Validation of a procedure for the numerical simulations of gas–liquid stirred tanks by means of a computational fluid dynamics approach. (9th August 2022)

• Record Type: Journal Article
• Authors: Maluta, Francesco; Alberini, Federico; Montante, Giuseppina; Paglianti, Alessandro
• Is Part Of: Canadian journal of chemical engineering. Volume 100:Number 12(2022)
• Page Start: 3472
• Page End: 3485
• Publication Date: 2022-08-09
• Subjects: BT‐6 -- gas−liquid stirred tanks -- CFD -- electrical resistance tomography -- kLa; Chemical engineering -- Periodicals; Technology -- Periodicals; 660.05
• DOI: 10.1002/cjce.24548
• Languages: English
• ISSNs: 0008-4034

Abstract

Abstract: Impellers with concave and vertically asymmetric blades proved superior gas dispersion capabilities and power consumption characteristics under gassed turbulent conditions with respect to traditional flat‐blade turbines in aerated fermenters. In this study, a pilot‐size gas–liquid tank stirred with an asymmetric blade disk impeller is numerically investigated by means of a Reynolds averaged two‐fluid model combined with a simplified population balance model without adjustable parameters. This work aims at increasing the predictive capabilities of computational fluid dynamics multiphase modelling by validating a computational approach for the realistic simulation of industrial aerated fermenters and thus allowing for a more reliable scale‐up. A methodology for achieving fully predictive results on fundamental variables for gas–liquid stirred tanks such as gassed power consumption, overall gas hold‐up, and volumetric mass transfer coefficient, with affordable computational requirements at pilot and industrial scale is presented. Two‐phase results are compared with the experimental data collected in a geometry matching the computational domain equipped with 3 planes of 16 sensors to enable electro‐resistance tomography measurements and with suitable correlations from the literature. The limits and strength of the numerical procedure are discussed, starting from the comparison between computational predictions and experimental measurements.