OCR Text |
Show acceleration of the flow as it traveled through a restriction (tube bank), and the k factor model accounted for the flow resistance across the restriction. The grid was subdivided into eleven (11) sub-grids containing 325 defined porous regions. The porous regions in Grid 1 and Grid 3 represent the superheater and preheater pendants that have 18" centres. These porous regions were assigned a porosity of 7 0 % . The porous regions defined in Grid 4, Grid 6, Grid 8 and Grid 10 represent the tube banks having 6" centres. These regions were assigned a porosity of 9 0 % . These porosity values are roughly equivalent to the open cross-sectional area in these various regions. The k factor model accounts for the flow resistance important for turbulent flows and it represents a pressure drop by: dP/dXi = ki{l/2pVi |Ui| } In these simulations the pressure losses were calculated in the x, y and z directions. The A'-factors used for the porous media model were based on empirical data given in the literature for heat exchanger design (Kays and London, 1964). From this analysis flow across the tube banks would have a friction loss term about ten times that of flow along the tubes. For the pendant sections with 18" centres (Grid 1 and 3) a A-factor, for the x and z directions (perpendicular to the tubes), of 0.1 was used and the A-factor for the y direction (parallel to the tubes) was 0.01. For the remaining 6" centre heat exchangers the A-factor in the x and z directions was 0.5 while for the y direction 0.05 was used. These values result in a pressure drop up to the inlet of Grid 10 (from the superheater inlet to the "outlet plane" of interest) of 1280 Pa. This is in good qualitative agreement with published values (Singer, 1981; Babcock & Wilcox, 1978) of draft losses through the superheater/reheater sections of about 4" H 2 O (~ 1000 Pa). N o quantitative boiler operating values from the Keephills plant were available for comparison. 1.2 Heat Loss Model In order to apply a sorbent sulphur capture model accurate flow and temperature predictions were required as these would influence the particle residence time and temperature history. A heat loss model was developed that accurately mapped the temperature gradients in the sections of the boiler of interest. This model was based on the volume average heat loss for each heat exchanger section where measured temperature values were known. This average value was applied to the individual computational elements in the porous regions across which the temperature change was observed and was scaled by the local gas temperature and the log-mean temperature difference ( L M T D ) for the heat exchange section being modelled. For the heat loss model the grid was subdivided into five different sections, as described above, where measured temperature information existed. The first section consisted of Grid 1, 18" centre superheater section with an inlet gas temperature of 1200 °C and an outlet gas temperature of 1130 °C, the second section consisted of Grids 3 and 4, 18" and 6" centre reheat sections with a temperature drop from 1130 °C to 867 °C, the third section being Grid 6, a 6" centre superheater with a temperature drop from 867 °C to 735 °C, the fourth section was the cross over tubes, Grid 8, with 6" centres and a temperature drop from 735 °C to 710 °C, the fifth section being the down flow section past the "outlet plane" or Grid |