A volume of fluid approach to model injection of highly viscous fluids

Despite the transition to renewable energy sources, fossil fuels will still play a significant role in satisfying the world's energy needs shortly. In addition, the rise in the demand for light distillates and the depletion of light crude oil reservoirs are shifting the interest toward the conversion of heavy fuel oils. The gasification process converts solid or liquid organic mixtures into lighter and cleaner components. Liquids gasification is often performed in entrained-flow gasifiers: a spray is injected from the top into an oxidizing environment in these reactors. One of the significant challenges of gasification of a liquid feed is to achieve adequate atomization, which results in increased yield of the process and minimization of solid residue formation. Both phase change and reactivity of the liquid are proportional to the surface exposed to the hot oxidizing environment. In this context, it is crucial to model the injection process. The fuels processed in gasification reactors consist of heavy and generally viscous mixtures. The peculiar physical properties of those fuels need to be adequately modelled in CFD simulations of the injection. This work demonstrates the application of the volume of fluid (VoF) method to simulate injection in an entrained-flow gasifier. The VoF allows capturing the interface between liquid and gas through a variable, called volume fraction, which is the non dimensional volume of one phase in the computational cell in the grid. The software adopted for the simulations is based on the OpenFOAM library. It combines a series of state-of-the-art techniques to address the difficulties of the problem in the framework of finite volume large eddy simulations. The volume fraction of liquid is advected geometrically using the isoAdvector algorithm with piecewise linear interface construction (PLIC) to accurately resolve ligaments down to small droplets. The latter is essential for accurate curvature estimation. Adaptive grid refinement (AGR) helps this purpose. Also, a transition of smaller droplet to Lagrangian particles unlocks large-scale simulations while optimizing computational cost. The physical properties were assumed to be constant, not a function of the liquid temperature pressure or concentration, since the injection is a relatively short phenomena confined to few diameters downstream the nozzle. The same assumptions would not be valid outside the considered region. The goal of the work in progress is to identify the proper injection mechanism and highlight which Sauter mean diameter distribution would be best to set up proper inlet boundary conditions in a reduced-order gasifier model.