Velocity measurements have been made of the complex flow in the freeboard region about bubbling fluidised beds. Most research was carried out using a 5 m-tall bed, which had a 0.3m x 0.3m square-sided cross-section. Velocity profiles have been measured at various heights above the bed surface. These velocity profiles should allow elutriation calculations to be made, determining the carry-over of fine particles from bubbling fluidised beds. The technique used, Particle Image Velocimetry (PIV), is relatively new and is reviewed here. PIV enables simultaneous velocity measurements to be made at thousands of different points in a fluid flow, allowing transient flow structures to be seen (<I>e.g.</I> short-lived vortices and jets of rapid gas flow). In contrast, most previous research on the topic has used single-point measurement techniques such as Laser Doppler Anemometry, which yield a time series of velocity measurements at a given point but no information about the velocities at nearby points during this time (making it difficult to detect unsteady features of the flow). In this work , velocities have been calculated from PIV flow recordings using a digital analysis method, based on the calculation of an autocorrelation plane for small square-sided area elements from the flow record. This technique is assessed in detail, by means of Monte Carlo simulations. An optimum diameter has been obtained for digital particle images. Recommendations are also given for the minimum and maximum separations of these images, if a good yield of highly accurate vectors is to be obtained after analysis. For digital analysis using square-sided interrogation areas, it is felt that these new recommendations are more appropriate than the older values given by Keane & Adrian (1990, 1991) using circular interrogation spots. It was found that unexpectedly large errors can arise in these PIV measurements due to auto-correlation wrap-around, if the image separations and analysis parameters have certain values. The conditions under which these errors arise have been investigated. New research allowed a model to be developed, which appears to describe these wrap-around errors satisfactorily.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:661179 |
| Date | January 1998 |
| Creators | Rix, Steven J. L. |
| Publisher | University of Edinburgh |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://hdl.handle.net/1842/14287 |
Page generated in 0.0018 seconds