Deposition prediction in a pilot scale pulverized fuel-fired combustor

Abstract

Fossil fuels have traditionally been used in power generation systems and represent the main source of greenhouse gas emissions from this sector. Renewable fuels, especially biomass, are now being substituted for fossil fuels to reduce CO2 emissions. Co-firing biomass with coal, which has been widely practised in the UK and Europe, is one route to reduce the environmental impact of using coal. However, the deposition of ash particles and vapour species on heat exchanger surfaces during operation is a serious issue in pulverised coal and biomass fired power plant as this reduces the plant thermal efficiency and can cause fireside corrosion, which limits component lives. Deposit formation is difficult to predict as it varies with many factors including: boiler geometry, combustion conditions and fuel composition.

Computational Fluid Dynamics (i.e. CFD) is one of the best modelling tools to study the flow behaviour of gases and particles around heat exchanger tubes and predict deposition. This work used an Eulerian-Lagrangian model to describe the gas flow field around tubes and the solid ash particle trajectories respectively. User Defined Functions (i.e. UDFs) were developed for the CFD package to enable the prediction of deposit growth, deposit shape and temperature gradients around superheater/reheater tubes. Deposit build up insulates such tubes from the flow of the hot combusted gas stream and reduces heat transfer between this gas stream and the steam coolant following within the tubes, thus raising the deposit temperature. The CFD-based predictions generated were consistent with available literature data. The CFD deposition model has been applied to predict deposition on air cooled ceramic probes in a 100 kWth pilot-scale, pulverized fuel (PF) combustor and compared to deposition data measured after the combustor rig runs. For modelling purposes, the geometry was simplified to a two-dimensional domain with inert spherical ash particles dispersed in air and injected at an inlet plane. Experimental data from a series of rig runs have been used to test this CFD deposition modelling approach