Dynamic modelling and simulation of turbulent bubbly flow in bubble column reactors

Considerable progress in understand and predicting turbulent bubbly flow in bubble column reactors has been advanced over the last two decades or so using a combination of model development, computational techniques and well-designed experiments. However, there remain many modelling uncertainties mainly associated with inadequate physical prescriptions rather than numerical schemes. The present project addresses some of these questions, in particular in relation to the interactions between the deformable rising bubbles and the turbulent eddies, with the later which from liquid shear flow or in the wakes of bubbles. Recent literature on existing models and experimental studies of bubble column reactors is reviewed in Chapter 1. It appears that the correlations and phenomenal models developed from early-stage experimental studies have been implemented into CFD modelling, and in return, accelerates the developments of theoretical understandings of the flow characteristics in the bubble columns. The research efforts made from both CFD modelling and experimental studies to understand the complicated mechanisms of gas-liquid interactions have been summarised in this chapter. In chapter 2, the inlet conditions, as one of the important issues in the CFD simulations of bubble columns, have been addressed. A kinetic inlet model is proposed, which considers the effects of number and size of holes on the gas spargers, the volume flow rate, and the gas-phase velocity profile. The proposed model achieves similar accuracy as modelling the real sparger holes while the computational costs have been significantly reduced. Chapter 3 applies a CFD-PBM method to investigate the influence of various shapes of bubbles on the bubble breakage rate and bubble size distribution. Bubbles are classified into spherical, ellipsoidal and spherical-capped shapes, and explicitly calculated in the breakage kernel. The correlation of aspect ratio of ellipsoidal bubbles is developed base on dimensionless numbers, summarising the effect of buoyancy, surface tension, and viscosity. The surface energy and pressure head have been adopted as two competing breakage mechanisms with the energy density constraint has been used as the breakage criterion. The simulation results demonstrate improvements in the estimations of gas holdup, liquid velocity, and bubble size distribution, as well as strong enhancements in mass transfer prediction. The effects of the turbulent kinetic energy spectrum for the turbulent bubbly flow on the bubble breakage are considered in Chapter 4. The κ-3 power law scaling behaviour of bubble induced turbulence is considered together with the Kolmogorov -5/3 law to characterise the turbulent eddies that interact with the subsequent bubbles. A characteristic length scale Λ is used to approximately separate the shear turbulence and bubble induced turbulence. The implementation of the modified breakage model into CFD modelling shows a great improvement in the prediction of bubble breakage rate, which believes to be competitive to the results obtained from Chen et al. (2004) that has artificially increase of breakage rate by 10 times. In Chapter 5, the approaching velocities of collision bubbles that are under the influence of shear turbulence and bubble induced turbulence are clearly distinguished. The turbulence dissipation rate that strongly affects the estimation of collision time has been calculated by taking into account the turbulence generation and dissipation in the wakes of bubbles, especially considering the anisotropic feature of bubble induced turbulence in the Reynolds stress turbulence model by using extra source terms. The modified coalescence model properly addresses the coalescence rate for different sizes of binary bubble coalescence. Chapter 6 presents the experimental study of the spatial velocity fluctuations and the turbulence energy spectrum in the wakes of bubbles by using PIV and highspeed imaging techniques. The experimental results clearly demonstrate the existence of the κ-3 power law scaling region due to bubble induced turbulence. The theoretical analysis successfully shows that the scaling exponent of -3 to be robust from three different aspect. In sum, some important issues of the gas-liquid interactions in turbulent bubbly flows have been addressed in this project. The implication is that the liquid phase turbulence is strongly affected by the size and shape of rising bubbles. Meanwhile, it can be found from the turbulence energy spectrum that the behaviours of turbulent eddies in the wakes of bubbles are very different from those in shear flow, thereby strongly influencing the kernels of bubble coalescence and breakage and hence the model predicted bubble size distributions