Predicting microscale bubble to slug transition boundary using an Artificial Neural Network

Abstract

Microchannel heat sinks have potential applications in, for example, miniature refrigeration systems, the cooling of computer chips and power electronics and the cooling of fuel cells1. Implementing flow boiling in microchannel heat sinks promises to provide significantly greater heat transfer rates than single-phase flows due to the utilisation of latent heat. However, general predictive tools for heat transfer rates and pressure drops in microchannels must be derived and agreed to facilitate extensive adoption by industry. Microscale heat transfer rates and pressure drops are fundamentally dependent upon the prevailing flow patterns, which describe the geometry of the liquid-vapour interface.