UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Physical flow modelling of a kraft recovery boiler Ketler, Stephen Paul

Abstract

Measurements of the vertical, and one horizontal, components of velocity have been made in an isothermal scale model of a kraft black liquor recovery boiler. The model used water as the working fluid, and was a 1:28 scale representation of an operating recovery boiler in Plymouth, North Carolina. All of the air ports were represented in the model; however, the char bed shape, mass flow from the bed, and liquor flow were not. Laser-Doppler velocimetry was used to measure velocities and power spectral densities in the model on a 6x6 grid of points on three horizontal planes. Quantitative flow visualization was performed using laser sheet illumination of particles in the flow, with subsequent analysis of the particle images. Flow conditions simulating industrial arrangements and special configurations were run in the model. Industrial configurations employed flows from primary, starting burner, secondary, and either concentric load-burner or interlaced tertiary ports. Special configurations using only primary and secondary port flows were tested to investigate the sensitivity of the flowfield to variations in the lower furnace port flows. Orifice-plate flowmeters and valves were used to set the flowrates through manifolds connected to groups of ports. The error between the set and actual model total flowrate was shown to be dependent on the choice of flowmeters, and on the drift of valve settings with time. Considering both sources of error, for a typical model flowrate of 150 US gallons per minute, the expected difference between the set and actual total flowrate was —6.5 percent. Differences between the set mass flows, and the values calculated from the vertical velocities measured by the laser Doppler velocimeter, were less than ten percent on average. Larger variances between measured and predicted mass flows were explained on the basis of observed low frequency oscillations. Computer software was created for the computation of velocity power spectral densities using the output of the laser-Doppler velocimeter. A system of digital particle image velocimetry was also created for quantitative two-dimensional flow visualization. Laser light was spread into a planar sheet, and the motion of small polystyrene particles added to the flow was recorded on videotape at 30 frames per second. The information on the videotape was digitized, and cross-correlation analysis of successive image pairs yielded a grid of velocity vectors for each 1/30 second interval which could be animated. It was found that the flow in the model was extremely sensitive to any asymmetries in the secondary level port flows. An increase in secondary flow velocity of 10% on one wall, and a corresponding decrease on the opposite wall, caused the core region of vertical upflow at the liquor gun level to occupy half of the model cross section near the wali with the lower velocity. It was very difficult to balance the flows accurately enough to have the core region of vertical upflow in the model centre. In the upper regions of the furnace, the majority of the upflow was near the walls, with down or stagnant flow in the model centre. The flow exhibited low frequency unsteadiness in addition to a high level of turbulence throughout. Periods on an order of 10-20 seconds were observed in the velocity power spectral densities, and in the particle image velocimetry results. The use of concentric load-burner port flows on the tertiary level, as opposed to a 2x2 interlaced arrangement of high velocity tertiary ports, was found to provide a somewhat more uniform distribution of turbulence kinetic energy in the upper furnace. Implications for design and operation of industrial recovery boilers are discussed.

Item Media

Item Citations and Data

Rights

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.