This thesis explores the modeling and analysis of hydrogen-air combustion in a double-swirl coaxial injector configuration, representative of modern low-emission combustion systems. LES simulations were carried out using a Dynamic Thickened Flame (DTF) combustion model implemented within the OpenFOAM framework. The study focused on two flame configurations observed in the HYdrogen LOw NOx (HYLON) injector: attached and lifted. The model performance was evaluated against experimental and numerical data available in the literature, with particular attention to capturing the flame topology, stabilization behavior, and thermochemical dynamics of the reactions. In addition, post-processing based on the Computational Singular Perturbation (CSP) method was used to investigate the dominant reaction processes and identify the species that kinetically control the process. Studies were also carried out on the influence of analytical and numerical formulations of the Jacobian matrix for chemistry in terms of accuracy and computational cost. The simulations successfully reproduced the overall flame dynamics in the attached configuration and the stabilization mechanisms observed experimentally, confirming the suitability of the DTF approach for modeling hydrogen combustion. However, discrepancies due to incorrect local flame thickening induced by the DTFLES model revealed limitations in the algorithm, indicating the need for a correction based on the Takeno index to improve physical consistency. The CSP analysis provided new insights into the link between turbulence and chemical kinetics, supporting future modeling efforts of NOx formation. Overall, this work advances both the numerical modeling framework and the fundamental understanding of hydrogen combustion processes.
Questa tesi esplora la modellazione e l’analisi della combustione idrogeno-aria in una configurazione di iniettore coassiale a doppio swirl, rappresentativa dei moderni sistemi di combustione a basse emissioni. Sono state condotte simulazioni LES utilizzando un modello di combustione Dynamic Thickened Flame (DTF) implementato all’interno del framework OpenFOAM. Lo studio si è concentrato su due configurazioni di fiamma osservate nell’iniettore HYdrogen LOw NOx (HYLON): attaccata e sollevata. Le prestazioni del modello sono state valutate rispetto ai dati sperimentali e numerici forniti in letteratura, con particolare attenzione alla cattura della topologia della fiamma, al comportamento di stabilizzazione e alla dinamica termochimica delle reazioni. Inoltre, è stato utilizzato un post-processing basato sul metodo Computational Singular Perturbation (CSP) per indagare i processi di reazione dominanti e identificare le specie che controllano cineticamente il processo. Sono stati anche effettuati studi sull’influenza delle formulazioni analitiche e numeriche della matrice Jacobiana per la chimica in termini di accuratezza e costo computazionale. Le simulazioni hanno riprodotto con successo la dinamica generale della fiamma nella configurazione attaccata e i meccanismi di stabilizzazione osservati sperimentalmente, confermando l’idoneità dell’approccio DTF per la modellazione della combustione dell’idrogeno. Tuttavia, discrepanze dovute ad un inspessimento locale errato della fiamma, indotto dal modello DTFLES, hanno rivelato limitazioni nell’algoritmo, indicando la necessità di una correzione basata sull’indice di Takeno per migliorare la coerenza fisica. L’analisi CSP ha fornito nuove intuizioni sul legame tra turbolenza e cinetica chimica, supportando futuri sforzi di modellazione della formazione di NOx . Nel complesso, questo lavoro fa progredire sia il framework di modellazione numerica sia la comprensione fondamentale dei processi di combustione dell’idrogeno.
LES and CSP-based analysis of hydrogen combustion kinetics in swirling flows
PIOMBONI, LUCA
2024/2025
Abstract
This thesis explores the modeling and analysis of hydrogen-air combustion in a double-swirl coaxial injector configuration, representative of modern low-emission combustion systems. LES simulations were carried out using a Dynamic Thickened Flame (DTF) combustion model implemented within the OpenFOAM framework. The study focused on two flame configurations observed in the HYdrogen LOw NOx (HYLON) injector: attached and lifted. The model performance was evaluated against experimental and numerical data available in the literature, with particular attention to capturing the flame topology, stabilization behavior, and thermochemical dynamics of the reactions. In addition, post-processing based on the Computational Singular Perturbation (CSP) method was used to investigate the dominant reaction processes and identify the species that kinetically control the process. Studies were also carried out on the influence of analytical and numerical formulations of the Jacobian matrix for chemistry in terms of accuracy and computational cost. The simulations successfully reproduced the overall flame dynamics in the attached configuration and the stabilization mechanisms observed experimentally, confirming the suitability of the DTF approach for modeling hydrogen combustion. However, discrepancies due to incorrect local flame thickening induced by the DTFLES model revealed limitations in the algorithm, indicating the need for a correction based on the Takeno index to improve physical consistency. The CSP analysis provided new insights into the link between turbulence and chemical kinetics, supporting future modeling efforts of NOx formation. Overall, this work advances both the numerical modeling framework and the fundamental understanding of hydrogen combustion processes.| File | Dimensione | Formato | |
|---|---|---|---|
|
2025_12_Piomboni_Tesi.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Tesi
Dimensione
18.64 MB
Formato
Adobe PDF
|
18.64 MB | Adobe PDF | Visualizza/Apri |
|
2025_12_Piomboni_Executive_Summary.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Executive Summary
Dimensione
2.08 MB
Formato
Adobe PDF
|
2.08 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/10589/246177