Identification data of a supercritical carbon dioxide extraction process

A pilot-size supercritical carbon dioxide batch extraction system (two-liter reactor) was modeled through data-driven system identification, forming a full simulator consisting of unit process control loops and energy consumption responses. The modeling of the unit processes was based on measurement data gathered by experimental design and careful preliminary examination of the system. The arrangement of the implemented experimental design (central composite design) allowed the exploitation of linear regression analysis. The gathered dynamical data was used in forming state-space representations of the unit processes and energy consumptions. A total of 31 runs were conducted for data gathering, including the 20 runs determined by the experimental design, 9 repetitions and 2 model testing runs. The total duration of the raw data is 34.6 hrs (avg. 1.1 hrs/run) with a sampling time of one second. The experiments determined by the CCD arrangement were split into two sets (runs 1-11 and runs 12-20). These two sets were carried out in consecutive days. Repetition runs of the second set and the two testing data runs were conducted later in discrete time instants. Detailed information of data contents can be found in readme.txt.