Abstract: This study investigates the dry reformation of methane (DRM) over Ni/Al2O3 catalysts in a dielectric barrier discharge (DBD) non-thermal plasma reactor. A novel hybrid machine learning (ML) model is developed to optimize the plasma-catalytic DRM reaction with limited experimental data. To address the non-linear and complex nature of the plasma-catalytic DRM process, the hybrid ML model integrates three well-established algorithms: regression trees, support vector regression, and artificial neural networks. A genetic algorithm (GA) is then used to optimize the hyperparameters of each algorithm within the hybrid ML model. The ML model achieved excellent agreement with the experimental data, demonstrating its efficacy in accurately predicting and optimizing the DRM process. The model was subsequently used to investigate the impact of various operating parameters on the plasma-catalytic DRM performance. We found that the optimal discharge power (20 W), CO2/CH4 molar ratio (1.5), and Ni loading (7.8 wt%) resulted in the maximum energy yield at a total flow rate of
51 mL/min. Furthermore, we investigated the relative significance of each operating parameter on the performance of the plasmacatalytic DRM process. The results show that the total flow rate had the greatest influence on the conversion, with a significance exceeding 35% for each output, while the Ni loading had the least impact on the overall reaction performance. This hybrid model demonstrates a remarkable ability to extract valuable insights from limited datasets, enabling the development and optimization of more efficient and selective plasma-catalytic chemical processes.

Abstract: Electrification and carbon capture technologies are essential for achieving net-zero emissions in the chemical sector. A crucial strategy involves converting captured CO2 into CO, a valuable chemical feedstock. This study evaluates the feasibility of two innovative methods: plasma activation and electrolysis, using clean electricity and captured CO2. Specifically, it compares a gliding arc plasma reactor with an embedded novel carbon bed system to a modern zero-gap type low-temperature electrolyser. The plasma method stood out with an energy cost of 19.5 GJ per tonne CO, marking a 43% reduction compared to electrolysis and conventional methods. CO production costs for plasma- and electrolysis-based plants were $671 and $962 per tonne, respectively. However, due to high uncertainty regarding electrolyser costs, the CO production costs in electrolysis-based plants may actually range from $570 to $1392 per tonne. The carbon bed system in the plasma method was a key factor in facilitating additional CO generation from O-2 and enhancing CO2 conversion, contributing to its cost-effectiveness. Challenges for electrolysis included high costs of equipment and low current densities. Addressing these limitations could significantly decrease production costs, but challenges arise from the mutual relationship between intrinsic parameters, such as CO2 conversion, CO2 input flow, or energy cost. In a future scenario with affordable feedstocks and equipment, costs could drop below $500 per tonne for both methods. While this may be more challenging for electrolysis due to complexity and expensive catalysts, plasma-based CO production appears more viable and competitive.

Abstract: Despite the recent promising potential of plasma-based nitrogen fixation, the technology faces significant challenges in efficient upscaling. To tackle this challenge, we investigate two reactors, i.e., a small one, operating in a flow rate range of 5-20 ln min-1 and current range of 200-500 mA, and a larger one, operating at higher flow rate (100-300 ln min-1) and current (400-1000 mA). Both reactors operate in a pin-to-pin configuration and are powered by direct current (DC) from the same power supply unit, to allow easy comparison and evaluate the effect of upscaling. In the small reactor, we achieve the lowest energy cost (EC) of 2.8 MJ mol-1, for a NOx concentration of 1.72%, at a flow rate of 20 ln min-1, yielding a production rate (PR) of 33 g h-1. These values are obtained in air; in oxygen-enriched air, the results are typically better, at the cost of producing oxygen-enriched air. In the large reactor, the higher flow rates reduce the NOx concentration due to lower SEI, while maintaining a similar EC. This stresses the important effect of the geometrical configuration of the arc, which is typically concentrated in the center of the reactor, resulting in limited coverage of the reacting gas flow, and this is identified as the limiting factor for upscaling. However, our experiments reveal that by changing the reactor configuration, and thus the plasma geometry and power deposition mechanisms, the amount of gas treated by the plasma can be enhanced, leading to successful upscaling. To obtain more insights in our experiments, we performed thermodynamic equilibrium calculations. First of all, they show that our measured lowest EC closely aligns with the calculated minimum thermodynamic equilibrium at atmospheric pressure. In addition, they reveal that the limited NOx production in the large reactor results from the contracted nature of the plasma. To solve this limitation, we let the large reactor operate in so-called torch configuration. Indeed, the latter enhances the NOx concentrations compared to the pin-to-pin configuration, yielding a PR of 80 g h-1 at an EC of 2.9 MJ mol-1 and NOx concentration of 0.31%. This illustrates the importance of reactor design in upscaling. With the focus on feasibility evaluation of scaling-up plasma-based nitrogen fixation by combined experiments and thermodynamic modelling, we aim to tackle the challenge of design and development of an energy-efficient and scaled-up plasma reactor.

Abstract: Multi-objective optimization is an important decision-making tool for energy processes, as multiple targets need to be achieved. These objectives are usually conflicting since a single solution cannot be optimal for all objectives, resulting in a set of Pareto-optimal solutions. Multiple indicators might be available to describe a sustainability objective, such as the environmental impact which is commonly evaluated by performing a life cycle assessment. In this study, Pareto aggregation is proposed as a method which employs a novel multi-objective optimization-based approach as an alternative to the classically used aggregation in life cycle assessment. This method identifies conflicting environmental indicators and performs an aggregation among those that require a trade-off. An environmental-economic optimization of a second-generation bioethanol plant is used to illustrate and evaluate the proposed method. Process parameters from a biochemical conversion pathway flowsheet simulation model are chosen as optimization variables. To reduce the computational time, surrogate models, based on artificial neural networks, are used. Out of the eighteen ReCiPe Midpoint environmental indicators, five were identified as conflicting, resulting in an aggregated environmental objective, which was then traded off with the economic objective function, chosen as the levelized cost of ethanol. Comparison with the widely used single-score EcoIndicator99 showed that the Pareto aggregation method can reduce most of the environmental indicators by up to 6.5%. This research provides an insight on non-redundant objective functions, aiming at reducing the dimensionality of multi-objective optimization problems, while taking into consideration decision-makers’ preferences.

Abstract: This study introduces solid-state tuning of a mesostructured cellular foam (MCF) to enhance hydrogen (H-2) storage in clathrate hydrates. Grafting of promoter-like molecules (e.g., tetrahydrofuran) at the internal surface of the MCF resulted in a substantial improvement in the kinetics of formation of binary H-2-THF clathrate hydrate. Identification of the confined hydrate as sII clathrate hydrate and enclathration of H-2 in its small cages was performed using XRD and high-pressure H-1 NMR spectroscopy respectively. Experimental findings show that modified MCF materials exhibit a similar to 1.3 times higher H-2 storage capacity as compared to non-modified MCF under the same conditions (7 MPa, 265 K, 100% pore volume saturation with a 5.56 mol% THF solution). The enhancement in H-2 storage is attributed to the hydrophobicity originating from grafting organic molecules onto pristine MCF, thereby influencing water interactions and fostering an environment conducive to H-2 enclathration. Gas uptake curves indicate an optimal tuning point for higher H-2 storage, favoring a lower density of carbon per nm(2). Furthermore, a direct correlation emerges between higher driving forces and increased H-2 storage capacity, culminating at 0.52 wt% (46.77 mmoles of H-2 per mole of H2O and 39.78% water-to-hydrate conversions) at 262 K for the modified MCF material with fewer carbons per nm(2). Notably, the substantial H-2 storage capacity achieved without energy-intensive processes underscores solid-state tuning's potential for H-2 storage in the synthesized hydrates. This study evaluated two distinct kinetic models to describe hydrate growth in MCF. The multistage kinetic model showed better predictive capabilities for experimental data and maintained a low average absolute deviation. This research provides valuable insights into augmenting H-2 storage capabilities and holds promising implications for future advancements.