Numerical approaches for the simulation of the yeast distribution in a fermentation tank

Abstract

AbstractIn recent years, in addition to the experimental investigation of a flow phenomena, numerical investigation has become increasingly established. Especially in investigations where conventional flow measurement techniques such as PIV or LDA fail, numerical computation can offer new possibilities. Models for the simulation of the natural convection flow in a fermentation tank and in second step a superposition of the natural convection by gas bubbles have already been presented in earlier publications. Based on these results, this paper examines the simulation of yeast using the Euler‐Lagrange model. The aim is to extend the existing simulation, which is based on the Eulerian‐Eulerian time‐averaged model and to validate it with measurements from real experiments with the Ultrasonic Doppler Velocimetry (UDV). Starting point is the Dense Discrete Phase Model (DDPM), which is based on the Kinetic Theory of Granular Flow (KTGF). Especially the physical interaction models, for example the lift force, virtual mass force and collision models, play an important role. These fundamental studies must be conducted because the behaviour of the real yeast is really complex. During real fermentation, the yeast increases over time and also combines with other yeast particles. As a result, the proportion of yeast in the process increases and the equivalent particle size also increases due to the combination of the yeast. The goal of this paper is a numerical simulation of the flow processes in the fermentation tank, based on the Eulerian‐Eulerian model with the additional DDPM as realistic as possible, taking into account the natural convection flow and the actual bubble behaviour as well as the yeast distribution.