Abstract in Englisch:

This paper presents a detailed validation of a modeling framework and its implementation for the simulation of flame spray pyrolysis (FSP) using different nozzle geometries of the so-called SpraySyn burner. Gas, liquid and particulate dynamics are compared against experimental data from literature as well as our own laser-Doppler anemometry, phase-Doppler anemometry and multi-line OH laser-induced fluorescence thermometry imaging measurements. The modeling framework consists of large eddy simulations (LES) coupled to the sparse-Lagrangian multiple mapping conditioning (MMC) model, a Lagrangian spray solver and a sectional description of the population balance equation. Simulations start downstream of the burner exit planes where turbulent inlet data for the gas and liquid phases are generated by independent LES that use a 1-Fluid method to capture the dynamics of the liquid jet break-up. The gas and liquid dynamics are validated and analyzed for an ethanol spray flame in the SpraySyn1 and SpraySyn2 burners. In these burners iron(III) oxide particulates are synthesized from iron pentacarbonyl (IPC)-ethanol solutions with varying IPC concentrations. The results demonstrate the very good predictive capabilities of the MMC modeling framework for the gas and liquid phases. The predictions of the particulate formation are validated by comparison of elastic light scattering (ELS) signals from experiments against synthetic ELS signals calculated from the simulations. Results are of reasonable accuracy for the SpraySyn1 FSP series but indicate an imbalance between particulate growth and dilution with the surrounding gas. Predictions for the SpraySyn2 FSP series are consistent with SpraySyn1 results and indicate an increased product particulate size due to an increased residence time downstream of the heat release zone.