Date of Award
Doctor of Philosophy (PhD)
Bubble columns and slurry bubble columns are multiphase reactors used for a wide range of applications in the biochemical, chemical, petrochemical, and metallurgical industries. In spite of their widespread usage, the scale-up of bubble columns remains an ongoing challenge. Various scale-up approaches, based on concepts ranging from ideal mixing to complex 3-D multiphase CFD models, have been used for assessing the effect of column size and gas and liquid flow rates on column hydrodynamics and reactor performance. Among these approaches, phenomenological models based on either single-class or multi-class bubbles that were validated on cold flow systems have been successful in predicting the residence time distributions of gas and liquid in pilot-scale bubble columns (Chen et al., 2004) (Gupta, 2002). However, such models are not entirely predictive, since they are validated using columns having the same size as hot operating units. To provide better predictive capability, we need prior knowledge of local hold-up, transport coefficients, and bubble dynamics. This dissertation provides an improved understanding of the key design parameters (gas hold-up, volumetric mass transfer coefficients, gas-liquid interfacial area, and their spatial distribution) for predictive scale-up of bubble columns.
In this work, a 4-point optical probe is used to estimate local gas hold-up and bubble dynamics (specific interfacial area, frequency, bubble velocity, and bubble chord-lengths) and their radial profiles in a cold-flow slurry bubble column and a bubble column photo-bioreactor. Along with local bubble dynamics, the effect of superficial gas velocity on volumetric mass transport coefficients in several sizes of bubble columns, with and without internals, and in slurry bubble columns and photo-bioreactors are studied.
In the bubbly flow regime, bubble dynamics in photo-bioreactors with suspended algae were dominated by the physicochemical properties of the liquid, as distinguished from the churn-turbulent flow regime in the slurry bubble columns, where bubble dynamics were mainly affected by turbulent intensities.
In the bubbly-flow regime, volumetric mass transfer coefficients increased with an increase in superficial gas velocity. However, in the churn-turbulent flow regime, they approached a constant value with an increase in the superficial gas velocity.
A new methodology was proposed to identify the flow regime from optical probe signals based on the support vector machine algorithm, which can uniquely classify flow regimes for various systems on a single flow regime map.
A new model for the liquid phase mixing, that with a proper choice of the mass transfer coefficients enables a good match of the predicted and measured tracer response is described. This model provides a better prediction of volumetric mass transfer coefficients than the currently used well mixed model for the liquid phase (CSTR).
The dissertation improves the fundamental understanding of the connection between bubble dynamics and mass transfer. Using the 4-point optical probe as a tool, it demonstrates a connection between bubble dynamics and volumetric mass transfer coefficients. Present work addresses the need of industries to have a method that can be used as an online process control tool to identify flow regime, this method has been tested at cold flow conditions and needs to be implemented at hot flow conditions. The parameters (radial distributions of gas hold-up, bubble velocities, and volumetric mass transfer coefficient) that are evaluated in the present work can be used to validate phenomenological models and CFD results at cold flow conditions, which can later be combined with process chemistry to accomplish scale-up (Chen et al., 2004).
The open literature on multiphase reactors is mainly limited to cold flow condition, and techniques such as the optical probe need to be extended to hot flow conditions. The optical probe described here can withstand high temperature and pressure, but for hot flow conditions it requires a better binding agent to hold the probe tips together, one that will not dissolve in industrial solvents.
Milorad P. Dudukovi
Humberto Gonzalez, Patrick L. Mills, Palghat Ramachandran, Yinjie Tang,