Artificial intelligence applications for the carbon dioxide capture process.

摘要

In this thesis, recently developed artificial intelligence (AI)-based systems for carbon dioxide capture (CO2) data management and analysis have been reported. Research on post combustion CO2 capture has been ongoing over the past two decades with the heightened interest in CO2 capture. The post-combustion capture process requires monitoring and analysis of a vast amount of data, as well as an understanding of the intricate relationships among the parameters involved in CO2 capture. This study applied artificial intelligence (AI) with software engineering technologies to assist industrial researchers and operators to understand and improve the operation of the CO2 capture process at the International Test Centre for Carbon Dioxide Capture (ITC). During the research, a data analysis decision support system for the CO2 capture process was developed for managing the data generated from the complex CO2 absorption process and automating the data filtering process. An enhanced version of the decision support system, which is a platform-independent web-based system that can overcome the prior system's limitations, is also presented in this thesis. In addition, the thesis describes a process of data modeling and analysis for unraveling the nature of the relationships between key parameters using artificial neural networks and sensitivity analysis. In the project, four prediction models for the dependent parameters have been generated. The results from the study indicate that the combined approach is able to capture the nonlinear or linear relationships among parameters in the CO2 capture process with a high degree of accuracy.