Development and Validation of a Computational Fluid Dynamics (CFD) Solver for Droplet-Fibre Systems

Abstract

Droplet-fibre interactions are found in many natural and anthropogenic systems. A common industrial example is fibrous filtration - used to capture liquid (e.g. oil) mists. The filters used consist mostly of highly porous arrays of randomly layered fibres. Given the random (complex) nature of these filters, the existing models describing their behaviour are mainly empirical in nature and thus only applicable over a narrow range of parameters and operating conditions. Therefore simulation of these filters using computational fluid dynamics offers a viable alternative to the existing models. In this work we will detail the development of a solver that couples the Lagrangian tracking of particles with a volume-of-fluid (VOF) solver. This solver is built on the existing open-source OpenFOAM CFD libraries, which have been modified to allow the physically accurate modelling of small particles. The solver also models the collection of these particles, where there is a transition from the discrete treatment (as in the Lagrangian tracking of the particles) to the volume-of-fluid treatment. The solver allows the simulation of the motion of small liquid droplets, the capture of these droplets by filter fibres, the coalescence of these captured droplets, into films and the subsequent break up of these films into droplet arrays by Plateau-Rayleigh instability. Also simulated is the movement of these coalesced droplets within the filter, leading to the drainage of oil from the filter.A validation of the fundamental physical mechanisms in the filter was performed, by comparing the simulated conformation of liquid droplets and films on the fibre to Plateau-Rayleigh instability theory. The model showed general agreement with both theory and observations. The simulated capture efficiency was also compared to capture efficiencies predicted by the single fibre efficiency (SFE) theory. A good agreement between the two was found.