Author: Pritheesh Gnanaselvam

Publisher:

Published: 2020

Total Pages: 140

ISBN-13:

The simulations and experiments outlined are designed to explore the effect of existing turbulent dispersion models in predicting particle deposition characteristics at higher temperatures. The continuous phase solution was obtained from using a Reynolds-Averaged Navier Stokes (RANS) turbulence model and the turbulent dispersion was modeled using a Continuous Random Walk (CRW) model. Euler-Maruyama scheme was implemented to solve the non-dimensional Langevin equation to model the stochastic nature of the equation appropriately. Previous studies have shown that the particle deposition characteristics depend greatly on the time step of integration. With the Euler-Maruyama scheme, the CFD results were shown to be less sensitive to the time step of integration and with decrease in time step more stable results were obtained. Direct comparison with the Discrete Random Walk (DRW) model shows that DRW fails to predict flow fluctuations seen by particles in the diffusion-impaction regime. Previous studies of this phenomenon were all performed at ambient conditions. The CRW model was shown to predict impact velocities reasonably well, when the chosen time step of integrations is such that the stochastic and damping term are comparable in magnitude. Presented here are pipe-flow experiments conducted in the High Temperature Deposition Facility (HTDF) with a mean jet velocity of 150 m/s – 200 m/s with exit centerline temperature of 1525K to assess the capability of CRW in predicting particle deposition characteristics at high temperatures. The flow temperature was chosen in such a way that the temperature inside the pipe at any point is higher than the melting point of dust used, so that an `all stick’ condition can be used to model particle-wall interactions. The derivation and the effect of the drift correction and the stochastic terms in the normalized Langevin equation were discussed in detail. Simulations were performed trying to reproduce experimental results with and without injection line. The CFD model without injection line was shown to follow the predicted pipe deposition based on the effective drift correction, whereas the CFD model with injection line did not show significant change in pipe deposition results with velocity. CFD results using the OSU-CRW model posited that the model is more appropriate for turbulent pipe flows in the fully developed region.