(647a) Practically Achievable Process Limits for a Temperature Swing Adsorption Process for CO2 Capture from Ngcc Flue Gas

Climate change is being tackled in many ways. Most prominent pathway includes a combination of replacing intensive fuels such as coal with natural gas and carbon capture technology. Owing to the dilute nature of flue gas emitted from natural gas fired plants (NGCC), absorption-based capture has found to be energy intensive . As an alternative, temperature swing adsorption (TSA) processes are found to be a widely acknowledged separation technology associated with dilute fuel sources. Scientists have also been involved in development of appropriate materials such as zeolites, carbons, and metal-organic frameworks (MOFs) for effective capture of CO₂. Notably, metal-organic frameworks (MOFs), that have long eluded large-scale applications have been commercialized for industrial CO₂ capture (Lin, 2021.). Conventional selection of adsorbents was based on process metrics evaluated from equilibrium measurements. Recent studies have clearly articulated the need to combine process design and optimization techniques to effectively evaluate the true potential for various adsorbents (Pai, 2021). Zeroing in on suitable material-process combination can be obtained via screening or by process inversion. This work addresses the limits that can be achieved for maximum possible performance indicators such as productivity (Pr), recovery (Re), purity (Pu), and minimum possible energy (En) for a 3-step TSA process separating CO₂ from NGCC flue gas. The pure component CO₂ and N₂ adsorption isotherms were described using single site Langmuir equations. The competition was described using extended Langmuirian model. A 3-step steam assisted TSA process was considered. Several case studies were designed and evaluated for the key performance indicators Pu, Re, En, and Pr. The case studies were curated aptly to recognize process limits as well as adsorbent limits. To overcome the cumbersome task of a conventional process optimization, a Machine assisted Adsorption Process Learner and Emulator (MAPLE) framework was employed (Pai, 2021). This technique trains surrogate models, based on artificial neural networks (ANN), to predict large-scale process performance for a given adsorbent. The ANN models were tweaked appropriately with training-testing ratio, number of neurons, number of hidden layers, and activation functions to emulate the detailed model with proximity. For each of the case studies that were considered, a large design space of MOFS and process conditions were searched with a genetic algorithm whose fitness was scored based on the MAPLE process models. Results of this study will be discussed at the meeting.

Keywords:, carbon capture, temperature swing adsorption, MAPLE, process inversion

References

Pai, K.N. et.al. (2021). Practically Achievable Process Performance Limits for Pressure- Vacuum Swing Adsorption-Based Postcombustion CO2 Capture. ACS Sustainable Chemistry & Engineering, 3838-3849.

Lin, J. et.al. (2021.). A scalable metal-organic framework as a durable physisorbent for carbon dioxide capture. Science, 1464-1469.