In-situ diagnostics of the packed bed in post-combustion CO₂ capture using electrical capacitance tomography

Abstract

Climate change is commonly acknowledged as a result of increased greenhouse gas levels in the atmosphere. The excess use of fossil fuels and consequential emission of greenhouse gases is causing global warming. Global warming or climate change is a global issue that requires concerted efforts on a planetary scale to reduce the amount of carbon accumulated into the atmosphere. Post-combustion capture with amine solvents is the most matured technology and plays an essential role in addressing the issue of CO₂ emission. Flooding phenomenon is an operation problem exists in post-combustion process, which reduces CO₂ capture efficiency and causes potential damage to the equipment. This process must be avoided to ensure continued carbon dioxide absorption. However, knowledge of the flow regime transition process involved in flooding is currently limited due to a lack of an in situ diagnostic technique for process monitoring and a lack of comprehensive datasets which reflect the practical carbon capture process. Furthermore, the nature of interface of air and liquid during flooding onset is not well known and according to correlations for flooding point and packed column capacity estimation, which have uncertainties of up to ±30%. The inability to accurately operate the carbon capture process could lead to a decrease in CO₂ capture efficiency, safety issues, oversizing of the packed column and increased CO₂ capture costs.

The work described in this thesis proposes to diagnose packed columns in the post-combustion capture process through electrical capacitance tomography (ECT) as a real-time qualitative and quantitative imaging method. Comprehensive packed column models have been built for each post-combustion carbon capture system, using COMSOL Multiphysics with MATLAB as the simulation software. To ensure an objective simulation of the performance of ECT in different post-combustion carbon capture systems, the representation of packed column geometry structure and material electrical permittivity have been verified and calibrated based on real-scenario test conditions. For each carbon capture system, many parametric simulations were performed to evaluate the effect on tomographic imaging. The correctional images of post-combustion carbon capture packed column were reconstructed from the simulation data and the feasibility of ECT for packed column diagnosis was evaluated.

To further validate the feasibility of ECT, a laboratory-scale test campaign on two different types of post-combustion test rigs were conducted to investigate the effect of different packed column process variables on tomographic images. The column constitutes a transparent pipe filled with packing material, in which is easy to observe the flooding onset and flow regime transition. The liquid distribution and holdup are monitored through ECT, which allows variations in the predominant characteristics of flooding events to be investigated in greater detail than in previous research. Furthermore, combined with Convolutional Neural Networks (CNN), ECT enables a high degree of accuracy with only ±1% error on liquid holdup calculation and greater robustness in environments affected by flooding with strong turbulence flow. In addition, ECT technology has attracted great interest in the field of cryogenic carbon capture, which captures CO2 by cryogenically desubliming CO₂ out of the flue gas as CO₂ frost on the cold surfaces of the heat exchangers. The technique was found to provide half the cost and energy of the state-of-the-art carbon capture methods. However, this process could lead to several operational problems, such as plugging. ECT could help a process engineer to understand the process of frost formation and provide key tomographic images and process parameters to facilitate cryogenic carbon capture implementation. The successful development of an ECT system would result in an in situ, agile, non-intrusive and low-cost monitoring tool for the requirements of packed column diagnosis.