A novel ANN-CFD model for simulating flow in a vortex mixer. (12th October 2022)

• Record Type: Journal Article
• Authors: Sarkar, Sourav; Singh, K.K.; Suresh Kumar, K.; Sreekumar, G.; Shenoy, K.T.
• Is Part Of: Chemical engineering science. Volume 260(2022)
• Publication Date: 2022-10-12
• Subjects: ANN -- CFD -- LES -- Mixer -- Turbulence -- Vortex; Chemical engineering -- Periodicals; Génie chimique -- Périodiques; Chemical engineering; Periodicals; Electronic journals; 660
• DOI: 10.1016/j.ces.2022.117819
• Languages: English
• ISSNs: 0009-2509

Abstract

Highlights: A novel ANN-CFD based modelling methodology reported for simulating flow in a batch vortex mixer. Two-phase flow problem requiring interface tracking simulations reduced to a single-phase flow problem. ANN model trained with a large number of experimental data is used to predict vortex shape. Validation of ANN-CFD model with the PIV data reported in literature. Velocity profiles obtained from ANN-CFD model show good match with the profiles obtained from VOF simulations. Validated ANN-CFD model used to have detailed insights into the flow and turbulence structures. Abstract: The study presents a novel ANN-CFD model for simulating flow in a vortex mixer (an unbaffled vessel stirred by a magnetic stirrer). Large eddy simulations (LES) are performed to simulate flow and turbulence considering presence of both air and liquid phases. The flow fields in air and liquid phases are coupled by applying suitable boundary conditions at the air–liquid interface represented by the vortex. The shape of the vortex is built-in in the computational domain and obtained by using an artificial neural network (ANN) model trained and validated with the data obtained from experiments carried out to capture the vortex shape under different parametric conditions. ANN-CFD model is validated with PIV data reported in literature. The ANN-CFD model reduces the computational time significantly by obviating the need of performing computationally very intensive interface tracking simulations. The model is used to understand hydrodynamics and turbulence characteristics of flow in the vortex mixer.