Conference Agenda

Enhanced weathering (EW) of minerals has recently been recognised as a promising strategy for gigaton-level large-scale carbon dioxide removal. However, prior to the practical application of EW-based CO2 capture, significant acceleration is necessary by optimising the local environment where solid, liquid, and gas phases interact. This optimisation is crucial as the drive for more efficient CO2 capture technologies continues to escalate, underscored by the increasing need for operational flexibility to adapt to substantial variations in the flow rate and concentration of CO2-rich flue gases. Additionally, integrating renewable energy sources into CO2 capture strategies enhances environmental benefits; however, the inherently intermittent power output due to meteorological and seasonal variations poses a considerable challenge.

The efficacy of chemical reactors to enhance the mass transport and reaction rates in the EW-based CO2 capture process has not yet been thoroughly evaluated. In response, our study in the past a few years conducted detailed mechanistic modelling, validated rigorously through experimental data, focusing on three distinct reactor configurations designed specifically for EW-based CO2 capture: trickle bed reactors, packed bubble column reactors, and stirred slurry reactors. We selected CO2 capture rate, energy consumption, and water consumption as three pivotal performance indicators under a range of operating conditions. Given the complex computational demands of these mechanistic models, we developed advanced machine learning-based models and optimisation approaches, including Response Surface Methodology (RSM), Support Vector Regression (SVR), and hybrid algorithms. These methodologies facilitated the rigorous optimisation of each reactor type to minimise competing objectives effectively, allowing for the generation and comparative analysis of Pareto fronts for the three types of reactors. Our findings illustrate that the mass transport of CO2 into the aqueous phase, considered as the rate-limiting step, substantially influences the capture performance. Notably, when substituting calcite with forsterite, the process's controlling mechanism transitions from gas/liquid mass transfer to solid dissolution due to forsterite's lower dissolution rate.

Furthermore, our research extended into the dynamic adaptive of the CO2 capture process within a renewable energy framework, where CO2 emitted from power plant flue gases is converted into bicarbonate and subsequently stored in the ocean. We developed data-driven surrogate dynamic models to accurately predict the CO2 capture rate and energy consumption of a selected reactor. Utilising a multi-objective NSGA-II genetic algorithm, we pre-emptively optimised reactor conditions based on forecasts of inlet flue gas CO2 concentration and available wind energy. This proactive approach aimed to maximise the carbon capture rate while minimising the reliance on non-renewable energy sources. The adaptive multi-objective optimisation framework demonstrated a significant improvement, increasing the CO2 capture rate by 16.7% and reducing the use of non-renewable energy by 36.3%.

The methodologies we have developed and refined throughout this study hold considerable potential for widespread application across the industry, thereby advancing progress toward achieving UN Sustainability Goals 7 (Affordable and Clean Energy), 9 (Industry, Innovation, and Infrastructure), 12 (Responsible Consumption and Production), and 13 (Climate Action). This research not only provides a foundation for future innovations in CO2 capture technologies but also serves as a vital stepping stone toward sustainable industrial practices.

Power-to-X (PtX) is one of the most promising solutions for long-term storage of renewable energy such as wind and solar power (Miehling et al., 2022). In particular, Power-to-Ammonia (PtA) has attracted significant interest due to its ability to store and recover energy without carbon dioxide emissions. The major challenge for Power-to-Ammonia (PtA) lies in managing the highly variable power supply from renewable sources, as the Haber-Bosch process is designed for stable operation. The fluctuating nature of renewable energy requires the development of new flexible operating strategies across a wide operating envelope from 10 % to 120 % of the nominal load (Armijo & Philibert, 2020). Rosbo et al. (2023) introduced a dynamic model for an adiabatic quench-cooled reactor (AQCR). We investigated open-loop transients, optimal operating points, and basic control strategies across a broad operating window. Following this, we extended our model library to include adiabatic indirect cooled (AICR) and direct-cooled reactors (IDCR), evaluating the stability and conversion performance of these reactors for a PtA process (Rosbo et al., 2024).

In this work, we expand to a plantwide model of the ammonia synthesis loop, incorporating catalytic beds, heat exchangers, compressors, steam turbines, and flash separators. We define a function for the total electrical power utility of the PtA plant composed of electrolysers, air separation, compressors, separator cooling, and steam turbines. The power function is adopted as the objective function for optimising the PtA plant across its operating envelope. Operating constraints include maximum reactor temperatures, compressor choke and stall, minimum steam temperature, and maximum loop pressure. Six degrees of freedom are considered for optimisation: three reactor temperatures, N₂/H₂ feed ratio, separator temperature, and loop pressure. We perform the optimisation by minimising the power function for a given hydrogen make-up feed flow. Across the operating envelope, the optimisation results reveal different active constraint regions. We assess the sensitivity of the optimisation results for relevant process disturbances such as feed argon content, catalyst activity, cooling water temperature, and hydrogen production cost. Additionally, we evaluate the loss in the objective function associated with fixing individual optimisation variables across the operating envelope. This identifies potential self-optimising variables that result in minimal loss when controlled to a constant value across the operating range.

References
Armijo, J., & Philibert, C. (2020). Flexible production of green hydrogen and ammonia from variable solar and wind energy: Case study of Chile and Argentina. International Journal of Hydrogen Energy, 45(3), 1541–1558.

Miehling, S., Fendt, S., & Spliethoff, H. (2022). Optimal integration of Power-to-X plants in a future European energy system and the resulting dynamic requirements. Energy Conversion and Management, 251(July 2021),

Rosbo, J. W., Jensen, A. D., Jørgensen, J. B., & Huusom, J. K. (2024). Comparison, operation and cooling design of three general reactor types for Power-to-Ammonia processes. Chemical Engineering Journal, 496, 153660.

Rosbo, J. W., Ritschel, T. K. S., Hørstholt, S., Huusom, J. K., & Jørgensen, J. B. (2023). Flexible operation, optimisation and stabilising control of a quench cooled ammonia reactor for Power-to-Ammonia. Computers & Chemical Engineering, 176(108316).