is attributed to the increased accumula-
tion of nitrite. The inhibition of the methanogens by the nitrate and nitrite as well as
the substrate competition between nitrate reducers and methanogens has been identified
as the primary mechanisms involved in the suppression of the methanogenesis (Akunna
et al., 1994). The continuous increase in the methane after the sixth day was observed even
with zero concentrations of acetate. The kinetic constants values suggested in the ADM1
Seção 4.3. Results and Discussion 93
and the extension were estimated without isolating each stage of the anaerobic digestion
process. In this case, several syntrophic relationships and a competition in the process
are disregarded; consequently, the acetate derived from the decomposition of the biomass
is not considered in the model. However, the input acetate is quickly consumed, and the
continuous production of methane in the reactor is the result of the use of this substrate
by the methanogenic Archaea acetoclastic and the hydrogenotrophic. A slight increase in
the acetate concentration occurred concurrent to the complete nitrate and nitrite reduc-
tion, which was attributed to the cessation of the acetate utilization by the denitrifiers.
This result is corroborated by a net increase in the percentage of the denitrifying after
the nitrate and nitrite reductions. A net increase in the percentage of the acetoclastic
methanogens was predicted at 0.5 and 1 days, which coincided with the complete nitrate
and nitrite reduction as well as the beginning of the acetate utilization as a result of the
methane production.
4.3.2 Simulation continuous reactor
4.3.2.1 Analysis of scenarios
Figures 4.4, 4.5 and 4.6 demonstrate the simulation results for the six scenarios descri-
bed in Table 4.3. The variables analyzed were the COD removal, nitrate concentration,
methane production, nitrogen gas, volatile fatty acids (VFA) concentration, pH and per-
centage of COD used by the denitrifying (DP) and methanogenic (MP) microorganisms.
From the results obtained from Scenarios S1, S2 and S3, it is observed that there was
a lower organic matter removal efficiency in Scenarios S1 and S2 (Figure 4.4a). These
results are attributed to the insufficient organic matter availability for the simultaneous
AD/DN process. In this case, denitrification represents the primary route for the organic
matter consumption (Figure 4.6b). The fastest nitrate reduction rates were observed in S3
(Figure 4.5a). However, a significant nitrate and nitrite accumulation (Figure 4.5a and b)
was observed in S1 and S2 as a consequence of the inhibition of the methanogenic activity
acetoclastic (Figure 4.4b). This inhibition caused an accumulation of the acetate and a
reduction in the organic matter removal efficiency. This statement is corroborated by the
lower methane production in S1 and S2 (Figure 4.4b). In S3, the initial suppression of
the methanogenesis and its recovery is observed. In several studies, the nitrate and nitrite
have shown to be inhibitory to the methanogenesis, partly due to the elevated level of the
redox potential caused by the reduction in nitrate and partly to a direct toxic effect of
the compound itself (Chen e Lin, 1993; Akunna et al., 1994). A slower reduction in the
94 wastewater
(a ) 5
N itr a te ( m g /L )
4 E x p e r im e n ta l
0 ( C o n tr o l)
[m L ]
3 1 0 S im u la te d
3 7 .5 0 ( C o n tr o l)
4
2
C H
1 0
3 7 .5
1
(b ) 0
A c e ta te [m g C O D /L ]
1 3 5 0 0 ( C o n tr o l)
1 0
9 0 0 3 7 .5
4 5 0
0
(c ) 4 0
-
N O 3 1 0
[m g /L ]
3 0
-
N O 2
1 0
2 0 -
-
2
N O 3
3 7 .5
, N O
-
1 0 N O 3 7 .5
-
3
2
N O
(d ) 0
1 0 0
D P 1 0
[% ]
7 5
M P 1 0
D P , D M
5 0 M P 3 7 .5
D P 3 7 .5
2 5
0
0 1 2 3 4 5 6 7 8
T im e [D a y s ]
Figure 4.3: Model calibration for a batch reactor inoculated with strict methanogenic sludge after the
addition of different concentrations of nitrate 0 (control), 10 and 37.5 mg N/L: (a) methane production;
(b) production/consumption profiles of acetate; (c) consumption of nitrate; and (d) (DP) and (MP) % of
COD used for the denitrification and methanogenic process, respectively.
nitrate and nitrite associated with S1 and S2, could be the reason for the slower nitrogen
gas production (Figure 4.5c).
(a ) 1 0 0
8 0
R e m o v a l C O D [% ]
6 0
4 0
2 0
(b ) 0
/d a y ]
7 5 0
4
[k g C H
5 0 0
4
C H
2 5 0
(c ) 0
S 1
6 0 0 S 2
V F A [ m g C O D /L ]
S 3
S 4 ( C o n tr o l)
4 0 0 S 5
S 6
2 0 0
0
0 2 5 5 0 7 5 1 0 0 1 2 5 1 5 0
T im e [d a y s ]
Figure 4.4: Results of the model simulation for continuous flow systems treating domestic wastewater:
(a) removal COD; (b) methane production; and (c) VFA
The alkalinity can be generated in the system during the partial denitrification and
causes a significant increase in the pH higher than of the control condition, thus confirming
the occurrence of the denitrification process. Furthermore, Ramakrishnan and Gupta
(2008) reported the increased alkalinity in the effluent as a result of the denitrification.
This increase is due to the use of a strong acid (acetate) and a weak acid production
(CO2 , aq) for the methanogenesis and denitrification processes. The pH value (Figure
4.6a) remains within the range of 6-9, as recommended in the literature. A pH value lower
than six and higher than nine can inhibit both the acetogenic stage (Campos e Flotats,
2003; Huiliñir et al., 2008) and the denitrification process (Estuardo et al., 2008).
96 wastewater
(a ) 1 2 0
1 2 0
1 0 0 1 0 0
N itr a te [m g /L ]
8 0 8 0
6 0
6 0
4 0
4 0 2 0
0
2 0 0 5 1 0 1 5 2 0
(b ) 1 2 0
2 0
S 1
1 0 0 S 2
1 5
S 3
8 0
N itr ite [m g /L ]
S 4
S 5 1 0
6 0
S 6
4 0 5
2 0 0
0 5 1 0 1 5 2 0
(c ) 0
2 0 0
/d a y ]
1 5 0
2
[k g N
1 0 0
2
N
5 0
0
0 2 5 5 0 7 5 1 0 0 1 2 5 1 5 0
T im e [D a y s ]
(a) consumption of nitrate; (b) consumption of nitrite; and (c) nitrogen gas production
The control scenario has the highest removal efficiency of the organic matter. However,
the simulation results for the strong domestic wastewater indicate that in the presence of
nitrate, the S3, S5 and S6 obtained the highest removal efficiency of the organic matter
(Figure 4.4a). This finding could be explained by the largest consumption of both the VFAs
(Figure 4.4c) by denitrifying and the acetate by acetoclastic methanogenic, resulting in the
largest methane production and an increasing denitrification rate, as indicated in Figures
4.4b and 4.5a. In Figures 4.6b and c, it can be observed that for Scenarios S3 and S6, the
primary route for the soluble organic matter consumption is the methanogenesis. Initially,
the VFA is used for denitrifying, and the acetate is consumed then methanogenic process.
The acetate is the predominant product, representing 45% of the total VFAs. This result
agrees with those obtained by Akunna et al. (1994), Bernet et al. (1996) and Mosquera-
Corral et al. (2001). However, in S5, the primary route consumption of the organic matter
is the denitrification, consequently, the methane production is lower. Additionally, it is
possible to observe a rapid nitrate reduction rate during the initial days with respect
to Scenarios S3, S5 and S6 (Figure 4.5a). A faster denitrification rate results in a pH
within the ranges recommended for a high survival rate of the anaerobic and denitrifying
microorganisms, even with the accumulation of acid in the reactor during the initial days.
(a ) 8 ,0
7 ,5
7 ,0
p H
6 ,5
6 ,0
(b ) 1 0 0
7 5
D .P . [% ]
5 0
2 5
(c ) 1 0 0
7 5
M .P . [% ]
5 0
2 5
0
0 2 5 5 0 7 5 1 0 0 1 2 5 1 5 0
T im e [D a y s ]
(a) pH; (b) (DP) % of COD used for denitrification; and (c) (MP) % of COD used for methanogenic.
Although Scenarios S2 and S5 have the same COD/NO−

3 ratio, it is observed that the
organic matter removal efficiency and the gas nitrogen production in S5 were higher due
to increased availability of the electron donor to perform the denitrification process. The
nitrate and nitrite accumulation resulted in a slower rate of the initial methane production.
Therefore, although Scenarios S3, S5 and S6 have similar values of organic matter removal
efficiency, it can be observed that Scenario S6 is more favorable for achieving the AD/DN
98 wastewater
processes in terms of the SCNR. In this scenario, approximately 70% of the organic mat-
ter is removed, and the methane production and nitrate reduction rates are higher. This
behavior has also been experimentally observed by Akunna et al. (1992). They repor-
ted that the simultaneous denitrification and methanogenesis process occurred at 8.86 ≤
COD/NO− −
3 ≤ 53, while only denitrification was observed at COD/NO3 ≤ 8.86. Although
S6 has a higher methane production, a few researchers reported that the considerable losses
of methane in the anaerobic reactors treating domestic wastewater are primarily due to
the inappropriate gas-liquid-solid (GLS) separator device design, construction, and ope-
ration. Moreover, the inevitable loss of the dissolved methane in the effluent represents
a considerable fraction of the total methane produced from the low-strength wastewaters
(Kobayashi et al., 1983; Foresti et al., 2006).
4.3.2.2 The effect of HRT on SCNR
Figures 4.7, 4.8 and 4.9 illustrate the effects of the HRT on the following properties:
COD removal, nitrate concentration, methane production, pH, VFA production, nitrogen
production, DP and MP. The HRT was gradually decreased from 15 to 4 hours as follows:
15, 10, and 4 hours on day 0, 50, and 100, respectively.
Figure 4.7a depicts a decrease in the organic matter removal efficiency when the HRT
decreases from 15 to 4 hours. The reason for the reduction in the COD removal efficiency
was observed because of the insufficient contact time available for the microorganisms to
mineralize the organic matters, allowing a part of the influent COD to leave the reactor
without proper treatment (Leitao et al., 2006). The organic matter removal efficiency
obtained in all of the scenarios was observed to be similar, with the exception of the
control scenario. This finding indicates that for the same increase in the HRT from 6
to 15 hours during the SCNR for strong domestic wastewater, the soluble organic matter
supply is limited at the beginning of the experiment, thus incomplete denitrification occurs,
resulting in the accumulation of the acetate and a decrease in the COD removal. In S5,
the soluble organic matter removal rate is higher in the first day of the 15 hour HRT.
This result may be attributed to the faster utilization of the valerate, butyrate, propionate
and acetate by the denitrifying microorganisms (Figure 4.9b) in the initial days as well as
increasing the fermentation rate of higher order VFAs due to an increase in the population
size of the fermenting species after all of the N-oxides were consumed. The results of the
effluent VFA concentration and DP support this observation: at longer HRTs (15 hour),
an increased VFA consumption rate and a reduction in the denitrifiers from 75% to 37%
was observed.
(a ) 1 0 0 1 6
8 0
R e m o v a l C O D [% ]
1 2
H R T [h o u rs ]
6 0
8
4 0
4
2 0
0 0
(b ) 1 6
1 2 5 0
1 2
/d a y ]
1 0 0 0
H R T [h o u rs ]
7 5 0
4
8
[k g C H
5 0 0
4
4
C H
2 5 0
(c ) 0 0
1 6
1 0 0 0 S 3
S 4 ( C o n tr o l)
V F A [ m g C O D /L ]
S 5 1 2
H R T [h o u rs ]
7 5 0
S 6
5 0 0 8
2 5 0 4
0 0
0 2 5 5 0 7 5 1 0 0 1 2 5 1 5 0
T im e [d a y s ]
Figure 4.7: Results of the model simulation for continuous flow systems treating strong domestic was-
tewater experiencing three step decrements in the HRT from 15 to 4 hours: (a) removal COD; (b) methane
production and (c) VFA
Figure 4.7b depicts a decline in the methane production due to the decreasing HRT in
all of the scenarios, which caused a significant impact in S5. This result could be attributed
to the inhibition of the acetoclastic methanogens as a result of the accumulation of the
nitrates and nitrites (Figures 4.8a and b). The decrease in the nitrate reduction rate and
the lowest methane production is confirmed.
Figure 4.9a indicates that there was a small reduction in the pH in the initial moments,
where the HRT decreased, and remained without any variations. This results is primarily
related to the accumulation of the VFA. However, when the VFA consumption rate is higher
(Scenario S5), a increase in the pH during the initial hours is observed. It is interesting
to note that when the HRT is equal to 15 hours, the VFA consumption rate is lower in
Scenario S6, and the pH decrease and then increase thereafter. The decrease in the pH
was caused by fermentation, and the increase in the pH was caused by denitrification.
100 wastewater
(a ) 1 2 0
5 1 6
4
1 0 0
3
2 1 2
H R T [h o u rs ]
8 0
1
N itr a te [m g /L ]
6 0
0
S 3 0 2 5 5 0 7 5 1 0 0 1 2 5 1 5 0 8
S 4 ( C o n tr o l)
4 0
S 5
S 6 4
2 0
0
(b ) 1 2 0 1 0 10 6
8
1 0 0 6
4
N itr ite [m g /L ]
1 2
H R T [h o u rs ]
8 0 2
0
0 2 5 5 0 7 5 1 0 0 1 2 5 1 5 0
6 0 8
4 0
4
2 0
(c ) 4 0 0 10 6
3 0 0 1 2
/d a y ]
H R T [h o u rs ]
2 0 0
2
8
[k g N 2
1 0 0 4
N
0 0
0 2 5 5 0 7 5 1 0 0 1 2 5 1 5 0
T im e [d a y s ]
tewater experiencing three step decrements in the HRT from 15 to 4 hours: (a) consumption of nitrate;
(b) consumption of nitrite; and (c) nitrogen gas production
Thus, the pH varied due to the balance of the fermentation and denitrification activities
in simultaneous methanogenesis and denitrification processes in the domestic wastewater.
Generally, it can be observed that the alkalinity can be generated during the denitrification
process with an increase in the pH, which leads to a increase in the pH compared to that
of the control.
4.4 Conclusions
A mathematical model of the AD/DN processes for domestic wastewater based on the
ADM1 was presented. The model simulations for batch experiments conducted with the
sludge of domestic wastewater fed with acetate for various initial nitrate concentrations
agreed well with the experimental data. The simulation results under continuous conditi-
ons clearly indicate that the best scenario for the AD/DN processes in reactors treating
Seção 4.4. Conclusions 101
(a ) 8 1 6
1 2
H R T [h o u rs ]
7
p H
6
4
(b ) 5 10 6
7 5
H R T [h o u rs ]
1 2
D .P . [% ]
5 0
8
2 5 4
(c ) 0 01 6
7 5
1 2
H R T [h o u rs ]
M .P . [% ]
5 0
8
2 5 4
0 0
0 2 5 5 0 7 5 1 0 0 1 2 5 1 5 0
T im e [D a y s ]
tewater experiencing three step decrements in the HRT from 15 to 4 hours: (a) pH; (b) (DP) % of COD
used for denitrification; and (c) (MP) % of COD used for methanogenic
domestic wastewater was S5 where the COD/NO−

3 ratio equaled 10 and the HRT equaled
15 hours. However, when the COD/NO−

3 ratio is higher (> 10), as in S3 and S6, it is ne-
cessary to increase the HRT to improve the SCNR. In this case, there is a longer substrate
contact time with the microorganism a greater efficiency of the removal of organic matter
and nitrogen into the reactor. When the organic matter is limited, accumulation of the
denitrification intermediates occurs, and a lower efficiency is observed in removing the or-
ganic matter. To treat the strong domestic wastewater, the primary route of consumption
for the organic matter is denitrification and, consequently, the accumulation of the VFA
acetate as the prime inhibition product of the methane production. Model calibration is
needed to use the proposed mathematical model to model the full-scale anaerobic reactors
fed with domestic wastewater.
102 wastewater
Capı́tulo 5
Modeling of simultaneous carbon and nitrogen removal
in an anaerobic-anoxic reactor treating salmon fishery
wastewater
Artigo submetido para publicação na revista
Water Research (0043-1354),
Abstract
This chapter presents the modeling of the SCNR, hereby extending the Anaerobic Digestion
Model No. 1 (ADM1) with denitrification processes for a synthetic wastewater generated
by a fish processing industry. The model was calibrated and validated using previously
published experimental data from a batch reactor with suspended biomass fed with a
mixture of a synthetic substrate and a saline protein-rich salmon-plant effluent. Based
on a sensitivity analysis, five parameters of the original ADM1 and of the denitrification
processes were estimated using iterative methods, which optimized the parameters with
experimental results. Simulation with calibrated parameters indicated good agreement
with experimental data, with the mean absolute error below 40%. The differences between
the experimental and simulated data in the first 5 hours were observed because lag or a
death phase are not considered in the ADM1 model. Cross-validation results indicated
that the calibrated model was able to predict the experimental results of effluent COD,
acetate, propionate, butyrate, nitrate and nitrite flows with good accuracy. This study
demonstrates a useful approach for the implementation of the model for the treatment of
a specific industrial wastewater.
keywords
Simulation; Denitrification; Anaerobic Digestion Model No. 1; salmon fishery wastewater

Capı́tulo 5. Modeling of simultaneous carbon and nitrogen removal in an anaerobic-anoxic reactor treating salmon fishery
104 wastewater
5.1 Introduction
The wastewater composition from the fish processing industry varies widely depending
on the mode of operation of the processing plants. It contains large amounts of biodegra-
dable organic matter, mainly in the form of proteins and lipids and high levels of salts, e.g.
Cl− , Na+ , K+ , and Mg+ , especially when sea water is used in the processing. In addition,
a high ammonia concentration is sometimes observed due to the high blood content of
the wastewater. Among the different types of treatments for these wastewaters, biological
processes are the most appropriate (Chowdhury et al., 2010).
Although aerobic processes are traditionally used in the treatment of industrial efflu-
ents, anaerobic processes are gaining prominence due to their advantages, such as lower
sludge production, renewable energy production (biogas), smaller area requirements and
suitability for the further degradation of pollutants from high-load effluents (COD > 4000
mg/L), such as those produced by the fish industry (Chernicharo, 2007; Appels et al.,
2011). However, the nitrogen removal in the anaerobic system is practically negligible,
occurring only in the transformation of the organic nitrogen into ammonium nitrogen
(ammonification).
The biological treatment method to remove organic matter and nutrients is usually ac-
complished through the anaerobic-anoxic-aerobic (A2 O) process. The process for nitrogen
removal, includes autotrophic nitrification and heterotrophic denitrification, which conse-
cutively convert nitrate (NO− −
3 ) to nitrite (NO2 ), nitric oxide (NO), nitrous oxide (N2 O)
and finally to nitrogen gas (N2 ) by denitrifying bacteria under anoxic conditions. This
process is characterized by high energy demand, operational costs and the problems of
sludge disposal (Kim et al., 2013). The integration of anaerobic digestion and denitrifi-
cation (AD/DN) processes in an anaerobic pre-treatment system for simultaneous carbon
and nitrogen removal (SCNR) is an approach that can change the traditional concept of
biological nitrogen removal. This configuration could lead to a SCNR reduction in treat-
ment costs, the removal of toxic compounds, lower energy consumption and reduced sludge
production (Chuan et al., 2009; Winkler et al., 2015).
Recent studies have reported the use of integrated AD/DN processes in anaerobic pre-
treatment systems to SCNR treat wastewaters generated by the fish processing industry
(Huiliñir et al., 2011, 2012). As can be seen, there have been several experimental efforts
in published, but, aspects such as the modeling of the process have scarcely been studied.
Tugtas et al. (2006) presented a first attempt, including the denitrification process in four
steps (NO− −
3 → NO2 → NO → N2 O → N2 ) to the Anaerobic Digestion Model No. 1
(ADM1) model; however, this work does not included important variables such as the pH.
Huiliñir et al. (2009, 2011) have proposed models, however, the models were developed in
a steady state and specially for biofilm processes. Therefore, there is still a necessity for
improving the modeling of SCNR for fishery wastewater.
Mathematical models have proven their effectiveness for process design and operation.
With respect to anaerobic digestion, the ADM1 model was proposed as a common platform
for designing, developing, and validating models of anaerobic digestion processes (Bats-
tone et al., 2002a). Then, over the last ten years, ADM1 has become a practical tool for
dynamic modeling of anaerobic digestion. This model allows for the simulation and predic-
tion of anaerobic processes under different states and operating conditions, and has been
frequently expanded to accommodate an increasing range of processes (Batstone, 2006c;
Penumathsa et al., 2008; Girault et al., 2012; Souza et al., 2013). Although the inclusion
of denitrification in ADM1 has been proposed (Tugtas et al., 2006), the application of this
modified ADM1 to fishery wastewater has not yet been proposed. The use of a widely
accepted model could help the fishery industry in the application and implementation of
new biological technology such as SCNR.
Thus, the present chapter proposes an approach for modeling the SCNR in an anaero-
bic/anoxic reactor fed with a synthetic substrate to mimic the wastewaters generated by
the fish processing industry. The model was calibrated and validated by using lab-scale
data from a batch reactor, as proposed by (Huiliñir et al., 2012).
5.2 Methodology
5.2.1 Model description and implementation
The original ADM1 is described in a scientific and technical report prepared by an

International Water Association (IWA) Task Group (Batstone et al., 2002a). This model
takes into account seven bacterial groups. The biological degradation processes are descri-
bed using Monod kinetics, while the extracellular processes (disintegration and hydrolysis)
and the biomass decay are described by using first order kinetics.
The ADM1 extension with denitrification for salmon fishery wastewater was coded
using the Matlab/Simulink
R 2010a toolset following the original ADM1 by Batstone
et al. (2002a). The model was implemented as a set of ordinary differential equations
using the ODE15s as the numerical solver.
Based on the experimental observations previously presented in Huiliñir et al. (2012)
carbohydrates and their degradation intermediates were neglected as the organic compo-
106 wastewater
nents of the substrate particulate. Therefore, the complex substrate used in this work
is defined as protein and lipids. As protein and lipids in the fishery wastewater have a
recalcitrant fraction (Aspe et al., 1997), and based on the ADM1 model and experimental
evidence, it was assumed that the protein and lipids has a biodegradable part and an inert
part.
The proposed model considered only a two-step denitrification process (NO− −
3 → NO2
→ N2 ). This assumption is consistent with many nitrification and denitrification models,

including those proposed by Viraj e Bruce (2000), Henze et al. (2000), and Huiliñir et al.
(2009). Denitrifier biomass (facultative anaerobes) was introduced as a fraction of the fer-
mentative consortia that is capable of fermenting valerate, butyrate and propionate in the
absence of N-oxides (nitrate and nitrite) (Tugtas et al., 2006). In the presence of N-oxides,
it was assumed that butyrate, valerate, propionate, and acetate were the electron donors
utilized by the denitrifiers (Kluber e Conrad, 1998a). In accordance with Aquino e Stuckey
(2008), the hydrogen (H2 ) consumption was omitted because of the slow consumption due
to the competition of the denitrifiers with the hydrogenotrophic methanogens. Further-
more, the consumption of sugars, amino acids, and long-chain fatty acids (LCFAs) by
denitrifiers was neglected also because of their slower consumption kinetics (Tugtas et al.,
2006).
The pH charge balance prediction proposed by Batstone et al. (2002a) was modified
by adding the concentration of NO− −
3 and NO2 ions. Denitrifying and anaerobic microor-
ganisms were assumed to reach optimal growth at a nearly neutral pH (Campos e Flotats,
2003; Estuardo et al., 2008).
Figure 5.1 indicate the proposed conceptual model for the simultaneous AD/DN pro-
cesses of salmon fishery wastewater. It includes the following biochemical processes: (1)
hydrolysis from protein; (2) hydrolysis from lipids; (3) acidogenesis from amino acids; (4)
acetogenesis from long-chain fatty acids; (5) propionate from acetogenesis; (6) butyrate
and valerate acetogenesis; (7) acetoclastic methanogenesis; (8) hydrogenotrophic metha-
nogenesis; (9) nitrate reduction; (10) nitrite reduction; and (9) the death and disintegration
of biomass into carbohydrate, protein, lipids, and inert substances. The biological kinetic
rate expressions and the stoichiometric parameters of the extended ADM1 model related
to the AD/DN process are reported in the Petersen matrix form as indicated in Table 5.1.
The degradation equations of the soluble substrate (valerate, butyrate, propionate and
acetate, i=4-7) in ADM1 including denitrification processes with the proposed modification
is given by:
Table 5.1 - Denitrification process included in the modified Anaerobic Digestion Model no1
Component → i 4 5 6 7 25 26 27 30 Process rate (ρj,n )
j Process Sva Sbu Spro Sac Sno3 Sno2 Sn2 Xf,nox (kg COD m−3 .d−1 )
↓
Sva Sno3
8-1 Uptake NO−
-1 - (1-Yc4 )/1.14 (1-Yc4 )/1.14 Yc4 kno3,va K Xf,nox k
3 no3,va +Sva no3 +Sno3
of vale-
rate
Sva
8-2 NO−
-1 - (1-Yc4 )/0.57 (1-Yc4 )/0.57 Yc4 kno2,va K Xf,nox k Sno2
2 no2,va +Sva no2 +Sno2
Sbu Sno3
9-1 Uptake NO−
-1 - (1-Yc4 )/1.14 (1-Yc4 )/1.14 Yc4 kno3,bu K X f,nox kno3 +Sno3
3 no3,bu +Sva
of
buty-
rate
Sbu
9-2 NO−
-1 - (1-Yc4 )/0.57 (1-Yc4 )/0.57 Yc4 kno2,bu K Xf,nox k Sno2
2 no2,bu +Sva no2 +Sno2
Spro
Seção 5.2. Methodology
10-1 Uptake NO−

-1 - (1-Ypro )/1.14 (1-Ypro )/1.14 Ypro kno3,pro K Xf,nox k Sno3
3 no3,pro +Spro no3 +Sno3
of pro-
pionate
Spro Sno2
10-2 NO−
-1 - (1-Ypro )/0.57 (1-Ypro )/0.57 Ypro kno2,pro K Xf,nox k
2 no2,pro +Spro no2 +Sno2
Sac Sno3
11-1 Uptake NO−
-1 - (1-Yac )/1.14 (1-Yac )/1.14 Yac kno3,ac K Xf,nox k
3 no3,ac +Sac no3 +Sno3
of
acetate
Sac Sno2
11-2 NO−
-1 - (1-Yac )/0.57 (1-Yac )/0.57 Yac kno2,ac K Xf,nox k
2 no2,ac +Sac no2 +Sno2
20 Decay -1 kdec,Xf,nox Xf,nox

107
108 wastewater
Inactive
biomass
Protein Lipids
NO3- -N 11 Disintegration
(Xpr) (Xli)
Xch Xpr Xli Inert
Hydrolysis 1 2
9
Denitrification Amino Acids Long Chain
(Saa) Fatty Acids (Sfa)
3 4
Acidogenesis Death
Propionate Valerate (Sva)

NO2- -N (Spro) Butyrate (Sbu)
6
Acetogenesis 5
0 Acetate (Sac) Hydrogen (Sh2)

10
Methanogenesis
N2 7 8
Methane (Sch4)
Figure 5.1: Conceptual model of the simultaneous AD/DN processes of salmon fishery wastewater.
dSliq,i q X X
= (Sin,i − Sliq,i ) − ρj νi,j − (ρj + ρj,nox ) νi,j − ρ11 ν7,11 Inox − ρj,n (5.1)
dt Vliq j=5−7 j=8−10
In Equation 5.1, the first term of the right side is the substrate mass balance for a
continuous stirred-tank reactor (CSTR). The second term represents the process equations
for the uptake of amino acids by fermentation, obtained from the ADM1 matrix. The
third term is the fermentation (acetogenesis) of soluble compounds i in the absence of N-
oxides by denitrifiers. The fourth term represents the methanogens obtained by the ADM1
matrix multiplied by an N-oxide inhibition factor (Inox ) to account for the inhibitory effect
of N-oxide on acetoclastic methanogens. The inhibitory effect was simulated using non-
competitive inhibition functions. The fifth term represents the uptake of soluble substrates
by denitrifiers in the presence of nitrate and nitrite. The general rate equation for the
utilization of butyrate, valerate and propionate by facultative anaerobes in the absence of
N-oxides is expressed as:

Si KS,no3 KS,no2 Si
ρj,nox = km,j × Xf,nox , (5.2)
KS + Si KS,no3 + Sno3 KS,no2 + Sno2 Sva + Sbu + Spro
where km,j is the rate constant for the fermentation of process j (kg substrate COD/kg
biomass-COD.d), and Xf,nox is the population facultative anaerobes (kg biomass-COD/m3 ).
The general rate equation for the utilization of N-oxides (nitrate, nitrite and nitrogen gas,
i = 25-27) is defined as follows:
dSi,nox q
= (Sin,i,nox − Sliq,i,nox ) + ρj,n , (5.3)
dt Vliq
where Si,nox is the concentration of N-oxide i (kg N/m3 ) and ρi,nox is the rate of produc-
tion/consumption of N-compound i as a result of denitrification (kg N/m3 .d). Equation 5.4
represents the mass balance used by the ADM1 model to describe the dynamic behavior
of particulates in the liquid phase:
dXliq,i qXin,i Xliq,i X

= − + ρj νi,j , j = 1, 3, 4, 6 − 12, 14 − 20. (5.4)
dt Vliq Vliq
To predict the pH temporal profile and associated potential inhibition in acid-base

equilibrium, the charge balances of the ADM1 model were then modified by adding the
SNO−3 and SNO−2 components (Huiliñir e Montalvo, 2013), see Equation 5.5,
SAc− SPr− SBu− SVa−

SCat+ + SNH+ + SH+ − SHCO− − − − − − S− −
NO3 − SNO2 − SOH− − SAn− = 0 , (5.5)
4 3 64 112 160 208
where, SAc− , SPr− , SBu− , SVa− , SNH+4 , SHCO−3 , SNO−3 and SNO−2 are the concentrations of
ionized forms of buffer components.
Whereas the nitrogen gas is assumed to obey the ideal gas law and exists at a tempe-
rature that is equivalent to the liquid phase temperature in a constant volume (completely
mixed) and a constant pressure headspace (Batstone et al., 2002a), the general dynamic
gas phase concentration equation of nitrogen gas is written by Equation 5.6:
dSgas,N2 Sgas,N2 qgas Vliq
= + ρT,N2 , (5.6)
dt Vgas Vgas
where Sgas,N2 is the gas phase concentration of gas of nitrogen; qgas is the gas flow outside
the reactor; and Vgas is the reactor gas volume (headspace volume). ρT,N2 is the specific
mass transfer rate of gas nitrogen expressed as follows:
ρT,N2 = kL a(Sliq,N2 − 28KH,N2 pgas,N2 ) , (5.7)
where kL a is the volumetric gas-liquid mass transfer coefficient, KH,N2 is the Henry’s law
coefficient of nitrogen, Sliq,N2 is the liquid phase concentration of nitrogen gas and pgas,N2
is the gas phase pressure of nitrogen calculated from the ideal gas law as follows:
pgas,N2 = Sgas,N2 RT . (5.8)
By summing the gas phase concentration of all gas components, a dynamic behavior
is shows (CH4 , CO2 , H2 and N2 ), and the total pressure of the gas phase above the liquid
110 wastewater
is equal to one atmosphere minus the vapor pressure of water at the specified reactor
temperature. The gas production rate can be estimated by Equation 5.9:
RT ρ
T,H2 ρT,CH4
qgas = Vliq + + ρT,CO2 + ρT,N2 . (5.9)
Pgas − pgas,H2 O 16 64
The set of governing equations is solved simultaneously to simulate the dynamics and
biological kinetics in reactors. Details about model governing equations, input parameters,
and underlying assumptions are described in Batstone et al. (2002a) and Tugtas et al.
(2006).
5.2.2 Experimental setup and initial conditions
Experimental observations from a characterization study of the SCNR process using

salmon effluents carried out by (Huiliñir et al., 2012) were used for modelling, calibration
and validation. During these experiments, an inoculum was obtained from a fed batch
reactor, which was fed weekly to obtain an HRT of 125 days.
During two years of operation, the removal efficiencies of nitrate and total organic
carbon (TOC) reached 100 and 79%, respectively. In addition to peptone substrate, 24
g L−1 of NaCl was added to mimic a fish and salmon effluent. The experiments were
performed at a temperature of 37 ◦ C and at a constant pH of 7.5 in batch reactors with
liquid and gas volume equal to 800 and 195 mL, respectively. More details about the
experimental set-up, analytical methods and operating condition can be found in (Huiliñir
et al., 2012).
In the experiments conducted by Huiliñir et al. (2012), the initial nitrate concentration
was varied to obtain C/N ratios ranging from 2.0 to 63.0 mg TOC/ mg NO−
3 . In this study,
two sets of experimental data were chosen for calibration and direct validation (data set D1)
and cross validation (data set D2) of the proposed model. Furthermore, experimental data
measured in TOC were converted to COD because this is the ADM1 unit. The COD/TOC
conversion factor of 2.185 mg COD/ mg TOC was based on experiments conducted by
(Aspé et al., 2001) with wastewater from the salmon industry. The authors found a linear
correlation with a regression coefficient (R2 ) equal to 0.921.
The influent COD concentrations of data set D1 and D2 were 2,720 mg COD /L and
2,904 mg COD/L, respectively. The initial nitrate concentrations of data set D1 and
D2 were 34.95 mg NO− −
3 / L and 85.95 mg NO3 / L, respectively. The initial biomass
concentration of the seven microbial populations contained in the ADM1 (Xaa , Xc4 , Xpro ,
Xac and Xh2 ) are defined as follows:
Yi
Xi = Xtotal , i = 17, ..., 23 , (5.10)
Ynox
where Yi is the observed yield coefficient of each microbial group, Ynox is the overall observed
yield coefficient, and Xtotal is the total biomass concentration in the reactor measured as
volatile suspended solids. The Yi and Ynox values were obtained from the yield coefficients
and decay rate (0.02 d−1 ) as suggested by Batstone et al. (2002a) and Tugtas et al. (2006),
respectively. Denitrifiers were assumed to be 40% of the butyrate, valerate and propionate
degraders in the baseline model.
5.2.3 Sensitivity analysis
Sensitivity analysis has been widely applied to reduce the complexity of parameter
estimation procedure, to determine the significance of model parameters and to identify
the dominant parameters (Tartakovsky et al., 2008). Therefore, a parametric sensitivity
analysis was performed to define the most sensitive parameters of the ADM1 extension.
The parameters evaluated were hydrolysis constant rates for proteins (khyd,pr ), Monod ma-
ximum specific substrate uptake rate (km ), and half-saturation constant (KS ). Although
Jeong et al. (2005) mentioned that the yield of product on substrate (f ) and yield of bio-
mass production (Y ) values showed high sensitivities to components of the ADM1 exten-
sion analyzed, in this study these parameters were negligible due to their small variations.
Concerning to parameters denitrification only the half saturation coefficient of component
i on process j (KS,j ) and Monod maximum specific uptake rate for process j (km,j ) were
evaluated.
The sensitivities were quantified in terms of the variation of measurable process va-
riables under the perturbation of model parameters in their neighborhood domain. The
average of absolute differences between simulation results with prior determined parame-
ters values suggested by Soto et al. (2002), Gonzalez et al. (2005), Tugtas et al. (2006), and
Batstone et al. (2002a) and with parameters with a relative change of target parameter
was used as the sensitivity index (SI), as presented in the following equation:
P
|CST D (t) − CSEN S (t)|
SI = , (5.11)
N
where CST D (t) and CSEN S (t) are the simulation results with the suggested parameter values
and the parameters with a relative change of target parameter, respectively, for each given
time (t), and N is the number of data. The sensitivity analysis of kinetic parameters for
four components was carried out by changing the value of a target parameter from 50% to
50% with respect to their suggested values (Jeong et al., 2005).
112 wastewater
5.2.4 Model calibration and validation procedure
Calibration of the more sensitive model parameters was carried out. Model calibration
was performed on an expert basis by a trial and error approach, driven by knowledge
from the sensitivity analysis and using the parameter ranges reported in the literature as
constraints. The iterative procedure reported by Mendes et al. (2015) was applied.
Direct and cross validation are usually considered as to be steps of the model vali-
dation procedure. The data then were then divided into two subsets as recommended by
Donoso-Bravo et al. (2011): (1) data used during model calibration (data set D1) for direct
validation, and (2) unseen data (data set D2) for cross validation. The accuracy of the
predictions for direct and cross validation were evaluated by using the mean absolute error,
which was classified as a quantitative prediction with high (±10%), medium (10-40%) or
low (40-70%) accuracy (Batstone e Keller, 2003).
Model calibration was performed by comparing the model results with the experimental
measurements for the nitrate, nitrite, COD, butyrate, propionate, and acetate concentrati-
ons and adjusting the unknown parameter until the model results adequately fit the expe-
rimental observations. Optimization was made using the fminsearch optimization function
in the MATLAB
R toolbox, which is based on the Nelder-Mead simplex algorithm. The
biomass initial values are key parameters in the model calibration. The typical strategy
proposed by Donoso-Bravo et al. (2011) was used to optimize the initial biomass values. It
consists of simulating the process several times until the biomass concentrations are cons-
tant. After the calibration and direct validation process were complete, cross-validation
was performed to verify the agreement between the model results and the experimental
measurements using the previously calibrated value parameter. The cross-validation was
performed using independent data set D2.
5.3 Results and Discussion
5.3.1 Sensitivity analysis and model calibration
The results of the sensitive analysis for organic matter (COD) and nitrate reduction
(NO−
3 ) are indicated in the Figures 5.2 and 5.3, respectively. As indicated in Figures
5.2a,b,c and 5.3a,b,c, the parameters, proteins hydrolysis constant (khyd,pr ), Monod ma-
ximum specific uptake rate for amino acids utilizers (km,aa ) and half saturation value for
amino acids utilizers (KS,aa ) indicate more influence in the COD and NO−
3 . High sensitivity
values were also observed for Monod maximum specific uptake rate for acetate utilization
rates by the denitrifiers (kno3,ac ) and half saturation value for the utilization of the acetate
by the denitrifiers (KS,no3,ac ) (5.2d and 5.3d). Based on the sensitivity results, we decided
to use khyd,pr , km,aa , KS,aa , kno3,ac and KS,no3,ac for parameter estimation.
Table 5.2 shows the initial values of the kinetic parameters and their respective opti-
mized values. The initial values of the protein hydrolysis parameters (khyd,pr ) was found in
the range of values found in the literature (Gonzalez et al., 2005). khyd,pr parameter was
chosen based on Gonzalez et al. (2005) because the similarity of the substrates (fishing
wastewaters). Concerning the calibration of the denitrification parameters, the KS,no3 and
KS,no2 values for the utilization of the nitrate and nitrite, respectively, were chosen based
on data from the literature (Soto et al., 2002). The others, such as, the km values of the
electron donors by the denitrifiers, KS the values of electron donors by the denitrifiers and
yield of electron donor denitrifiers (Y ) were maintained the same as those suggested by
Tugtas et al. (2006).
(a ) 3 0 (b ) 3 0
k h y d ,p r
k m ,a a
2 5 2 5 k m ,c 4
k m ,p ro
2 0 2 0 k
]
m ,a c
-1
-1
k
S I [m g C O D .L
S I [m g C O D .L
m ,h 2
1 5 1 5
1 0 1 0
5 5
0 0
-5 0 -2 5 0 2 5 5 0 -5 0 -2 5 0 2 5 5 0
R a n g e [% ] R a n g e [% ]
(c ) 3 0 (d ) 3 0
K k n o 3 ,b u
S ,a a
2 5 K K
S ,c 4 2 5 S ,b u ,n o 3
K k n o 3 ,p ro
S ,p ro
K K
2 0 S ,a c 2 0 S ,p ro ,n o 3
]
k
-1
-1
K n o 3 ,a c
S I [m g C O D .L
S ,h 2
S I [m g C O D .L
K S ,a c ,n o 3
1 5 1 5
K S ,n o 3
1 0 1 0
5 5
0 0
-5 0 -2 5 0 2 5 5 0 -5 0 -2 5 0 2 5 5 0
R a n g e [% ] R a n g e [% ]
Figure 5.2: Sensitivity indices (SI) of the organic matter calculated for different ranges of (a) hydrolysis
constant rate; (b) Monod maximum specific uptake rates; (c) half-saturation constants; and (d) parameters
denitrification
In the calibration procedure, the disintegration constant was first estimated by mat-
ching the model outputs with the measured outputs. Then, the model outputs for the
concentrations of acetic acid, propionic acid and butyrate acid were changed by changing
the half saturation constants and maximum uptake rates.
114 wastewater
(a ) 3 (b ) 3
k h y d ,p r
k m ,a a
k m ,c 4
k
]
-1
m ,p ro
.L
2 2 k m ,a c
]
-
3
-1
S I [m g N O
m ,h 2
.L
-
3
S I [m g N O
1 1
0 0
-5 0 -2 5 0 2 5 5 0 -5 0 -2 5 0 2 5 5 0
R a n g e [% ] R a n g e [% ]
(c ) 3 (d ) 3
K S ,a a
k n o 3 ,b u
K S ,c 4
K S ,b u ,n o 3
K S ,p ro
k n o 3 ,p ro
]
-1
2 K 2 K
.L
]
S ,a c S ,p ro ,n o 3
-1
K k
-
3
.L
S ,h 2 n o 3 ,a c
S I [m g N O
-
3
S I [m g N O K S ,a c ,n o 3
K S ,n o 3
1 1
0 0
-5 0 -2 5 0 2 5 5 0 -5 0 -2 5 0 2 5 5 0
R a n g e [% ] R a n g e [% ]
Figure 5.3: Sensitivity indices (SI) of the nitrate reduction calculated for different ranges of (a) hydrolysis
constant rate; (b) Monod maximum specific uptake rates; (c) half-saturation constants; and (d) parameters
denitrification
The optimum values for the parameters obtained throughout calibration are given in
Table 5.2. The parameters as, km and KS for acetic acid, propionic acid and butyrate acid
were optimized together as they exhibit the lowest correlation and the highest relevance
(Batstone et al., 2003).
Table 5.2 - Initial and estimated kinetic parameters of the implemented extended ADM1.
Parameter Units Initial values Calibrated values Reference
khyd,pr h−1 0.09 0.05 Gonzalez et al. (2005)
km,aa h−1 2.08 1.87 Batstone et al. (2002a)
kno3,ac h−1 0.12 0.09 Tugtas et al. (2006)
KS,aa kg COD m−3 0.30 0.23 Batstone et al. (2002a)
KS,no3,ac kg COD m−3 0.50 0.38 Tugtas et al. (2006)
KS,no3 kg NO3 m−3 0.00047 0.00047 Soto et al. (2002)
KS,no2 kg NO2 m−3 0.00036 0.00036 Soto et al. (2002)
5.3.2 Direct validation
Figure 5.4 shows the simulation results after calibration for data set D1 in comparison
to the experimental values for COD, VFAs, nitrate and for nitrite. It is expected that
the simulation results merge with the observed data by the end of the calibration period.
In general, the results of the calibration indicated a good fit between the model and the
measured data, exhibiting a mean absolute error below 40% (15%-38%), which is considered
as a mean accuracy quantitative prediction (Batstone e Keller, 2003). However, a higher
relative deviation between the measured and predicted values of the COD, VFAs, nitrate
and nitrite and quality have been found in the 5 first hours. This behavior can be explained
because in the ADM1 model, a lag or death phase is not considered. It is assumed that
there is unrestricted growth of the biomass.
The complete reduction of nitrate and nitrite occurred in an period of 11 h. The ma-
ximum nitrate consumption rate was found between 6 and 8 h, when nitrite accumulation
also took place, reaching a maximum concentration of 20 mg NO−
2 -N/L. A lag phase in
the reactor was observed in the first 5 hours of operation. During this period, the organic
matter in the reactor (protein) was not degraded by the microorganisms. This phenome-
non likely occurs because of the changing of the substrate used at the beginning of the
experiment. The substrate for the inocula reactor was proposed by Akunna et al. (1992),
where K2 HPO4 and KH2 PO4 buffers were discarded and glucose as the organic matter
source was replaced by peptone. As the peptone degradation rate is slower than glucose,
microorganisms would have to adapt to the new conditions of the reactor. Thus, the re-
sulting protein degradation VFAs and nitrite to nitrate reduction only occur after 5 hours
of operation. In the period from 5 to 10 hours, the VFAs are consumed simultaneously
with nitrite and nitrate degradation. Therefore, its consumption was directly associated
with the denitrification process. After the nitrite and nitrate concentration became cons-
tant, the COD concentration became almost constant, and a VFA accumulated up to a
concentration of 1002 mg COD L−1 . Acetate is the main VFA, representing 45% of the
total VFA.
5.3.3 Cross validation
A model validation study was performed to assess the quality and applicability of the
calibrated model. The model outputs were compared with data set D2 without changing
the previously optimized parameter set.
Figure 5.5 shows the comparison of model predictions and experimental values for the
process variables during the validation study. As seen, the COD, VFAs, nitrate and nitrite
levels were moderately predicted by the model, showing a mean absolute error below 40%
(19%-36%).
116 wastewater
(a ) 1 0 0 0 3 0 0 0
7 5 0 2 2 5 0
]
]
-1
-1
[m g C O D .L
V F A s [m g C O D .L
C O D e x p e r im e n ta l
S a c
e x p e r im e n ta l
5 0 0 S p ro
e x p e r im e n ta l 1 5 0 0
S b u
S s im u la te d
C O D
a c
S p ro
s im u la te d
2 5 0 S b u
s im u la te d 7 5 0
C O D s im u la te d
(b ) 0 0
-
4 0 E x p e r im e n ta l N O 3
-
E x p e r im e n ta l N O 2
-
S im u la te d N O
]
3
-1
3 0
[m g N .L
-
S im u la te d N O 2
-
2
2 0
, N O
-
3
N O
1 0
0
0 5 1 0 1 5 2 0 2 5
T im e [h ]
Figure 5.4: Comparison between simulations and experimental data for D1 set after calibration
(a ) 1 0 0 0 3 6 0 0
7 5 0 2 7 0 0
]
-1
]
C O D e x p e r im e n ta l
-1
V F A s [m g C O D .L
C O D [m g C O D .L
S a c
S p ro
5 0 0 S b u
e x p e r im e n ta l 1 8 0 0
S a c
s im u la te d
S p ro
s im u la te d
S b u
s im u la te d
2 5 0 C O D s im u la te d 9 0 0
(b ) 0 0
-
E x p e r im e n ta l N O 3
-
7 5 E x p e r im e n ta l N O 2
-
S im u la te d N O 3
]
-1
-
S im u la te d N O
[m g N .L
5 0
-
2
, N O
-
3
N O
2 5
0
0 5 1 0 1 5 2 0 2 5
T im e [h ]
Figure 5.5: Validation of simulations with experimental results of D2 set

118 wastewater
5.4 Conclusions
An extension of ADM1 with nitrate reduction was proposed, calibrated and validated
to the describe the anaerobic digestion of salmon fishery wastewater. Sensitivity analysis
showed that among all kinetic parameters, khyd , km,process , and KS,process showed high sen-
sitivities to almost all components. Important parameters including khyd,pr , km,aa , KS,aa ,
kno3,ac and KS,no3,ac were estimated using iterative methods, which optimized the parame-
ters with experimental results. The model predictions regarding COD, acetate, propio-
nate, butyrate, nitrate and nitrite were in good agreement with the experimental data
produced during the model validation. The kinetic parameters and the model structure
proposed in this work can be considered valuable for describing the SCNR process in an
anaerobic/anoxic reactor fed with wastewaters generated by the fish processing industry,
increasing the applicability of ADM1 to a specific industrial wastewaters.
Capı́tulo 6
Conclusões e sugestões para trabalhos futuros
6.1 Conclusões
A compreensão dos processos de fermentação, metanogênese e desnitrificação para re-

mover simultaneamente carbono e nitrogênio em reatores anaeróbio-anóxico tratando águas
residuárias diluı́das e concentradas, é uma tarefa desafiadora. Nesse sentido, quando sa-
tisfatoriamente formulados, os modelos matemáticos constituem uma ferramenta útil ao
estudo do comportamento desses processos. Neste trabalho buscou-se avaliar a RSCN em
reator anaeróbio-anóxico tratando águas residuárias diluı́das (águas residuárias doméstica)
e concentradas (águas residuárias de processamento de salmão) por meio da modelagem
matemática e simulação.
As conclusões finais deste trabalho são apresentadas seguindo a mesma lógica de seu
desenvolvimento ao longo do texto.
No Capı́tulo 3 foram apresentados os resultados da aplicação do ADM1 para predizer o
comportamento de um digestor anaeróbio, utilizado para estabilização do lodo de sistema
de lodos ativados convencional operado sob condições mesofı́licas. Os resultados da análise
de sensibilidade permitiu identificar os parâmetros mais significativos para calibração do
ADM1, como: km,ac , km,h2 , km,su e km,aa . Os resultados da calibração mostram boa con-
cordância entre os dados experimentais e as previsões do modelo para as concentração de
propionato, acetato, produção e composição de biogás e pH. Os resultados da simulação
para avaliar o efeito do choque de carga orgânica no desempenho do processo anaeróbio
revelaram que valores de carga orgânica volumétrica acima de 35 kg DQO/m3 .dia causa
instabilidade e falha no processo de fermentação e consequentemente na metanogênese,
levando o acúmulo de AGV, falha no pH e redução na produção de metano. Conclui-se
que os resultados da aplicação do ADM1 corroboraram a potencial utilização do modelo
para prever o comportamento dos processos biológico de digestão anaeróbia.
120 Capı́tulo 6. Conclusões e sugestões para trabalhos futuros
No Capı́tulo 4 foram apresentados os resultados da calibração e simulação do modelo

matemático baseado no ADM1 para remoção simultânea de carbono e nitrogênio em reator
anaeróbio-anóxico tratando esgoto doméstico. Os resultados simulados aderiram bem aos
dados experimentais obtidos a partir de testes realizados em bateladas. As simulações
realizadas em condições de regime contı́nuo indicaram que o melhor cenário para RSCN
em reatores anaeróbio-anóxico tratando águas residuárias doméstica foi o com a razão
DQO/NO−
3 e TDH igual a 10 e 15 horas, respectivamente. No entanto, quando a razão
é menor que 10, é necessário aumentar o TDH para melhorar a RSCN. Os resultados das
simulações mostraram que quando a matéria orgânica é limitada, ocorre um acúmulo de
nitrato e nitrito e uma redução na eficiência foi observada.
No Capı́tulo 5 foram apresentados os resultados da calibração e validação do modelo
matemático baseado no ADM1 para remoção simultânea de carbono e nitrogênio em rea-
tor anaeróbio-anóxico tratando águas residuárias de processamento de salmão. Os dados
experimentais foram obtidos a partir de um reator em batelada anaeróbio-anóxico ali-
mentado com substrato sintético para simular um efluente de processamento de pescado.
Uma análise de sensibilidade foi realizada para identificar os parâmetros mais importantes
do modelo. Os resultados simulados das variáveis DQO, butirato, propionato, acetato,
nitrato e nitrito aderiram satisfatoriamente aos dados experimentais na calibração e va-
lidação do modelo com um erro médio absoluto variando entre 15 e 38%, respectivamente.
No entanto, foi observado um desvio entre os valores dos dados experimentais e dos si-
mulados nas cinco primeiras horas. Este comportamento esta relacionado à ausência da
modelagem cinética da fase lag no ADM1. Os resultados demonstraram que o modelo
matemático proposto nesse estudo pode ser considerado valioso para avaliar a RSCN em
reatores anaeróbio-anóxico tratando águas residuárias de processamento de pescado.
6.2 Sugestões para trabalhos futuros
A pesquisa realizada e os resultados obtidos despertam a abertura de várias frentes de

trabalho, dentre as quais destacam-se:
1. Estudos mais sofisticados e aprofundados para desenvolver extensões e modificações

do ADM1 contemplando:
• os processos simultâneos de fermentação, metanogênese, sulfetogênese e desni-

trificação heterotrófica;
• a remoção simultânea de carbono, nitrogênio e fósforo;

Seção 6.2. Sugestões para trabalhos futuros 121
• os processos simultâneos de fermentação, metanogênese, desnitrificação hete-

rotrófica e autotrófica considerando metano e compostos de enxofre como doa-
dores de elétrons; e
• o perı́odo de adaptação dos micro-organismos no meio (fase lag).
2. Determinar os parâmetros cinéticos intrı́nsecos aos processos de digestão anaeróbia

e desnitrificação em reatores tratando águas residuárias de origem doméstica;
3. Avaliar e aplicar a técnica de análise de componentes principais para simplificação

dos modelos matemáticos proposto nesse trabalho;
4. Avaliar os testes de biodegradabilidade anaeróbia para caracterização afluente da

extensão do ADM1, visando identificar as frações de DQO e parâmetros cinéticos
associados à sua degradação por meio da interpretação da curva da taxa de produção
de metano.
5. Desenvolver um modelo matemático com parâmetros distribuı́dos baseado noADM1

para simular os processos simultâneos de fermentação, metanogênese e desnitrificação
em um reator UASB alimentado com águas residuárias de origem doméstica.
6. Calibrar e validar o modelo matemático proposto no Capı́tulo 4 em um reator

anaeróbio-anóxico operado em regime contı́nuo;
7. Acoplar as adaptações do ADM1 para o processo de desnitrificação no ASM para

simular a RSCN através da recirculação efluente aeróbio nitrificado para o sistema
integrado anaeróbio-anóxico;
8. Aplicar métodos para avaliação da incerteza de medição associadas às variáveis de en-
trada e aos parâmetros dos modelos matemáticos proposto neste trabalho em regime
estacionário e transiente;
9. Aplicar estimadores de estado para inferir as composições (variáveis de estado) dos

modelos proposto neste trabalho.
122 Capı́tulo 6. Conclusões e sugestões para trabalhos futuros
Referências Bibliográficas
Aboutboul Y., Arviv R., van Rijn J., Anaerobic treatment of intensive fish culture effluents:
volatile fatty acid mediated denitrification, Aquaculture, 1995, vol. 133, p. 21
Akunna J., Bizeau C., Moletta R., Nitrate reduction by anaerobic sludge using glucose
at various nitrate concentrations - ammonification, denitrification and methanogenic
activities., Environmental Technology, 1994, vol. 15, p. 41
Akunna J., Bizeau C., Moletta R., Héduit A., Combined organic carbon and complete
nitrogen removal using anaerobic and aerobic upflow filters., Water Science and Tech-
nology, 1994, vol. 30, p. 297
Akunna J., Bizeau C., R M., Denitrification in anaerobic digesters: Possibilities and in-
fluence of wastewater COD/N-NOx ratio, Environmental Technology, 1992, vol. 13, p.
825
Andrews J. F., Dynamic Model of the Anaerobic Digestion Process, Sanitary Engineering
Division, 1969, vol. 21, p. 95
Andrews J. F., Kinetic models of biological waste treatment processes, Biotechnology and
bioengineering symposium, 1971, vol. 2, p. 25
Angelidaki I., Ellegaard L., Ahring B. K., A comprehensive model of anaerobic bioconver-
sion of complex substrates to biogas, Biotechnology and Bioengineering, 1999, vol. 63,
p. 363
Angelidaki I., Ellegard E., Ahring B., A mathematical model for dynamic simulation of
anaerobic digestion of complex substrates: focusing on ammonia inhibition, Biotechno-
logy and Bioengineering, 1993, vol. 42, p. 159
124 Referências Bibliográficas
Appels L., Lauwers J., Degreve J., Helsen L., Lievens B., Willems K., Anaerobic digestion
in global bio-energy production: potential and research challenges, Renewable Sustaina-
ble Energy Review, 2011, vol. 15, p. 4295
Aquino S., Stuckey D., Integrated model of the production of soluble microbial products
(SMP) and extracellular polymeric substances (EPS) in anaerobic chemostats during
transient conditions, Biochemical Engineering, 2008, vol. 38, p. 138
Aspé E., Jara M., Roeckel M., Ammonia inhibition in the anaerobic treatment of fishery
effluents., Water Environment Research, 2001, vol. 73, p. 154
Aspe E., Marti M., Roeckel M., Anaerobic treatment of fishery wastewater using a marine
sediment inoculum, Water Research, 1997, vol. 31, p. 2147
Baloch M. I., Akunna J. C., Collier P. J., Carbon and Nitrogen Removal in a Granular
Bed Baffled Reactor, Environmental Technology, 2006, vol. 27, p. 201
Banihani Q., Sierra-Alvarez R., Field J., Nitrate and nitrite inhibition of methanogenesis
during denitrification in granular biofilms and digested domestic sludges, Biodegrada-
tion, 2009, vol. 20, p. 801
Barrera E. L., Spanjers H., Solon K., Amerlinck Y., Nopens I., Jo D., Modeling the anaero-
bic digestion of canemolasses vinasse: Extension of the Anaerobic Digestion Model No.
1 (ADM1) with sulfate reduction for a very high strength and sulfate rich wastewater,
Water Research, 2015, vol. 71, p. 42
Barrou O., Karama A., Lakhal E., Bernard O., Pons M., Corriou J., Estimation of a
reduced model of the BSM1 activated sludge wastewater treatment plant., International
Journal of Chemical Reactor Engineering, 2008, vol. 6, p. A63
Batstone D., Jensen P., , 2011 in Wilderer P., ed., , Treatise on Water Science. Elsevier
Oxford pp 615 – 639
Batstone D., Keller J., Industrial applications of the IWA anaerobic digestion model no. 1
(ADM1), Water Science Technology, 2003, vol. 47, p. 199
Batstone D., Keller J., Newel B., Newland M., Modelling anaerobic digestion of complex
wastewater I: Model development, Bioresource Technology, 2000, vol. 75, p. 67
Referências Bibliográficas 125
Batstone D., Pind P., Angelidaki I., Kinetics of thermophilic, anaerobic oxidation of
straight and branched chain butyrate and valerate, Biotechnology and Bioengineering,
2003, vol. 84, p. 195
Batstone D., Tait S., Starrenburg D., Estimation of hydrolysis parameters in full-scale
anerobic digesters, Biotechnology and Bioengineering, 2009, vol. 102, p. 1513
Batstone D. J., Mathematical modelling of anaerobic reactors treating domestic was-

tewater: Rational criteria for model us, Reviews in Environmental Science and
Bio/Technology, 2006a, vol. 5, p. 57
Batstone D. J., Mathematical modelling of anaerobic reactors treating domestic wastewa-

ter: Rational criteria for model use, Reviews in Environmental Science and Bio Techno-
logy, 2006b, vol. 5, p. 57
Batstone D. J., Mathematical modelling of anaerobic reactors treating domestic wastewa-

ter: Rational criteria for model use, Reviews in Environmental Science and Bio Techno-
logy, 2006c, vol. 5, p. 57
Batstone D. J., Keller J., Angelidaki I., Kalyuzhny S., Pavlostathis S., Rozzi A., Sanders
W., Siegrist H., Vavilin V., Anaerobic digestion model No. 1, Scientific and Technical
Report No. 13. IWA Publishing, London, 2002a
Batstone D. J., Keller J., Angelidaki I., Kalyuzhny S. V., Pavlostathis S. G., Rozzi A.,
Sanders W. T. M., Siegrist H., Vavilin V. A., The IWA Anaerobic digestion model no
1. (ADM1), Water Science and Technology, 2002b, vol. 45, p. 65
Batstone D. J., Puyol D., Flores-Alsina X., Rodrı́guez J., Mathematical modelling of ana-
erobic digestion processes: applications and future needs, Reviews In Environmental
Science and Bio/Technology, 2015, vol. 11, p. 1
Bernard O., Bastin G., Identification of reaction networks for bioprocesses: determina-
tion of a partially unknown pseudo-stoichiometric matrix., Bioprocess and Biosystems
Engineering, 2005a, vol. 193, p. 51
Bernard O., Bastin G., On the estimation of the pseudo-stoichiometric matrix for macros-
copic mass balance modelling of biotechnological processes, Mathematical Biosciences,
2005b, vol. 193, p. 51
Bernard O., Chachuat B., Helias A., Rodriguez J., Can we assess the model complexity for
a bioprocess: theory and example of the anaerobic digestion process ?, Water Science
and Technology, 2006, vol. 53, p. 85
Bernard O., Hadj-Sadok Z., Dochain D., Genovesi A., Steyer J., Dynamical model deve-
lopment and parameter identification for an anaerobic wastewater treatment process,
Biotechnology and Bioengineering, 2001, vol. 75, p. 424
Bernet N., Delgenes J., Delgenes J., Moletta R., SBR as a relevant technology to combine
anaerobic digestion and denitrification in a single reactor, Water Science and Technology,
2001, vol. 43, p. 209
Bernet N., Delgenes N., Akunna J., Delgenes J., Moletta R., Combined anaerobic - aerobic
SBR for the treatment of piggery wastewater, Water Research, 2000, vol. 34, p. 611
Bernet N., Habouzit F., Moletta R., Use of an industrial effluent as a carbon source for
denitrification of a high-strength wastewater., Applied Microbiology and Biotechnology,
1996, vol. 46, p. 92
Blumensaat F., Keller J., Modelling of two-stage anaerobic digestion using the IWA Ana-
erobic Digestion Model No. 1 (ADM1), Water Research, 2005, vol. 39, p. 171
Bollon J., Le-Hyaric R., Benbelkacem H., Buffière P., Development of a kinetic model for
anaerobic dry digestion processes: focus on acetate degradation and moisture content,
Biochemical Engineering Journal, 2011, vol. 56, p. 212
Buendı́a I., Fernández F., Villasenor J., Rodrı́guez L., Biodegradability of meat industry
wastes under anaerobic and aerobic conditions, Water research, 2008, vol. 42, p. 67
Campos E., Flotats X., Dynamic simulation of pH in anaerobic processes, Biochemistry

and Biotechnology, 2003, vol. 109, p. 63
Cardoso R., Sierra-Alvarez R. Rowlette P., Flores E., Gómez J., Field J., Sulfide oxida-
tion under chemolithoautotrophic denitrifying conditions., Biotechnology Bioenginee-
ring, 2006, vol. 95, p. 1148
Chachuat B., Roche N., Latifi M., Reduction of the ASM 1 model for optimal control
of small-size activated sludge treatment plants - Réduction du modèle ASM 1 pour la
commande optimale des petites stations d epuration à boues activées, Revue des Sciences
de leau, 2003, vol. 16, p. 5
Chaudhry M., Beg S., Modelling of simultaneous methanogenesis and denitrification in an

upflow packed-bed biofilm reactor, Journal of Chemical Technology and Biotechnology,
1997, vol. 70, p. 267
Chen K., Lin Y., The relationship between denitrifying bacteria and methanogenic bacteria
in a mixed culture system of acclimated sludges., Water Research, 1993, vol. 27, p. 1749
Chen S., Suna D., Chung J.-S., Simultaneous methanogenesis and denitrification of ani-
line wastewater by using anaerobic-aerobic biofilm system with recirculation, Journal of
Hazardous Materials, 2009, vol. 169, p. 575
Chernicharo C., Biological wastewater treatment series. Anaerobic reactors. London, 2007
Chowdhury P., Viraraghavan T., Srinivasan A., Biological treatment processes for fish
processing wastewater: a review, Bioresource Technolology, 2010, vol. 101, p. 439
Christ O., Wilderer P., Angerhofer R., Faulstich M., Mathematical modelling of the hy-
drolysis of anaerobic processes., Water Science and Technology, 2000, vol. 41, p. 61
Chuan C., Aijie W., Nanqi R., Duu-Jong L., Juin-Yih L., High-rate denitrifying sulfide
removal process in expanded granular sludge bed reactor, Bioresource Technology, 2009,
vol. 100, p. 2316
Copp J., Jeppsson U., Rosen C., Towards an ASM1 e ADM1 state variable interface for
plant-wide wastewater treatment modeling. In Proceedings of the 76th Annual WEF
Conference and Exposition (WEFTEC), Los Angeles, USA , 2003
Costello D., Greenfield P., Lee P., Dynamic modeling of a single-stage high rate anaerobic
reactor .1. Model derivation, Water Research, 1991, vol. 25, p. 847
Dalmau J., Comas J., Rodrı́guez-Roda I., Pagilla K., Steyer J.-P., Model development and
simulation for predicting risk of foaming in anaerobic digestion systems, Bioresource
Technology, 2010, vol. 101, p. 4306
Del Pozo R., Diez V., Integrated anaerobic-aerobic fixed-film reactor for slaughterhouse
wastewater treatment, Water Research, 2005, vol. 39, p. 1114
Dereli R. K., Ersahin M. E., Ozgun H., Ozturk I., Aydin A. F., Applicability of Anaerobic
Digestion Model No. 1 (ADM1) for a specific industrial wastewater: Opium alkaloid
effluents, Chemical Engineering Journal, 2010, vol. 165, p. 89
Donoso-Bravo A., Mailier J., Martin C., Rodrı́guez J., Aceves-Lara C. A., Wouwer A. V.,
Model selection, identification and validation in anaerobic digestion: A review, Water
Research, 2011, vol. 45, p. 5347
Eastman J., Ferguson J., Solubilization of Particulate Organic-Carbon During the Acid
Phase of Anaerobic-Digestion., journal water pollution control federation, 1981, vol. 53,
p. 352
Elefsiniotis P., Wareham D. G and S. M., Use of volatile fatty acids from an acid-phase
digester for denitrification, Biotechnology, 2004, vol. 114, p. 289
Esposito G., Frunzo L., Panico A., Pirozzi F., Model calibration and validation for OFMSW
and sewage sludge co-digestion reactors., Waste Management, 2011, vol. 109, p. 2527
Estuardo C., Marti M., Huilinir C., Lillo E., von Bennewitz M., Improvement of nitrate
and nitrite reduction rates prediction, Electronic Journal of Biotechnology, 2008, vol. 11,
p. 1
Fedorovich V., Lens P., Kalyuzhnyi S., Extension of anaerobic digestion model no. 1 with
processes of sulfate reduction, Applied Biochemistry and Biotechnology, 2003, vol. 109,
p. 33
Fezzani B., Cheikh R., Extension of the anaerobic digestion model no. 1 (ADM1) to include
phenolic compounds biodegradation processes for the anaerobic co-digestion of olive mill
wastes at thermophilic temperature, Journal of Hazardous Materials, 2009, vol. 162, p.
1563
Fezzani B., Cheikh R. B., Implementation of IWA anaerobic digestion model No. 1 (ADM
1) for simulating the thermophilic anaerobic co-digestion of olive mill wastewater with
olive mill solid waste in a semi-continuous tubular digester, Chemical Engineering Jour-
nal, 2008, vol. 141, p. 75
Foresti E., Zaiat M., Vallero M., Anaerobic processes as the core technology for sustainable
domestic wastewater treatment: Consolidated applications, new trends, perspectives,
and challenges, Reviews in Environmental Science and Bio/Technology, 2006, vol. 5, p.
3
Fountoulakis M., Stamatelatou K., Batstone D., Lyberatos G., Simulation of DEHP bi-
odegradation and sorption during the anaerobic digestion of secondary sludge, Water
Science and Technology, 2006, vol. 54, p. 119
Galı́ A., Benabdallah T., Astals S., Mata-Alvarez J., Modified version of ADM1 model for
agro-waste application, Bioresource Technology, 2009, vol. 100, p. 2783
Gao D., Peng Y., Wu W., Kinetic model for biological nitrogen removal using shortcut
nitrification-denitrification process in sequencing batch reactor, Environmental Science
Garca-Diéguez C., Bernard O., Roca E., Reducing the Anaerobic Digestion Model No. 1
for its application to an industrial wastewater treatment plant treating winery effluent
wastewater, Bioresource technology, 2013, vol. 132, p. 244
Garibay-Orijel C., Ahring B., Rinderknecht-Seijas N., Poggi-Varaldo H., A simple model
for simultaneous methanogenic-denitrification systems., Journal of Chemical Technology
and Biotechnology, 2006, vol. 81, p. 173
Gavala H. N., Angelidaki I., Ahring B. K., Kinetics and Modeling of Anaerobic Digestion
Process, Advances in Biochemical Engineering Biotechnology, 2003, vol. 81, p. 57
Gernaey K. V., van Loosdrecht M. C., Henze M., Lind M., Jorgensen S. B., Activated
sludge wastewater treatment plant modelling and simulation: state of the art, Environ-
mental Modelling & Software, 2004, vol. 19, p. 763
Ghaniyari-Benis S., Borja R., Bagheri M., Ali Monemian S., Goodarzi V., Tooyserkani Z.,
Effect of adding nitrate on the performance of a multistage biofilter used for anaerobic
treatment of high-strength wastewater, Chemical Engineering Journal, 2010, vol. 156,
p. 250
Giraldo-Gomez E., Goodwin S., Switzenbaum M. S., Influence of mass transfer limitations
on determination of the half saturation constant for hydrogen uptake in a mixed-culture
CH(4)-producing enrichment, Biotechnology and Bioengineering, 1992, vol. 40, p. 768
Girault R and; Bridoux G., Nauleau F., Poullain C., Buffet J., Steyer J., Ag S., Béline
F. A., Waste characterisation procedure for ADM1 implementation based on degradation
kinetics, Water Research, 2012, vol. 46, p. 4626
Gonzalez G., Urrutia H., Roeckel M., Aspe E., Protein hydrolysis under anaerobic, saline
conditions in presence of acetic acid, Journal of Chemical Technology and Biotechnology,
2005, vol. 80, p. 151
Graef S., Andrews J., Stability and Control of Anaerobic Digestion, Journal Water Pollu-
tion Control Federetion, 1974, vol. 46, p. 666
Hao L.-P., Fan L., Pin-Jing H., Lei L., Li-Ming S., Predominant Contribution of Syntrophic
Acetate Oxidation to Thermophilic Methane Formation at High Acetate Concentrations,
Environmental Science and Technology, 2011, vol. 45, p. 508
Hendriksen H. V., Ahring B. K., Integrated removal of nitrate and carbon in an upflow
anaerobic sludge blanket (UASB) reactor: Operating performance, Water Research,
1996, vol. 30, p. 1451
Henze M., Gujer W., Mino T.and van Loosdrecht M., , 2000 Technical report Activated
sludge models ASM1, ASM2, ASM2d and ASM3. Scientific and technical report no. 9
IWA, London.
Henze M., van Loosdrecht M. C. M., Ekama G. A., Biological Wastewater Treatment:
Principles, Modelling and Design. IWA Publishing, London, UK, 2008
Hierholtzer A., Akunna J., Modelling sodium inhibition on the anaerobic digestion process,
Water Science and Technology, 2012, vol. 66, p. 1565
Hinken L., Huber M., Weichgrebe D., Rosenwinkel K., Modified ADM1 for modelling an
UASB reactor laboratory plant treating starch wastewater and synthetic substrate load
tests, Water Research, 2014, vol. 64, p. 82
Huang J.-S., Chou H.-H., Chen C.-M., Chiang C.-M., Effect of recycle-to-influent ratio
on activities of nitrifiers and denitrifiers in a combined UASB activated sludge reactor
system, Chemosphere, 2007, vol. 68, p. 382
Huete E., De Gracia M., Ayesa E., Garcia-Heras J., ADM1 based methodology for the cha-
racterisation of the influent sludge in anaerobic reactors, Water Science and Technology,
2006, vol. 54, p. 157
Huiliñir C., Aspe E., Roeckel M., Model of simultaneous denitrification and methanogenesis
in an upflow packed-bed biofilm reactor: Nitrogen compounds inhibition and pseudo two-
dimensional biofilm model, Journal of Chemical Technology and Biotechnology, 2009,
vol. 84, p. 254
Huiliñir C., Estrell A., Roeckel M., Modeling of the denitrification/anaerobic digestion pro-
cess of salmon fishery wastewater in a biofilm tubular reactor, Journal of Environmental
Management, 2011, vol. 92, p. 1591
Huiliñir C., Hernández S., Aspe E., Roeckel M., Simultaneous nitrate and organic matter
removal from salmon industry wastewater: The effect of C/N ratio, nitrate concentra-
tion and organic load rate on batch and continuous process, Journal of Environmental
Management, 2012, vol. 101, p. 82
Huiliñir C., Montalvo S., Modeling of an anoxic/methanogenic biofilm: effect of pH cal-

culation within the biofilm., Bioprocess and Biosystems Engineering, 2013, vol. 36, p.
1675
Huiliñir C., Roa E., Vargas D., Roeckel M., Aspe E., Kinetics of syntrophic acetogenesis
in a saline medium, Journal of Chemical Technology and Biotechnology, 2008, vol. 83,
p. 1433
Jeong H.-S., Suh C.-W., Lim J.-L., Lee S.-H., Shin H.-S., Analysis and application of
ADM1 for anaerobic methane production, Bioprocess and Biosystems Engineering, 2005,
vol. 27, p. 81
Kalyuzhnyi S., Batch anaerobic digestion of glucose and its mathematical modeling.
II. Description, verification and application of model, Bioresource Technology, 1997a,
vol. 59, p. 249
Kalyuzhnyi S., Batch anaerobic digestion of glucose and its mathematical modeling II. Des-
cription, verification and application of model, Bioresource Technology, 1997b, vol. 59,
p. 249
Kalyuzhnyi S., Fedorovich V., Lens P., Hulshoff Pol L., Lettinga G., Mathematical model-
ling as a tool to study population dynamics between sulfate reducing and methanogenic
bacteria, Biodegradation, 1998, vol. 9, p. 187
Kassab G. N., Combined carbon and nitrogen removal in integrated anaerobic/anoxic

sludge bed reactors for the treatment of domestic sewage, Wageningen University, Wa-
geningen, The Netherlands, 2009, Tese de Doutorado
Kerroum D., Mossab B.-L., Abdssalam Hassen M., Use of ADM1 model to simulate the
anaerobic digestion process used for sludge waste treatment in thermophilic conditions,
Engineering and Environmental Sciences, 2010, vol. 34, p. 121
Khan A. A., Mehrotra I., Kazmi A., Sludge profiling at varied organic loadings and per-
formance evaluation of UASB reactor treating sewage, Biosystems Engineering, 2015,
vol. 131, p. 32
Kim B., Kim S., Shin T., Kim H., Sang B., Comparison of the bacterial communities
in anaerobic, anoxic, and oxic chambers of a pilot A2O process using pyrosequencing
analysis, Current Microbiology, 2013, vol. 6, p. 555
Kleerebezem R., van Loosdrecht M., Critical analysis of some concepts proposed in ADM1,
Kluber H., Conrad R., Inhibitory effects of nitrate, nitrite, NO and N2O on methanoge-
nesis by Methanosarcina barkeri and Methanobacterium bryantii, FEMS Microbiology
Ecology, 1998a, vol. 25, p. 331
Kluber H., Conrad R., Inhibitory effects of nitrate, nitrite, NO, and N2O on methanoge-
nesis by Methanosarcina barkeri and Methanobacterium bryantii, FEMS Microbiology
Ecology, 1998b, vol. 25, p. 331
Kobayashi H., Stenstron M., Mah R., Treatment of low strength wastewater using the
anaerobic filter, Water Research, 1983, vol. 17, p. 903
Lee C., Shin H., Hwang S., Characteristics of granular sludge in a single upflow sludge
blanket reactor treating high levels of nitrate and simple organic compounds, Water
Science Technology, 2004, vol. 50, p. 217
Lee M.-Y., Suh C.-W., Ahn Y.-T., Shin H.-S., Variation of ADM1 by using temperature-
phased anaerobic digestion (TPAD) operation, Bioresource Technology, 2009, vol. 100,
p. 2816
Leitao R., Van Haandel A., Zeeman G., Lettinga G., The effects of operational and envi-
ronmental variations on anaerobic wastewater treatment systems: a review, Bioresource
Lengeler J.W. Drews G., Schlege H., Biology of the prokaryotes. Stuttgart, Germany., 1999
Lubken M., Wichern M., Schlattmann M., Gronauer A., Horn H., Modelling the energy
balance of an anaerobic digester fed with cattle manure and renewable energy crops,
Lyberatos G., Skiadas I., Modelling of Anaerobic Digestion - A review, Global Nest: the
International Journal, 1999, vol. 1, p. 63
Mccarty P. L., The development of anaerobic treatment and its future, Water Science
Madigan M., Martinko J., Parker J., Microbiologia de Brock. São Paulo, 2004
Mairet F., Bernard O., Ras M., Lardon L., Steyer J.-P., Modeling anaerobic digestion of
microalgae using ADM1, Bioresource Technology, 2011, vol. 102, p. 6823
Marazioti C., Kornaros M., Lyberatos G., Kinetic modeling of a mixed culture of Pseudo-
monas denitrificans and Bacillus subtilis under aerobic and anoxic operating conditions.,
Mendes C., Esquerre K., Queiroz L. M., Application of Anaerobic Digestion Model No. 1
for simulating anaerobic mesophilic sludge digestion, Waste Management, 2015, vol. 35,
p. 89
Metcalf Eddy Wastewater engineering: treatment disposal reuse. New York, 2007
Moawad A., Mahmoud U., El-Khateeb M., El-Molla E., Coupling of sequencing batch reac-
tor and UASB reactor for domestic wastewater treatment, Desalination, 2009, vol. 242,
p. 325
Mosey F., Mathematical-Modeling of the Anaerobic-Digestion Process - Regulatory Me-

chanisms for the Formation of Short-Chain Volatile Acids from Glucose, Water Science
Mosquera-Corral A., Sánchez M., Campos J., Méndez R., Lema J., Simultaneous metha-
nogenesis and denitrification of pretreated effluents from a fish canning industry, Water
Research, 2001, vol. 35, p. 411
Mu S., Zeng Y., Wu P., Lou S., Tartakovsky B., Anaerobic digestion model no. 1-based
distributed parameter model of an anaerobic reactor: I. Model development, Bioresource
Mu S. J., Zeng Y., Tartakovsky B., Wu P., Simulation and Control of an Upflow Anaerobic
Sludge Blanket (UASB) Reactor Using an ADM1-Based Distributed Parameter Model,
Industrial and Engineering Chemistry Research, 2007, vol. 46, p. 1519
Muha I., Grillo A., Heisig M., Schönberg M., Linke B., Wittum G., Mathematical modeling
of process liquid flow and acetoclastic methanogenesis under mesophilic conditions in a
two-phase biogas reactor, Bioresource Technology, 2012, vol. 106, p. 1
Nopens I., Batstone D., Copp J., Jeppsson U., Volke E., Alex J., Vanrolleghem P., An
ASM/ADM model interface for dynamic plant-wide simulation, Water Research, 2009,
vol. 43, p. 1913
O’Rourke J., Kinetics of anaerobic treatment at reduced temperatures, Stanford University,

Stanford, CA, USA, 1968, Tese de Doutorado
Orozco A., Bioingenierı́a de aguas residuales: teorı́a y diseño. Acodal, 2005
Palatsi J., Illa J., Prenafeta-Boldú F., Laureni M., Fernandez B., Angelidaki I., Flotats
X., Long-chain fatty acids inhibition and adaptation process in anaerobic thermophilic
digestion: Batch tests, microbial community structure and mathematical modelling,
Bioresource Technology, 2010, vol. 101, p. 2243
Parker W., Wu G., Modifying ADM1 to include formation and emission of odourants,
Parker W. J., Application of the ADM1 model to advanced anaerobic digestion, Bioresource
Pavlostathis S., Giraldo-Gomez E., Kinetics of Anaerobic Treatment - a Critical-Review,

Critical Reviews in Environmental Control, 1991, vol. 21, p. 411
Pavlostathis S. G., Gossett J. M., A kinetic model for anaerobic digestion of biological
sludge, Biotechnology and Bioengineering, 1986, vol. 28, p. 1519
Peiris B., Rathnasiri P., Johansen J., Kuhn A., Bakke R., ADM1 simulation of hydrogen
production, Water Science and Technology, 2006, vol. 53, p. 129
Penumathsa B., Premier G., Kyazze G., Dinsdale R., Guwy A., Esteves S., Rodrı́guez J.,
ADM1 can be applied to continuous bio-hydrogen production using a variable stoichio-
metry approach, Water research, 2008, vol. 42, p. 4379
Philips S., Laanbroek H. J., Verstraete W., Origin, causes and effects of incresead ni-
trite concentrations in aquatic environments, Reviews in Environmental Science &
Bio/Technology, 2002, vol. 1, p. 115
Ramakrishnan A., Gupta S., Effect of hydraulic retention time on thebiodegradation of

complex phenolic mixture from simulated coal wastewater in hybrid UASB reactors,
Journal of Hazardous Materials, 2008, vol. 153, p. 843
Ramirez I., Mottet A., Carrère H., Déléris S., Vedrenne F., Steyer J., Modified ADM1
disintegration/hydrolysis structures for modeling batch thermophilic anaerobic digestion
of thermally pretreated waste activated sludge., Water Research, 2009, vol. 43, p. 3479
Rodrı́guez J., Lema J., Van Loosdrecht M., Kleerebezem R., Variable stoichiometry with
thermodynamic control in ADM1., Water Science and Technology, 2006, vol. 54, p. 101
Rodrı́guez J., Roca E., Lema J., Bernard O., Determination of the adequate minimum
model complexity required in anaerobic bioprocesses using experimental data., Journal
of Chemical Technology and Biotechnology, 2008, vol. 83, p. 1694
Rosen C., Jeppsson U., , 2002 Technical report Anaerobic COST benchmark model des-
cription. Deptament of Industrial Electrical Engineering and Automation (IEA), Lund
University, Lund, Sweden
Rousseau P., Steyer J., Volcke E., Bernet N., Béline F., Combined anaerobic digestion and
biological nitrogen removal for piggery wastewater treatment: a modelling approach,
Roy R., Conrad R., Effect of methanogenic precursors (acetate, hydrogen, propionate)
on the suppression of methane production by nitrate in anoxic rice field soil, FEMS,
Microbial Ecology, 1999, vol. 28, p. 49
Ruiz G., Jeison D., Chamy R., Development of denitrifying and methanogenic activities
in USB reactors for the treatment of wastewater: Effect of COD/N ratio, Process Bio-
chemistry, 2006, vol. 41, p. 1338
Scheid D., Stubner S., Conrad R., Effects of nitrate and sulfate amendment on the metha-
nogenic populations in rice root incubations, FEMS Microbial Ecology, 2003, vol. 43, p.
309
Shapleigh J. P., Denitrifying Prokaryotes. Springer, 2006
Shimada T., Zilles J., Raskin L., Morgenroth E., Carbohydrate storage in anaerobic se-
quencing batch reactors, Water Research, 2007, vol. 41, p. 4721
Siegrist H., Renggli D., Gujer W., Mathematical Modelling of Anaerobic Mesophilic Sewage
Sludge Treatment, Water Science and Technology, 1993, vol. 27, p. 25
Siegrist H., Vogt D., Garcia-Heras J., Gujer W., Mathematical model for meso and ther-
mophilic anaerobic sewage sludge digestion, Environmental Science and Technology,
2002, vol. 36, p. 1113
Soto O., Sanchez O., Aspe E., Roeckel M., Denitrification kinetics of simulated fish pro-
cessing wastewater at different ratios of nitrate to biomass, Biotechnology Letters, 2002,
vol. 24, p. 1173
Souza T., Carvajal A., Donoso-Bravo A., Pena M., Polanco F., ADM1 calibration using
BMP tests for modeling the effect of autohydrolysis pretreatment on the performance of
continuous sludge digesters, Water research, 2013, vol. 47, p. 3244
Steffens M., Lant P., Newell R., A systematic approach for reducing complex biological
wastewater treatment models, Water Research, 1997, vol. 31, p. 590
Tai C., Singh K., Grant S., Combined removal of carbon and nitrogen in an integrated
UASB-Jet loop reactor bioreactor system, Journal of Environmental Engineering-ASCE,
2006, vol. 132, p. 634
Tartakovsky B., Morel E., Steyer J., Guiot S., Application of a variable structure model in
observation and control of an anaerobic digestor, Biotechnology Progress, 2002, vol. 18,
p. 898
Tartakovsky B., Mu S., Zeng Y., Lou S., Guiot S., Wu P., Anaerobic digestion model
No. 1-based distributed parameter model of an anaerobic reactor: II. Model validation,
Tejasen S., Taruyanon K., Modelling of Two-Stage Anaerobic Treating Wastewater from
a Molasses-Based Ethanol Distillery with the IWA Anaerobic Digestion Model No.1,
Engineering Journal, 2010, vol. 14, p. 25
Tugtas A., Tezel U., Pavlostathis S., An extension of the anaerobic digestion model no.1
to include the effect of nitrate reduction processes, Water Science and Technology, 2006,
vol. 54, p. 41
Tugtas A. E., Pavlostathis S., Electron donor effect on nitrate reduction pathway and
kinetics in a mixed methanogenic culture, Biotechnology and Bioengineering, 2007a,
vol. 98, p. 756
Tugtas A. E., Pavlostathis S. G., Inhibitory effects of nitrogen oxides on a mixed metha-
nogenic culture, Biotechnology and Bioengineering, 2007b, vol. 96, p. 444
van Haandel A., Lettinga G., Anaerobic Sewage Treatment. John Wiley and Sons, 1994
Van Haandel A., Van Der Lubbev J., Handbook Biological Waste Water Treatment e
Design and Optimization of Activated Sludge Systems.. Quist Publishing, 2007
Van Haandel A. Marais G., O Comportamento do Sistema de Lodo Ativado: Teoria e

Aplicações para Projetos e Operações. Campina Grande: Epgraf, 142 p., 1999
van Lier J., High-rate anaerobic wastewater treatment: diversifying from end of the pipe
treatment to resource oriented conversion techniques, Chemosphere, 2008, vol. 57, p.
1137
Vanrolleghem P., Rosen C., Zaher U., Copp J., Benedetti L., Ayesa E., Jeppsson U.,
Continuity-based interfacing of models for wastewater systems described by Petersen
matrices, Water Science and Technology, 2005, vol. 52, p. 493
Vavilin V., Rytow S., Lokshina L., Modelling hydrogen partial pressure change as a re-
sult of competition between the butyric and propionic groups of acidogenic bacteria,
Vavilin V., Vasiliev V., Ponomarev A., Rytov S., Simulation model methane as a tool
for effective biogas production during anaerobic conversion of complex organic matter.,
Viraj S., Bruce E. R., Nonsteady-state modeling of multispecies activated sludge process,
Water Environment Research, 2000, vol. 72, p. 554
Von Munch E., Keller J., Lant P., Newell R., Mathematical modelling of prefermenters -
I. Model development and verification, Water Research, 1999, vol. 33, p. 2757
Wang L., Aziz T. N., de los Reyes F. L., Determining the limits of anaerobic co-digestion
of thickened waste activated sludge with grease interceptor waste, Water Research, 2013,
vol. 47, p. 3835
Wang Q., Noguchi C., Kuninobu M., Hara Y., Kakimoto K., Ogawa H., Kato Y., Influ-
ence of hydraulic retention tine on anaerobic digestion of pretreated sludge, Bioresource
Wang R., Li Y., Chen W., Zou J., Chen Y., Phosphate release involving PAOs activity
during anaerobic fermentation of EBPR sludge and the extension of ADM1, Chemical
Engineering Journal, 2015, vol. 26, p. 212
Wawrzynczyk J., Enzymatic treatment of wastewater sludge, sludge solubilisation, im-

provement of anaerobic digestion and extraction of extra cellular polymeric substances,
Lund University, 2007, Tese de Doutorado
Westerholm M., Dolfing J., Angela S., Gray N. D., Head I. M., Schnürer A., Quantification
of syntrophic acetate-oxidizing microbial communities in biogas processes, Environmen-
tal Microbiology Reports, 2011, vol. 3, p. 500
Wett B., Eladawy A., Ogurek M., Description of nitrogen incorporation and release in
ADM1, Water Science and Technology, 2006, vol. 54, p. 67
Wichern M., Gehring T., Fischer K., Andrade D., Lubken M., Koch K., Gronauer A., Horn
H., Monofermentation of grass silage under mesophilic conditions: measurements and
mathematical modeling with ADM 1, Bioresource Technology, 2009, vol. 100, p. 1675
Winkler M.-K., Ettwig K., Vannecke T., Stultiens K., Bogdan A., Kartal B., Volcke E.,
Modelling simultaneous anaerobic methane and ammonium removal in a granular sludge
reactor, Water Research, 2015, vol. 73, p. 323
Xie L., Chen J., Wang R., Zhou Q., Effect of carbon source and COD/NO3-N ratio on
anaerobic simultaneous denitrification and methanogenesis for high-strength wastewater
treatment, Journal of Bioscience and Bioengineering, 2012, vol. 113, p. 759
Yasui H., Goel R., Li Y., Noike T., Modified ADM1 structure for modelling municipal
primary sludge hydrolysis, Water Research, 2008, vol. 42, p. 249
Yingyu A., Fenglin Y., Fook S. W., Hwee C. C., Effect of Recirculation Ratio on Simul-
taneous Methanogenesis and Nitrogen Removal Using a Combined Up-Flow Anaerobic
Sludge Blanket Membrane Bioreactor, Environmental Engineering Science, 2009, vol. 26,
p. 1047
Yingyu A., Fenglin Y., Hwee C. C., Fook S. W., Bing W., The integration of methano-
genesis with shortcut nitrification and denitrification in a combined UASB with MBR,
Zaher U., Buffiere P., Steyer J., Chen S., A procedure to estimate proximate analysis of
mixed organic wastes, Water Environment Research, 2009, vol. 81, p. 407
Zaher U., Grau P., Benedetti L., Ayesa E., Vanrolleghem P., Transformers for interfa-
cing anaerobic digestion models to pre- and post-treatment processes in a plant-wide
modelling context, Environmental Modelling and Software, 2007, vol. 22, p. 40
Zumft W., Cell biology and molecular basis of denitrification, Microbiology and Molecular
Biology Reviews, 1997, vol. 61, p. 533

Silva (2016)

Enviado por

Direitos autorais:

Formatos disponíveis

Silva (2016)

Enviado por

Dados do documento

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Silva (2016)

Enviado por

Direitos autorais:

Formatos disponíveis

UNIVERSIDADE FEDERAL DA BAHIA

CARLOS EDUARDO PEREIRA MENDES DA SILVA

MODELAGEM MATEMÁTICA E SIMULAÇÃO DA

MODELAGEM MATEMÁTICA E SIMULAÇÃO DA

Tese apresentada ao programa de Pós-Graduação

Orientadores: Profa. Dra..Karla Patrícia S. O. R. Esquerre;

Tese (doutorado) – Universidade Federal da Bahia. Escola

1. Modelagem matemática. 2. Desnitrificação. 3. Carbono -

O doutorado possibilitou o aprofundamento do conhecimento adquirido na graduação em Engenha-

Aos colegas do Programa de Pós-Graduação em Engenharia Industrial da Escola Politécnica -

A Universidade Federal da Bahia, por fornecer infraestrutura para o desenvolvimento de pesquisas,

Bı́blia Sagrada - Provérbios 8:12

“Ó profundidade da riqueza, da sabedoria e do conhecimento de Deus! Quão insondáveis

Bı́blia Sagrada - Romanos 11:33

realizadas simulações para avaliar a RSCN em um reator anaeróbio-anóxico operado em

2.1 Sequência de processos metabólicos durante a digestão anaeróbia. Fonte:

3.1 Estimation procedure to identify most sensitive parameters. . . . . . . . . 74

4.1 Conceptual model of the simultaneous AD/DN processes of domestic was-

5.1 Conceptual model of the simultaneous AD/DN processes of salmon fishery

2.1 Reações do nitrogênio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3.1 Composition of the wastewater sludge . . . . . . . . . . . . . . . . . . . . . 71

4.1 Disintegration and hydrolysis kinetic constants . . . . . . . . . . . . . . . . 85

5.1 Denitrification process included in the modified Anaerobic Digestion Model

ACP Análise de Componentes Principais

Csu Conteúdo de carbono presente nos monossacarı́deos [kmolC

Principais publicações do autor durante o doutoramento.

Publicações associadas à tese

3. Application of Anaerobic Digestion Model No. 1 for simulating anae-

4. Modeling simultaneous carbon and nitrogen removal in an anaerobic-

5. Modeling of simultaneous carbon and nitrogen removal in an anaerobic-

6. Conclusões e sugestões para trabalhos futuros 119

1.1 Motivação e justificativa

ver a remoção de matéria orgânica e de nutrientes é composto de três estágios, anóxico,

influência de variáveis de processo tais como: variação da vazão volumétrica na corrente

trogênio de águas residuárias concentradas e diluı́das, ainda é incipiente e pouco explorada

1.2 Objetivos da tese

O objetivo principal desta pesquisa é avaliar a remoção simultânea de carbono e ni-

• Aplicar o ADM1 para simular o comportamento dinâmico de um reator anaeróbio

• Modelar e simular os processos simultâneos de fermentação, metanogênese e des-

• Modelar e simular os processos simultâneos de fermentação, metanogênese e des-

1.3 Estrutura da tese

é adotada a convenção de utilização do ponto como separador decimal, ao invés do uso da

operado em regime contı́nuo.

2.1 Processos biológicos de digestão anaeróbia

A digestão anaeróbia em reatores estrutura um ecossistema no qual micro-organismos

2.1.1 Fermentação e metanogênese

O processo de conversão da matéria orgânica em ambientes anaeróbios ocorre por meio

Para à ocorrência dos processos metabólicos de fermentação e metanogênese é necessária

os micro-organismos acetogênicos e metanogênicos. A maioria das bactérias fermentativas

CH3 COOO− + H2 O → CH4 + HCO3− ∆G◦ = −31, 0 kJ (2.1)

2.1.2 Redução de nitrato

a desnitrificação heterotrófica e autotrófica, respectivamente (desnitrificação) (Madigan

N O3− → N O2− → N O(g) → N2 O(g) → N2 (g) (2.3)

foram determinadas baseadas nos princı́pios bioenergéticos (Mccarty, 2001), a saber:

• Desnitrificação com glicose

∆G0 = −113, 3 kJ/eeq

• Desnitrificação com propionato

∆G0 = −99, 83 kJ/eeq

• Desnitrificação com acetato

∆G0 = −99, 60 kJ/eeq