TU Darmstadt / ULB / TUbiblio

Modelling Biofilm Systems for Wastewater Treatment: Impact of Microscale Features on Global Modelling Results

Acevedo Alonso, Vanessa (2022)
Modelling Biofilm Systems for Wastewater Treatment: Impact of Microscale Features on Global Modelling Results.
Technische Universität Darmstadt
doi: 10.26083/tuprints-00021670
Ph.D. Thesis, Primary publication, Publisher's Version

Abstract

Biofilm models are effective tools that allow the mathematical description of biofilm systems,the prediction of their removal performance as well as their conceptual exploration. Models have the advantage of being less resource and time consuming than laboratory experiments and being more flexible regarding the scenarios that can be analyzed. Due to the high complexity of biofilms, it is virtually impossible to develop a model that comprises all the phenomena occurring within the biofilm. Simplifications are considered a substantial part of the modelling process, and even the most comprehensive models developed until now make use of assumptions and simplifications. Nevertheless, the vast majority of models fulfill their purpose and are useful to both researches and practitioners (Wanner et al. 2006). The decision on which model is better for which modelling task relies on the data available, the level of understanding of the phenomena occurring in the system and the objective of the modelling task. In the specific case of biofilms, a trade-off between microscale features and global modelling results is often present. A more detailed description of the microscale features does not necessarily lead to more compelling modelling results, and occasionally it can compromise the identifiability and the determination of relevant parameters. However, neglecting microscale features can result in inaccurate representations of the system and thus the explanatory power of the model may be diminished. Depending on the system and its specific conditions, the modeler is presented with a dilemma: Which microscale features are worth including, which simplifications can be afforded and which simplifications are compulsory due to the lack of information. Two microscale features are of special interest in this work: microbial community composition and dimensionality. Three biofilm systems were used to explore the impact that these two microscale features have on the global modelling results and on the model’s explanatory power: A biologically active Granular Activated Carbon (bGAC) filter, a Moving Bed Bioreactor (MBBR) and a Membrane Aerated Biofilm Reactor (MABR). An individual publication is dedicated to each one of these systems. These systems were used to showcase alternative modelling approaches and to illustrate the effect of choosing simple or more complex descriptions of the microscale features of interest. The first and second publication, P1 and P2 respectively, focus on the microbial community composition. Publication P1 deals with DOC removal from WWTP effluents in a bGAC-filter. Within the filter, DOC is removed by simultaneous adsorption and biodegradation, therefore a suitable model should include both mechanisms. It proposes a model that integrates a traditional one-dimensional biofilm model with the ideal adsorbed solution theory that can be applied within the activated sludge model framework. A simplified microbial community composed solely composed of aerobic heterotrophic bacteria is selected. The developed model is able to describe the DOC breakthrough curves at different empty bed contact times and it also shows the relative contribution of biodegradation and adsorption to the total DOC removal. Publication P2 analyzes the behavior of heterotrophic bacteria in an MBBR reactor operating as a Partial Nitritation/ Anammox (PN/A) system. It discusses the growth strategies that hetrotrophic bacteria pursue when facing substrate scarcity. The effect of the yield-strategy is analyzed in two scenarios. In the first scenario a group of heterotrophic bacteria growing on endogenous COD is investigated, whereas in the second scenario, external COD is allowed into the system and a second group of heterotrophs (rate strategist) growing on the more available external COD is added. The competition between both groups over space, and electron acceptor is assessed. In both scenarios the pursuing of the yield strategy seems to be crucial for the diversity of the heterotrophic community. Effluent concentrations as well as heterotrophic produced dinitrogen gas is strongly affected by the growth-strategy that the heterotrophs selected, higher denitrification activities are observed when the yield strategy is selected. Finally dimensionality is the microscale feature of interest in the third publication P3. An MABR reactor used for PN/A is modelled. Publication P3 compares a traditional one-dimensional model and different pseudo two-dimension models that allow the implementation of concentration gradients in the bulk liquid and the gas phase, individually and simultaneously as well as in counter or parallel flow. The results show that the one-dimensional model underestimates the effluent’s total dissolved nitrogen concentration in comparison to the prediction delivered by the pseudo two-dimensional models. Differences in the axial gradients in the biofilm are also observed between the two evaluated modelling approaches. P3 also demonstrates that the concentration gradients in the gas phase have a more significant impact on the modelling results than the concentration gradients in the bulk liquid. The importance of the microscale features: microbial composition and dimensionality is explored. Alongside, the implications on the global modelling results, product of the simplifications of microscale features are investigated. In the case of microbial composition, although more information has been made available due to the new experimental techniques (molecularbiology, imagining etc.) there are still disparities between what can be determined experimentally and how this can be implemented into the existent modelling frameworks. In the case of dimensionality, longitudinal gradients seem to be more influential than it was previously assumed and need to be taken into account to better describe MABRs. The adequate level of complexity required for a microscale feature and in general for a model should be decided based on the modelling goals, the current understanding of the system and the available data.

Item Type: Ph.D. Thesis
Erschienen: 2022
Creators: Acevedo Alonso, Vanessa
Type of entry: Primary publication
Title: Modelling Biofilm Systems for Wastewater Treatment: Impact of Microscale Features on Global Modelling Results
Language: English
Referees: Lackner, Prof. Dr. Susanne ; Morgenroth, Prof. Dr. Eberhard
Date: 2022
Place of Publication: Darmstadt
Collation: 127 Seiten
Refereed: 24 June 2022
DOI: 10.26083/tuprints-00021670
URL / URN: https://tuprints.ulb.tu-darmstadt.de/21670
Abstract:

Biofilm models are effective tools that allow the mathematical description of biofilm systems,the prediction of their removal performance as well as their conceptual exploration. Models have the advantage of being less resource and time consuming than laboratory experiments and being more flexible regarding the scenarios that can be analyzed. Due to the high complexity of biofilms, it is virtually impossible to develop a model that comprises all the phenomena occurring within the biofilm. Simplifications are considered a substantial part of the modelling process, and even the most comprehensive models developed until now make use of assumptions and simplifications. Nevertheless, the vast majority of models fulfill their purpose and are useful to both researches and practitioners (Wanner et al. 2006). The decision on which model is better for which modelling task relies on the data available, the level of understanding of the phenomena occurring in the system and the objective of the modelling task. In the specific case of biofilms, a trade-off between microscale features and global modelling results is often present. A more detailed description of the microscale features does not necessarily lead to more compelling modelling results, and occasionally it can compromise the identifiability and the determination of relevant parameters. However, neglecting microscale features can result in inaccurate representations of the system and thus the explanatory power of the model may be diminished. Depending on the system and its specific conditions, the modeler is presented with a dilemma: Which microscale features are worth including, which simplifications can be afforded and which simplifications are compulsory due to the lack of information. Two microscale features are of special interest in this work: microbial community composition and dimensionality. Three biofilm systems were used to explore the impact that these two microscale features have on the global modelling results and on the model’s explanatory power: A biologically active Granular Activated Carbon (bGAC) filter, a Moving Bed Bioreactor (MBBR) and a Membrane Aerated Biofilm Reactor (MABR). An individual publication is dedicated to each one of these systems. These systems were used to showcase alternative modelling approaches and to illustrate the effect of choosing simple or more complex descriptions of the microscale features of interest. The first and second publication, P1 and P2 respectively, focus on the microbial community composition. Publication P1 deals with DOC removal from WWTP effluents in a bGAC-filter. Within the filter, DOC is removed by simultaneous adsorption and biodegradation, therefore a suitable model should include both mechanisms. It proposes a model that integrates a traditional one-dimensional biofilm model with the ideal adsorbed solution theory that can be applied within the activated sludge model framework. A simplified microbial community composed solely composed of aerobic heterotrophic bacteria is selected. The developed model is able to describe the DOC breakthrough curves at different empty bed contact times and it also shows the relative contribution of biodegradation and adsorption to the total DOC removal. Publication P2 analyzes the behavior of heterotrophic bacteria in an MBBR reactor operating as a Partial Nitritation/ Anammox (PN/A) system. It discusses the growth strategies that hetrotrophic bacteria pursue when facing substrate scarcity. The effect of the yield-strategy is analyzed in two scenarios. In the first scenario a group of heterotrophic bacteria growing on endogenous COD is investigated, whereas in the second scenario, external COD is allowed into the system and a second group of heterotrophs (rate strategist) growing on the more available external COD is added. The competition between both groups over space, and electron acceptor is assessed. In both scenarios the pursuing of the yield strategy seems to be crucial for the diversity of the heterotrophic community. Effluent concentrations as well as heterotrophic produced dinitrogen gas is strongly affected by the growth-strategy that the heterotrophs selected, higher denitrification activities are observed when the yield strategy is selected. Finally dimensionality is the microscale feature of interest in the third publication P3. An MABR reactor used for PN/A is modelled. Publication P3 compares a traditional one-dimensional model and different pseudo two-dimension models that allow the implementation of concentration gradients in the bulk liquid and the gas phase, individually and simultaneously as well as in counter or parallel flow. The results show that the one-dimensional model underestimates the effluent’s total dissolved nitrogen concentration in comparison to the prediction delivered by the pseudo two-dimensional models. Differences in the axial gradients in the biofilm are also observed between the two evaluated modelling approaches. P3 also demonstrates that the concentration gradients in the gas phase have a more significant impact on the modelling results than the concentration gradients in the bulk liquid. The importance of the microscale features: microbial composition and dimensionality is explored. Alongside, the implications on the global modelling results, product of the simplifications of microscale features are investigated. In the case of microbial composition, although more information has been made available due to the new experimental techniques (molecularbiology, imagining etc.) there are still disparities between what can be determined experimentally and how this can be implemented into the existent modelling frameworks. In the case of dimensionality, longitudinal gradients seem to be more influential than it was previously assumed and need to be taken into account to better describe MABRs. The adequate level of complexity required for a microscale feature and in general for a model should be decided based on the modelling goals, the current understanding of the system and the available data.

Alternative Abstract:
Alternative abstract Language

Biofilmmodelle sind wirksame Instrumente, die die mathematische Beschreibung von Biofilmsystemen, die Vorhersage ihrer Abbauleistung sowie die konzeptionelle Erforschung dieser Systeme ermöglichen. Modelle haben den Vorteil, dass sie weniger ressourcen- und zeitaufwändig sind als Laborexperimente und dass sie hinsichtlich der zu analysierenden Szenarien flexibler sind. Aufgrund der hohen Komplexität von Biofilmen ist es praktisch unmöglich, ein Modell zu entwickeln, das alle im Biofilm vorkommenden Phänomene umfasst. Vereinfachungen sind ein wesentlicher Bestandteil des Modellierungsprozesses. Selbst die umfassendsten Modelle, die bisher entwickelt wurden, treffen zahlreiche Annahmen. Dennoch erfüllt die überwiegende Mehrheit der Modelle ihren Zweck und so bleiben sie sowohl für die Forschung als auch für die Praxis nützlich (Wanner et al. 2006). Die Entscheidung, welches Modell für welche Modellierungsaufgabe besser geeignet ist, hängt von den verfügbaren Daten, dem Grad des Verständnisses der im System auftretenden Phänomene und dem Ziel der Modellierungsaufgabe ab. Im speziellen Fall von Biofilmen besteht häufig ein Kompromiss zwischen microscale features und globalen Modellierungsergebnissen. Eine detailliertere Beschreibung der microscale features führt nicht notwendigerweise zu verbesserten Modellierungsergebnissen und kann gelegentlich die Identifizierbarkeit des Biofilms und die Bestimmung der relevanten Parameter beeinträchtigen. Die Vernachlässigung von microscale features kann jedoch zu ungenauen Darstellungen des Systems führen und damit die Aussagekraft des Modells verringern. Je nach System und seinen spezifischen Bedingungen steht der Modellierer vor einem Dilemma: Welche microscale features sind es wert, einbezogen zu werden, welche Vereinfachungen kann man sich leisten und welche sind aufgrund des Mangels an Informationen zwingend erforderlich. Zwei microscale features sind in dieser Arbeit von besonderem Interesse: die Zusammensetzung der mikrobiellen Gemeinschaft und die Dimensionalität. Drei Biofilmsysteme wurden verwendet, um die Auswirkungen dieser beiden microscale features auf die globalen Modellierungsergebnisse und die Aussagekraft des Modells zu untersuchen: Eine biologisch aktiver granulierter Aktivkohle (bGAK) Filter, ein Moving Bed Bioreactor (MBBR) und ein Membrane Aerated Biofilm Reactor(MABR). Jedem dieser Systeme ist eine eigene Veröffentlichung gewidmet. Diese Systeme wurden verwendet, um unterschiedliche Modellierungsansätze vorzustellen und die Auswirkungen der Wahl einfacher oder komplexerer Beschreibungen der interessierenden micro features zu veranschaulichen. Die erste und zweite Veröffentlichung, P1 bzw. P2, konzentrierten sich auf die Zusammensetzung der mikrobiellen Gemeinschaft. Die Veröffentlichung P1 befasst sich mit der Entfernung von DOC aus Kläranlagenablauf in einem bGAK-Filter. Innerhalb des Filters wird DOC durch gleichzeitige Adsorption und biologischen Abbau entfernt, weshalb ein geeignetes Modell beide Mechanismen berücksichtigen sollte. Ein Modell wurde entwickelt, das ein traditionelles eindimensionales Biofilmmodell mit der ideal adsorbed solution Theorie verbindet, die im Rahmen der activated sludge models angewendet werden kann. Es wurde eine vereinfachte mikrobielle Gemeinschaft ausgewählt, die ausschließlich aus aeroben heterotrophen Bakterien besteht. Das entwickelte Modell ist in der Lage, die DOC-Durchbruchskurven bei unterschiedlichen Leerbett-Kontaktzeiten zu beschreiben und zeigt auch den relativen Beitrag von biologischem Abbau und Adsorption zur gesamten DOC-Entfernung. Publikation P2 analysiert das Verhalten heterotropher Bakterien in einem MBBR-Reaktor, der als partielle Nitritation/Anammox (PN/A) betrieben wird. Sie erforscht die Wachstumsstrategien, die heterotrophe Bakterien bei Substratknappheit verfolgen. Die Auswirkung der yield strategy wird in zwei Szenarien analysiert. Im ersten Szenario wurde eine Gruppe heterotropher Bakterien untersucht, die auf endogenem CSB wächst, während im zweiten Szenario den System externer CSB zugeführt wurde und eine zweite Gruppe heterotropher Bakterien (rate strategists) hinzugefügt wurde, die auf dem besser verfügbaren externen CSB wächst. Der Wettbewerb zwischen beiden Gruppen um Platz und Elektronenakzeptoren wurde untersucht. In beiden Szenarien schien die Verfolgung der yield strategy ausschlaggebend für die Diversität der heterotrophen Gemeinschaft zu sein. Die Abwasserkonzentrationen sowie der von den Heterotrophen produzierte gasförmige Stickstoff wurden stark von der von den Heterotrophen gewählten Wachstumsstrategie beeinflusst, wobei höhere Denitrifikationsaktivitäten beobachtet wurden, wenn die yield strategy gewählt wurde. In der dritten Veröffentlichung P3 war die Dimensionalität das microcale feature von Interesse. Es wurde ein MABR-Reaktor modelliert, der für PN/A verwendet wird. In der Veröffentlichung P3 werden ein traditionelles eindimensionales Modell und verschiedene pseudo-zweidimensionale Modelle verglichen, die die Implementierung von Konzentrationsgradienten in der Flüssigkeits- und der Gasphase ermöglichen, und zwar sowohl einzeln und gleichzeitig als auch im Gegen- oder Parallelstrom. Die Ergebnisse zeigen, dass das eindimensionale Modell die Gesamtkonzentration des gelösten Stickstoffs im Abwasser im Vergleich zu den Vorhersagen der pseudo-zweidimensionalen Modelle unterschätzt. Auch bei den axialen Gradienten im Biofilm wurden Unterschiede zwischen den beiden evaluierten Modellierungsansätzen festgestellt. P3 zeigte auch, dass die Konzentrationsgradienten in der Gasphase einen stärkeren Einfluss auf die Modellierungsergebnisse hatten als die Konzentrationsgradienten in der Flüssigkeitsphase. Die Bedeutung der microscale features: mikrobielle Zusammensetzung und Dimensionalität wird untersucht. Außerdem werden die Auswirkungen auf die globalen Modellierungsergebnisse untersucht, die sich aus den Vereinfachungen der microscale features ergeben. Im Fall der mikrobiellen Zusammensetzung stehen mehr Informationen zu Verfügung, dank der neuen experimentellen Techniken (Molekularbiologie, Imaging usw.), dennoch kann nicht alles, was experimentell bestimmt werden kann, direkt in herkömmliche Modelle implementiert werden. Was die Dimensionalität betrifft, haben Konzentration Gradienten in der Fließrichtung einen größeren Einfluss als bisher angenommen und müssen berücksichtigt werden, um MABRs besser zu beschreiben. Das angemessene Maß an Komplexität, das für ein microscale feature und allgemein für ein Modell erforderlich ist, sollte auf der Grundlage der Modellierungsziele, des aktuellen Verständnisses des Systems und der verfügbaren Daten entschieden werden.

German
Status: Publisher's Version
URN: urn:nbn:de:tuda-tuprints-216704
Classification DDC: 600 Technology, medicine, applied sciences > 620 Engineering and machine engineering
Divisions: 13 Department of Civil and Environmental Engineering Sciences
13 Department of Civil and Environmental Engineering Sciences > Institute IWAR
13 Department of Civil and Environmental Engineering Sciences > Institute IWAR > Wastewater Engineering
Date Deposited: 01 Nov 2022 12:36
Last Modified: 16 Dec 2022 08:33
PPN: 502426292
Referees: Lackner, Prof. Dr. Susanne ; Morgenroth, Prof. Dr. Eberhard
Refereed / Verteidigung / mdl. Prüfung: 24 June 2022
Export:
Suche nach Titel in: TUfind oder in Google
Send an inquiry Send an inquiry

Options (only for editors)
Show editorial Details Show editorial Details