Axial dispersion and CFD models for the extraction of levulinic acid from dilute aqueous solution in a Kühni column with 2-methyltetrahydrofuran solvent

Tutkimustuotos: Lehtiartikkelivertaisarvioitu

Tutkijat

  • Antero Laitinen
  • Karri Penttilä
  • Mikko Manninen
  • Jouni Syrjänen
  • Juha Kaunisto
  • Lasse Murtomäki

Organisaatiot

  • VTT Technical Research Centre of Finland

Kuvaus

1D axial dispersion and 3D CFD models for the extraction of levulinic acid from dilute aqueous solution by applying 2-methyltetrahydrofuran as a solvent are presented. The models are validated by comparison with the measured levulinic acid concentration profile data obtained in a bench-scale Kühni column. The 1D model contains NRTL parameters for the system levulinic acid-water-2MTHF. Correlations for drop size and hold-up for Kühni columns were taken from literature. The values for overall mass transfer coefficient ranged from 1.4E-5 to 2.2E-5 ms −1 , and increased as a function of the rotor speed. The fitting of the column performance resulted in a very good prediction of the solute concentration profiles in the extraction column, and the average absolute value of relative error for the 1D model was 23%. CFD model visualized the column performance at the column height of 150.5–160 cm giving valuable information on back mixing, phase velocities, dispersed phase volume fraction, and mass transfer. Dispersed phase volume fraction and mass transfer contours revealed, that the mass transfer rate (app. 0.25 g L −1 s −1 )is at its highest just below the rotor, and that there are blind spots in the compartments close to the extractor and just above each down comer. Values for the dispersed phase volume fraction are highest in the same area where the mass transfer reaches the highest values. The highest slip velocity values (app. 0.03 m −1 )are located in the tip of each compartment partition plates. General correlations, such as hold-up and drop size correlations, can successfully be applied in levulinic acid-water-2MTHF system reported in this work. The 1D axial dispersion model proved to be valuable tool for scale-up purposes, and CFD model, despite the long time needed for each simulation, gave useful information for the design purposes.

Yksityiskohdat

AlkuperäiskieliEnglanti
Sivut518-527
Sivumäärä10
JulkaisuChemical Engineering Research and Design
Vuosikerta146
TilaJulkaistu - 1 kesäkuuta 2019
OKM-julkaisutyyppiA1 Julkaistu artikkeli, soviteltu

ID: 33836358