Characterisation of gas-liquid dispersions for beer analogues
File(s)
Author(s)
Hepworth, Neil James
Type
Thesis
Abstract
The cloud or haze of microbubbles, which results when a glass of beer is dispensed (poured) through a nozzle from a pressurised vessel is an important characteristic of the product, which affects consumer satisfaction. How process conditions effect the ‘bubble haze’ in terms of bubble size distribution, the rise velocity of the haze and the surge time (the time for the haze to disappear) are not understood by the brewing industry. The aim of this study, therefore, is to understand how the following important process conditions affect the bubble haze characteristics: dissolved gas composition, beer flow rate and use of flow restrictors (’sparklers') in the nozzle. The study incorporated both experimental and modelling techniques. A model (synthetic) beer and rig were developed, tested using commercial beer and found to successfully produce ‘bubble haze’. This set-up was then used to investigate the ‘bubble haze’ bulk properties, in conjunction with video cameras. A novel CCD (Charge Coupled Device) camera based image analysis procedure was developed for use with a specifically designed thin-section receiving vessel to determine bubble size distributions and also individual bubble velocities. Images of bubbles produced in the glass and also in the beer nozzle only were analysed. The surge time was found to increase with increasing beer flow rate, nitrogen content in the supply gas and with the addition of a sparkler. The localised rise velocity of the bubble haze fell into two distinct sections, distinguished by a doubling in rise velocity. This doubling marked the point at which the liquid became sufficiently quiescent for the bubbles to rise unimpeded. The surge time increased with decreasing average bubble diameter and increasing gas hold-up as expected from consideration of the underlying physical properties and the findings of other researchers. Experimental results were compared with modelling three aspects of the dispense process: 1. Bubble nucleation rates were modelled by adapting existing nucleation models to incorporate liquid motion. 2. A combined bubble growth and motion model, successively predicted bubble growth rates and haze rise velocities. 3. Computational Fluid Dynamics simulations accurately modelled the development of the gas-liquid dispersion and the decay of the liquid flow patterns.
Version
Open Access
Date Awarded
2003
Copyright Statement
Attribution NoDerivatives 4.0 International Licence (CC BY-ND)
Advisor
Varley, Dr Julie
Sponsor
Dr John Hammond; BRi; UK Biotechnology and Biological Sciences Research Council.
Publisher Department
Chemical Engineering and Chemical Technology.
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)