Computational simulation of aerosol behaviour

Abstract

In this thesis, computational methods have been developed for the simulation of aerosol dynamics and transport. Two different coupled aerosol-computational fluid dynamics (CFD) models are discussed. One is a full Eulerian model and the other is a boundary layer-type streamtube model. The streamtube-based sectional model is able to provide an accurate solution of model equations within a reasonable computing time. For a number of studies, one-dimensional simulations are sufficient. Flow dependence can be taken from CFD simulations, flow correlations, or from experimentally-based estimates of time-temperature histories. While the focus is on the sectional method, a bivariate extension of the quadrature method of moments (QMOM) has also been tested. The method is shown to be able to provide reasonable computational representations of aerosol particle shape evolution. The models have been used to analyse various cases of aerosol formation, growth and deposition.

Aerosol formation and growth dynamics simulations are used in analyses of aerosol formation experiments in a laminar flow reactor, experiments of particle property evolution in a counterflow diffusion flame reactor, and aerosol formation mechanisms in recovery boilers. Computational simulations of recovery boilers demonstrate the feasibility of Na2SO4-route fume formation mechanism theory. The accordance between particle size distribution predictions and experimental data is fairly good.

Model studies of deposition provide insights into the transfer mechanisms of fly ash particles and inorganic vapours to the heat transfer surfaces of industrial boilers. Estimates of deposition velocities are obtained for particles of various sizes and inorganic vapours under various conditions. An area in which aerosol dynamics and transport processes are especially significant is the case of alkali chloride deposition. For these species, there seems to be a great deal of variation in the proportions of particle and vapour deposition in the typical temperature range of biofuel fired boiler superheater conditions. In recovery boilers, particle deposition is predicted to be the dominant final alkali chloride deposition mechanism for freshly soot-blown superheater surfaces, while direct vapour deposition can be equally important when there is an insulating deposit layer, and, consequently, higher temperatures at the deposit surface. The total rate of alkali chloride deposition does not have as strong a temperature dependence. While homogeneous nucleation is predicted to be negligible in recovery boiler superheater tube boundary layers, the reverse is true for conditions resembling fluidised beds during bio- and mixed fuel combustion. The relationship between deposit-related problems, such as fouling, slagging and corrosion, and the rate of deposition is not necessarily straightforward. Initial steps have been taken to simulate deposit transformations in order to provide computational demonstrations of the implications of alkali chloride deposition. A model that links alkali chloride deposition to the presence of HCl in the deposit-metal interface has been developed and illustrative simulations have been carried out.

Overall, the models are shown to provide an insight into aerosol dynamics and transport phenomena in various systems. It is noted that the development of a universal CFD-based aerosol model is difficult. Instead, it is often more appropriate to make suitable simplifying assumptions, such as a parabolic flow, a constant size distribution shape or a reduction to one dimension. Suitable assumptions allow the use of computationally efficient and accurate methods. These make it possible to extend the scope of modelling beyond aerosol issues, for a more extensive analysis.