| Description |
An ability to predict and control ash deposition are crucial to effective heat extraction from pulverized fuel (pf) units. Computation Fluid Dynamic (CFD) simulations in conjunction with fuel characterization data have been used in recent years to develop sub-models to predict slagging propensities under various operating conditions. A common methodology is to use particle and wall temperature predictions in conjunction with particle Weber numbers (obtained from Lagrangian tracking methods) to assess the slagging and fouling tendencies. However, a Weber number based capture criterion requires as inputs important slag properties like the critical sticking viscosity and surface tension that are needed to estimate the Weber number. Given that there is no consistency in reporting the values of critical sticking viscosity (with values as low as 1 Pa-s to higher than 109 Pa-s been reported in the literature), it is clear that a few key performance parameters (such as ash transport and sticking propensities such as the Weber number criterion) could be "tuned" to match the observations. Another inherent limitation of Lagrangian tracking methods is their inability to track changes to the ash composition and particle-size distribution (PSD) within the combustor which can subsequently impact the ash partitioning and deposition characteristics. This is attributed to the fact that the ash formation is a complex physio-chemical process consisting of: vaporization, condensation, melting, fragmentation, nucleation and coagulation of the mineral matter and organically bound metals in the parent fuel. In spite of these complexities, there are representative scenarios (depending on the location of ash deposition within the boiler) where its deposition rates can be predicted reasonably accurately based on simple Stokes number criterion. This is demonstrated here based on simulations of oxy-coal combustion in a 100 kW down-fired combustor where the deposition rate predictions of the "outer layer" of ash deposits are shown to agree reasonably well with experimental measurements. |
