Chemical Engineering Science, Vol.130, 264-274, 2015
Multiscale analysis and modelling of fluid flow within a photocatalytic textile
This paper deals with a multiscale strategy for the design of an original reactor that will be an assembly of light photocatalytic textiles. This process is intended to decontaminate industrial effluents such as water containing pesticides. The reactor comprises a fabric and optical fibres and the performance of such a complex textile has to be carefully analysed in terms of fluid flow and the resulting ability to degrade target pollutants. We present an experimental set-up based on classical 1D flow experiments to obtain data on fluid flow through the photocatalytic textile. We also propose a numerical model at the optical fibre scale using COMSOL Multiphysics to perform numerical simulations in a geometrical domain consisting of a Representative Volume Element (RVE) of the photocatalytic textile with periodic boundary conditions. A good fit is found between the permeability of the fabric given by the numerical model and that obtained from experimental measurements, which is also in agreement with the value calculated from an experimental determination of the fabric porosity using a permeability model for fibrous media. Then, the effect of geometrical parameters on fluid flow distribution in the textile is characterized numerically. A short optical fibre pitch maximizes the amount of fluid circulating per unit time in the neighbourhood of the region where the degradation reaction takes place. Finally, a simplified analytical model based on a combination of hydraulic resistances in series and in parallel is implemented and validated with the numerical model mentioned above. This simplified model can be advantageously used for further simulations at industrial reactor scale. Additional simulations on the whole textile are performed to better understand the edge effect currently encountered in 1D flow experiments, which could degrade experimental data. It is found that the edge effect can be neglected in the present experiments. (C) 2015 Elsevier Ltd. All rights reserved.