Analysis and improvement of simulated moving bed chromatography (SMB) were conducted to separate multi-components of C10–14 normal paraffin (n-paraffin) from kerosene. Owing to the complexities of the feed composition and operational variables, a mathematical dynamic model using limited experimental parameters and data-driven machine learning using real industrial data were both employed to evaluate SMB performance. The developed dynamic model could evaluate the purity, recovery, and solvent consumption. Under standard operating conditions, the simulation results indicated that the extracted n-paraffin purity was 99.6%; and the recoveries of nC10, nC12, and nC14 were 95.7%, 88.9%, and 81.6%, respectively. The analysis of the counter-current ratio in each zone suggested that recovery can be enhanced by decreasing the Zone I and IIb ratios and increasing the Zone III ratio. In particular, controlling the Zone III became more important as the carbon-chain-length increased. An additional second desorbent stream in the n-paraffin SMB, known as zone flush, played an important role in further improving the purity of the extracted n-paraffin. The machine learning results clearly presented the importance of zone flush on impurity removal, which could not be sufficiently estimated from the dynamic model. The increase in p-xylene content in the zone flush contributed to a reduction in aromatic impurities in the extract. Owing to the nature of the industrial data obtained in a limited operation range, it was recommended that the mathematical and data-driven models should be used complementarily. Since the subsequent fractionation section consumed a significantly larger amount of energy than the SMB itself, an exergy analysis was conducted on the overall n-paraffin SMB process integrated with the fractionation section and potential improvements were suggested. The results will contribute to the establishment of design and operation guidelines for improving n-paraffin SMB performance.
Bibliographical noteFunding Information:
This work was supported by the National Research Foundation of Korea (NRF) and funded by the Ministry of Science and ICT (NRF-2020K1A4A7A02095371).
© 2022 Elsevier B.V.
All Science Journal Classification (ASJC) codes
- Analytical Chemistry
- Filtration and Separation