Heat transfer in a pressurized fluidized bed with continuous addition of fines

Abstract

The goal of this work was to investigate the impact of pressure and the presence of fine particles (i.e., a surrogate for pulverized fuel) on the overall surface-to-bed heat transfer coefficient in relation to an oxygen-fired pressurized fluidized bed combustor. Experiments were conducted in a pilot-scale fluidized bed with an inner diameter of 0.15 m under cold flow conditions. A tube bundle consisting of five horizontal staggered rows was completely submerged in the bed. One of the tubes was replaced by a heating cartridge housed in a hollowed copper rod. Five thermocouples distributed at 45° intervals along the copper rod circumference measured the surface temperature and ensured that local effects were included. The bed material was glass beads of 1.0 mm in diameter while the fines were glass beads of 60 μm in diameter and thus susceptible to entrainment. The fine particles were continuously fed to the fluidized bed and then captured downstream by a filter system. Fluidization was conducted at 101, 600 and 1200 kPa (abs) with excess gas velocities (Ug-- Umf) of 0.21, 0.29 and 0.51 m/s. Fine particle feed rates were 0, 9.5 and 14.4 kg/h. Two heating rod positions (2nd row and 4th row) were studied. Overall, the heat transfer coefficient approximately doubled when pressure was increased from 101 to 1200 kPa. At atmospheric conditions, where the slug flow regime occurred, the maximum heat transfer coefficient was at the bottom of the rod, while it moved to the side of the rod at high pressures where the bubbling regime occurred. When the heating rod was moved from the 2nd row to the 4th row, the averaged heat transfer coefficient increased by 18%, 9% and 6% at 101, 600 and 1200 kPa. The addition of fine particles decreased the average heat transfer coefficient by 10 to 20 W/m2 K, where the averaged coefficients were approximately 220 and 450 W/m2 at 101 and 1200 kPa respectively, but there was no effect on the angular heat transfer profile around the tube surface. The results showed that average heat transfer coefficients matched the correlation developed by Molerus et al. (1995) within a 5% difference across all conditions when fines were not present.