Monitoring of carbon dioxide capture processes

This work demonstrates the use of multivariable statistical analysis for process transitions between different modes of operation. Continuous process analytics are used to define key variables, named “state variables”, to determine the current mode of operation. Several transition paths and statistical tests are visualized simultaneously using parallel coordinates. Such tests apply confidence limits that are appropriate to the current mode of operation. This methodology is successfully tested in the CO₂ capture plant at the University of Texas in Austin. The results show the effectiveness of using this type of application to detect abnormal operating conditions at different levels of operations.