For decades, fossil fuels have been the backbone of global energy systems. While they have supported industrial progress, their environmental impact—especially in terms of greenhouse gas emissions—has driven an urgent shift toward renewable energy. This transition is further motivated by geopolitical instability and the volatility of fossil fuel markets. Renewable sources like solar and wind are essential to a low-carbon future, but their inherent intermittency creates challenges for grid stability and energy supply. To overcome this, flexible and efficient energy storage solutions are needed. One promising approach is Power-to-Gas (PtG), which converts surplus renewable electricity into hydrogen via water electrolysis. This hydrogen can then be transformed into methane through biological methanation. Among various PtG strategies, ex-situ biological methanation has gained increasing attention. In this approach, hydrogenotrophic methanogens—a group of archaea—convert H2 and CO2 into CH4 under mild conditions, without the need for harsh catalysts or extreme operating conditions. Unlike in-situ systems, ex-situ reactors are physically separated from anaerobic digesters, allowing better control over microbial activity, gas flow, and operational parameters. The resulting biomethane is fully compatible with existing natural gas infrastructure. It can be injected into the grid, stored, or used as fuel for heat, electricity, and transport— making it a powerful bridge between the electricity and gas sectors. Moreover, this process not only utilizes excess renewable hydrogen but also recycles CO2, contributing to a circular carbon economy. Overall, ex-situ biological methanation offers a practical, scalable, and infrastructureready solution to support the transition toward a resilient and carbon-neutral energy system. This doctoral thesis focuses on experimental ex-situ biomethanation conducted in pilotscale trickle bed reactors. The experiments were carried out under both atmospheric and pressurized conditions, and the performance of the reactors was evaluated across a range of operating parameters. To gain deeper insight into the complex hydrodynamics and mass transfer phenomena occurring within the reactor, computational fluid dynamics (CFD) simulations were performed. These simulations allowed for the visualization and quantification of gas–liquid flow patterns and dispersion behavior, which are otherwise difficult to measure experimentally. In parallel, machine learning techniques were applied to analyze the experimental data and to develop predictive models of key performance indicators such as methane production rate, hydrogen transfer efficiency, and outlet gas composition. These data-driven models offered an efficient way to identify influential variables, optimize operating conditions, and support decision-making in the design and scaling of ex-situ biomethanation systems. The combined experimental, computational, and data-analytical approach provides a comprehensive understanding of reactor behavior and contributes to the development of more efficient and scalable biological power-to-gas technologies. Chapter 3 presents the literature review and outlines the scientific background relevant to ex-situ biomethanation. It highlights the current state of research, identifies existing knowledge gaps, and defines the objectives and scope of ex-situ biomethanation using hydrogenotrophic methanogens. Special emphasis is placed on different reactor configurations and mass transfer principles, with a particular focus on trickle bed reactors (TBRs). Chapters 4 and 5 investigate the experiments conducted in pilot-scale TBRs operated under atmospheric (R1) and pressurized (R2) conditions, respectively. In addition, the influence of sulfur—both its absence and presence—on methane purity and overall reactor performance was evaluated. The operational performance and the effects of various parameters were closely monitored over 849 days for R1 and 512 days for R2. Gas retention time (GRT) was identified as a key factor for achieving high methane purity in the outlet gas. The experiments demonstrated that high methane purity levels could be achieved, reaching up to 97.83% in R1 and 98.77% in R2 in the absence of sulfur, and up to 99.18% in R1 and 100% in R2 in the presence of sulfur. Furthermore, statistical analysis confirmed that the presence of sulfur not only enhanced methane purity and hydrogen transfer in both reactors but also improved process stability over the entire operational period. Chapter 6 investigates the key parameters influencing methane production rate (MPR), methane purity, and H2 transfer through the application of various machine learning models. Among the nine machine learning regression models evaluated, Random Forest (RF) and XGBoost demonstrated the highest predictive performance for both reactors. Gas retention time (GRT) and the H2/CO2 ratio were identified as the most influential features governing the key reactor parameters. A ML-based inverse design approach was subsequently employed to determine the optimal operating conditions, enabling efficient guidance for reactor operation. Chapter 7 presents the computational fluid dynamics (CFD) modeling of the TBR, providing insights into the assessment of key sources of uncertainty within the reactor. Interfacial momentum transfer and wetting efficiency correlations were analyzed as primary contributors to uncertainty. Water velocity, volume fraction, and water flux were used as metrics for this evaluation. The results indicate that the modified interfacial momentum transfer correlations provide a more robust prediction of flow behavior in the TBR. Additionally, the empirical wetting efficiency correlations closely reproduce the studied parameters, demonstrating their reliability for modeling the reactor.
Per decenni, i combustibili fossili hanno rappresentato la spina dorsale dei sistemi energetici globali. Pur avendo sostenuto il progresso industriale, il loro impatto ambientale— soprattutto in termini di emissioni di gas serra—ha spinto verso una transizione urgente verso le energie rinnovabili. Questo passaggio è ulteriormente motivato dall’instabilità geopolitica e dalla volatilità dei mercati dei combustibili fossili. Fonti rinnovabili come il solare e l’eolico sono essenziali per un futuro a basse emissioni di carbonio, ma la loro intrinseca intermittenza crea sfide per la stabilità della rete e l’approvvigionamento energetico. Per superare queste limitazioni, sono necessarie soluzioni di accumulo energetico flessibili ed efficienti. Un approccio promettente è il Power-to-Gas (PtG), che converte l’elettricità rinnovabile in eccesso in idrogeno tramite elettrolisi dell’acqua. Questo idrogeno può poi essere trasformato in metano attraverso la metanazione biologica. Tra le varie strategie PtG, la metanazione biologica ex-situ ha suscitato un interesse crescente. In questo approccio, i metanogeni idrogenotrofici—un gruppo di archei—convertono H2 e CO2 in CH4 in condizioni miti, senza necessità di catalizzatori aggressivi o condizioni operative estreme. Diversamente dai sistemi in-situ, i reattori ex-situ sono fisicamente separati dai digestori anaerobici, consentendo un controllo migliore sull’attività microbica, sul flusso dei gas e sui parametri operativi. Il biometano ottenuto è completamente compatibile con le infrastrutture del gas naturale esistenti. Può essere immesso nella rete, stoccato o utilizzato come combustibile per calore, elettricità e trasporti, rappresentando un ponte efficace tra i settori elettrico e del gas. Inoltre, questo processo non solo sfrutta l’idrogeno rinnovabile in eccesso, ma ricicla anche il CO2, contribuendo a un’economia circolare del carbonio. Complessivamente, la metanazione biologica ex-situ offre una soluzione pratica, scalabile e compatibile con le infrastrutture esistenti per supportare la transizione verso un sistema energetico resiliente e a zero emissioni di carbonio. Questa tesi dottorale si concentra sulla biomethanation ex-situ sperimentale condotta in reattori a letto gocciolante (TBR) a scala pilota. Gli esperimenti sono stati eseguiti sia in condizioni atmosferiche sia pressurizzate, e le prestazioni dei reattori sono state valutate su un ampio intervallo di parametri operativi. Per ottenere una comprensione più approfondita dei complessi fenomeni di idrodinamica e trasferimento di massa all’interno del reattore, sono state condotte simulazioni di fluidodinamica computazionale (CFD). Queste simulazioni hanno permesso di visualizzare e quantificare i modelli di flusso gas-liquido e il comportamento della dispersione, altrimenti difficili da misurare sperimentalmente. Parallelamente, sono state applicate tecniche di machine learning per analizzare i dati sperimentali e sviluppare modelli predittivi di indicatori chiave di prestazione, come il tasso di produzione di metano, l’efficienza di trasferimento dell’idrogeno e la composizione del gas in uscita. Questi modelli basati sui dati hanno offerto un metodo efficiente per identificare le variabili più influenti, ottimizzare le condizioni operative e supportare la progettazione e la scala dei sistemi di biomethanation ex-situ. L’approccio combinato sperimentale, computazionale e analitico fornisce una comprensione completa del comportamento del reattore e contribuisce allo sviluppo di tecnologie biologiche Power-to-Gas più efficienti e scalabili. Il Capitolo 3 presenta la revisione della letteratura e delinea il background scientifico pertinente alla biomethanation ex-situ. Viene evidenziato lo stato dell’arte della ricerca, identificate le lacune conoscitive esistenti e definiti gli obiettivi e l’ambito della biomethanation ex-situ mediante metanogeni idrogenotrofici. Particolare attenzione è rivolta alle diverse configurazioni dei reattori e ai principi di trasferimento di massa, con un focus specifico sui reattori a letto gocciolante (TBR). I Capitoli 4 e 5 descrivono gli esperimenti condotti nei TBR pilota operati rispettivamente in condizioni atmosferiche (R1) e pressurizzate (R2). Inoltre, è stata valutata l’influenza dello zolfo—sia in assenza sia in presenza—sulla purezza del metano e sulle prestazioni complessive del reattore. Le prestazioni operative e gli effetti dei vari parametri sono stati monitorati attentamente per 849 giorni per R1 e 512 giorni per R2. Il tempo di ritenzione del gas (GRT) è stato identificato come un fattore chiave per ottenere elevate percentuali di metano nel gas in uscita. Gli esperimenti hanno dimostrato che è possibile raggiungere alti livelli di purezza del metano, fino al 97,83% in R1 e al 98,77% in R2 in assenza di zolfo, e fino al 99,18% in R1 e al 100% in R2 in presenza di zolfo. Inoltre, l’analisi statistica ha confermato che la presenza di zolfo non solo migliora la purezza del metano e il trasferimento di idrogeno in entrambi i reattori, ma aumenta anche la stabilità del processo durante l’intero periodo operativo. Il Capitolo 6 analizza i principali parametri che influenzano il tasso di produzione di metano (MPR), la purezza del metano e il trasferimento di H2 attraverso l’applicazione di diversi modelli di machine learning. Tra i nove modelli di regressione machine learning valutati, Random Forest (RF) e XGBoost hanno mostrato le migliori prestazioni predittive per entrambi i reattori. Il tempo di ritenzione del gas (GRT) e il rapporto H2/CO2 sono stati identificati come le caratteristiche più influenti nel controllo dei principali parametri del reattore. Successivamente, è stato impiegato un approccio di progettazione inversa basato su ML per determinare le condizioni operative ottimali, fornendo così indicazioni efficaci per la gestione del reattore. Capitolo 7 presenta la modellazione della TBR mediante fluidodinamica computazionale (CFD), fornendo approfondimenti sulla valutazione delle principali fonti di incertezza nel reattore. Le correlazioni per il trasferimento di quantità di moto interfaciale e l’efficienza di bagnatura sono state analizzate come principali fattori di incertezza. Per questa valutazione sono stati utilizzati la velocità dell’acqua, la frazione volumica e il flusso d’acqua. I risultati indicano che le correlazioni modificate per il trasferimento di quantità di moto interfaciale forniscono una previsione più robusta del comportamento del flusso nella TBR. Inoltre, le correlazioni empiriche di efficienza di bagnatura riproducono accuratamente i parametri studiati, dimostrando la loro affidabilità per la modellazione del reattore.
Optimization of biological reduction of CO2 to methane in ex-situ reactors for power-to-gas conversion
Changizi, Maryam
2024/2025
Abstract
For decades, fossil fuels have been the backbone of global energy systems. While they have supported industrial progress, their environmental impact—especially in terms of greenhouse gas emissions—has driven an urgent shift toward renewable energy. This transition is further motivated by geopolitical instability and the volatility of fossil fuel markets. Renewable sources like solar and wind are essential to a low-carbon future, but their inherent intermittency creates challenges for grid stability and energy supply. To overcome this, flexible and efficient energy storage solutions are needed. One promising approach is Power-to-Gas (PtG), which converts surplus renewable electricity into hydrogen via water electrolysis. This hydrogen can then be transformed into methane through biological methanation. Among various PtG strategies, ex-situ biological methanation has gained increasing attention. In this approach, hydrogenotrophic methanogens—a group of archaea—convert H2 and CO2 into CH4 under mild conditions, without the need for harsh catalysts or extreme operating conditions. Unlike in-situ systems, ex-situ reactors are physically separated from anaerobic digesters, allowing better control over microbial activity, gas flow, and operational parameters. The resulting biomethane is fully compatible with existing natural gas infrastructure. It can be injected into the grid, stored, or used as fuel for heat, electricity, and transport— making it a powerful bridge between the electricity and gas sectors. Moreover, this process not only utilizes excess renewable hydrogen but also recycles CO2, contributing to a circular carbon economy. Overall, ex-situ biological methanation offers a practical, scalable, and infrastructureready solution to support the transition toward a resilient and carbon-neutral energy system. This doctoral thesis focuses on experimental ex-situ biomethanation conducted in pilotscale trickle bed reactors. The experiments were carried out under both atmospheric and pressurized conditions, and the performance of the reactors was evaluated across a range of operating parameters. To gain deeper insight into the complex hydrodynamics and mass transfer phenomena occurring within the reactor, computational fluid dynamics (CFD) simulations were performed. These simulations allowed for the visualization and quantification of gas–liquid flow patterns and dispersion behavior, which are otherwise difficult to measure experimentally. In parallel, machine learning techniques were applied to analyze the experimental data and to develop predictive models of key performance indicators such as methane production rate, hydrogen transfer efficiency, and outlet gas composition. These data-driven models offered an efficient way to identify influential variables, optimize operating conditions, and support decision-making in the design and scaling of ex-situ biomethanation systems. The combined experimental, computational, and data-analytical approach provides a comprehensive understanding of reactor behavior and contributes to the development of more efficient and scalable biological power-to-gas technologies. Chapter 3 presents the literature review and outlines the scientific background relevant to ex-situ biomethanation. It highlights the current state of research, identifies existing knowledge gaps, and defines the objectives and scope of ex-situ biomethanation using hydrogenotrophic methanogens. Special emphasis is placed on different reactor configurations and mass transfer principles, with a particular focus on trickle bed reactors (TBRs). Chapters 4 and 5 investigate the experiments conducted in pilot-scale TBRs operated under atmospheric (R1) and pressurized (R2) conditions, respectively. In addition, the influence of sulfur—both its absence and presence—on methane purity and overall reactor performance was evaluated. The operational performance and the effects of various parameters were closely monitored over 849 days for R1 and 512 days for R2. Gas retention time (GRT) was identified as a key factor for achieving high methane purity in the outlet gas. The experiments demonstrated that high methane purity levels could be achieved, reaching up to 97.83% in R1 and 98.77% in R2 in the absence of sulfur, and up to 99.18% in R1 and 100% in R2 in the presence of sulfur. Furthermore, statistical analysis confirmed that the presence of sulfur not only enhanced methane purity and hydrogen transfer in both reactors but also improved process stability over the entire operational period. Chapter 6 investigates the key parameters influencing methane production rate (MPR), methane purity, and H2 transfer through the application of various machine learning models. Among the nine machine learning regression models evaluated, Random Forest (RF) and XGBoost demonstrated the highest predictive performance for both reactors. Gas retention time (GRT) and the H2/CO2 ratio were identified as the most influential features governing the key reactor parameters. A ML-based inverse design approach was subsequently employed to determine the optimal operating conditions, enabling efficient guidance for reactor operation. Chapter 7 presents the computational fluid dynamics (CFD) modeling of the TBR, providing insights into the assessment of key sources of uncertainty within the reactor. Interfacial momentum transfer and wetting efficiency correlations were analyzed as primary contributors to uncertainty. Water velocity, volume fraction, and water flux were used as metrics for this evaluation. The results indicate that the modified interfacial momentum transfer correlations provide a more robust prediction of flow behavior in the TBR. Additionally, the empirical wetting efficiency correlations closely reproduce the studied parameters, demonstrating their reliability for modeling the reactor.| File | Dimensione | Formato | |
|---|---|---|---|
|
2025_10_Changizi.pdf
non accessibile
Dimensione
6.02 MB
Formato
Adobe PDF
|
6.02 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/244837