nNumerical investigation of bubble interaction mechanisms in gas-liquid bubbly flows: Harmonisation of bubble breakup and coalescence effects

Abstract

Numerical investigation of the dynamics of gas-liquid bubbly flow has been performed by using the population balance approach. Using the homogeneous MUSIG model coupled with the two-fluid model, the evolution of bubble size distribution in vertical pipes has been simulated with appropriate consideration of bubble interaction mechanisms. In particular, the study focuses on harmonising the imbalance often associated with breakup and coalescence rates by performing extensive numerical investigations for the examination and comparison of different model formulations to harmonise the coalescence calibration factor of Prince and Blanch (1990) model. To avoid the fine-tuning procedure of selecting calibration factors to harmonise the breakup and coalescence effects, a breakup calibration factor of 1.0 (default setting in ANSYS CFX) was used and three different expressions area-averaged and included as the coalescence calibration factor for the model of Prince and Blanch (1990) modelled using CFX Expression Language (CEL). The performance of the three different “cases” (modifications) has been investigated, compared with each other and validated against the experimental data reported by Monr´os et al. (2013). In general, the comparison has shown that all the “cases” yielded satisfactory results when the numerical results were compared with five primitive experimental parameters, namely: gas volume fraction, interfacial area concentration, Sauter mean bubble diameter, gas velocity and liquid velocity. The encouraging results obtained clearly show the capability of the “cases” modelled in capturing the evolution of the bubble sizes. The agreement with the gas volume fraction profiles indicates a level of confidence in the interfacial force models used. Similarly, the agreement seen with the interfacial area concentration indicates that the “cases” modelled for the birth and death processes are reasonably adequate to describe the bubble dynamics. Overall, the comparison shows that the model labelled as “case 2” presented the best results in most of the experimental conditions simulated.