Part 4b: Application of data modeling and analysis techniques to the CO2 capture process system. Issue 1 (February 2012)

• Record Type: Journal Article
• Authors: Zhou, Qing; Wu, Yuxiang; Chan, Christine W; Tontiwachwuthikul, Paitoon; Idem, Raphael O; Gelowitz, Don
• Is Part Of: Carbon management. Volume 3:Issue 1(2012)
• Page Start: 81
• Page End: 94
• Publication Date: 2012-02
• Subjects: Carbon dioxide mitigation -- Periodicals; Greenhouse gas mitigation -- Periodicals; Carbon dioxide -- Environmental aspects -- Periodicals; Greenhouse gases -- Environmental aspects -- Periodicals; 363.73874605
• DOI: 10.4155/cmt.11.77
• Languages: English
• ISSNs: 1758-3004

Abstract

The extensive literature review in this article showed that the research for improving efficiency of the CO2 capture process has focused on studying the features and performance of various aqueous amine solvents. Since improving efficiency of the CO2 capture process requires a good understanding of the intricate relationships among the key processes, the ultimate aim of our study is to enhance system efficiency by first explicating the relationships among the key parameters of the process system. The study presented in this article has two objectives: to identify and determine the significance of the process parameters that have influence on the performance of the CO2 capture process and to model the relationships among the process parameters in an attempt to explore the nature of their relationships. Our approach is to apply multiple data mining techniques to the 3-year operational data collected from the amine-based postcombustion CO2 capture process system at the International Test Centre of CO2 Capture located in Regina, Saskatchewan, Canada. The three data mining techniques adopted are statistical analysis, artificial neural network modeling combined with sensitivity analysis and adaptive network-based fuzzy inference system modeling. The data modeling based on the three methods was conducted and the strengths and weaknesses of each method was addressed. It was found that the adaptive network-based fuzzy inference system modeling was the most satisfactory method because it generated the interpretative models with high prediction accuracies. This article presents the process of data modeling and compares the results from each method.