• Citer

Predictive modeling of warm plasmas: from understanding chemistry to optimizing reactor design for H2 synthesis

par Maerivoet, Stein
Président du jury Perreault, Patrice
Promoteur Reniers, François ;Bogaerts, Annemie
Publication Non publié, 2026-09-04

Thèse de doctorat

• Accès en ligne
• Détails
• Contenu
• Statistiques

Résumé :

The industrial transition towards H₂ as both a reactant and an energy carrier has led to a rapidly increasing demand. Although H₂ is often described as a carbon-free fuel, this is only valid when it is produced without associated CO₂ emissions. Currently, H₂ is primarily generated via steam methane reforming, which results in approximately 13 kg of CO₂ emissions per kg of H₂ produced. Consequently, alternative production pathways must be explored to achieve truly carbon-free H₂.Plasma ((partially) ionized gas) has emerged as a promising medium for gas-phase chemistry. Its high-energy environment enables reaction pathways that are inaccessible under conventional conditions. However, warm plasmas typically reach temperatures of several thousand Kelvin, making some experimental diagnostics challenging.As a result, our understanding of plasma processes relies heavily on modeling. By incorporating the relevant physical and chemical phenomena into advanced numerical simulations, plasma behavior can be studied in detail, revealing the mechanisms that drive desired reactions. When sufficiently accurate, such models also enable the optimization of reactor conditions and design, leading to improved conversion efficiencies, reduced energy consumption, and enhanced product selectivity.This work aims to develop plasma models from first principles, without reliance on experimental input, and able to reproduce observed experimental results. These predictive models provide insight into complex plasma chemistry, and they support experimental research by offering guidance, not only on reaction pathways but also on gas flow dynamics and heat transfer.A comprehensive introduction to the underlying concepts is provided in Chapter 1, where a clear distinction is made between global models, which focus on detailed plasma kinetics, and multidimensional models, which additionally incorporate gas flow and heat transfer. Chapter 2 presents the governing equations used throughout this work for both modeling approaches.In Chapter 3, the first multidimensional plasma model with an extensive chemistry set is introduced. The results demonstrate significant differences compared to global models of the same system, highlighting the importance of spatial effects in warm plasmas. While this chapter represents an important step towards predictive multidimensional modeling, further refinement is required to improve agreement with experimental observations.Global modeling is employed in Chapter 4 to explore optimal operating conditions for plasma-based NH₃ conversion. Continuous warm plasma operation is found to yield the lowest energy cost, in agreement with experimental data. Additionally, the model predicts that preheating the feed gas reduces energy consumption in pulsed plasma systems and identifies plasma-assisted NH₃ cracking as a promising direction for future experimental investigation.Multidimensional models in Chapters 5 and 7 achieve improved agreement with experiments for warm NH₃ and CH₄ plasmas, respectively. These models provide insight into multidimensional physical effects, such as the detrimental impact of turbulent transport in NH₃ plasmas, and they allow initial recommendations for reactor design improvements.In Chapter 6, a two-zone global model is developed to study nanosecond pulsed CH₄ plasmas. By self-consistently calculating gas temperatures, the model reduces the need for user-defined inputs, while maintaining good agreement with experimental results. This approach enables detailed analysis of chemical pathways in both the plasma region and the surrounding colder gas, highlighting preheating as a strategy for reducing energy costs in CH₄ reforming.In conclusion, this thesis advances predictive plasma modeling in both global and multidimensional frameworks. These tools pave the way for the computational design and optimization of plasma reactors prior to experimental implementation, significantly reducing development time and cost.