Imaging and Pulse Light Velocimetry Applied to Bubbly Flows (Keynote) Available to Purchase

Bubbly flows are of great importance in many technical areas, such as chemical engineering or biotechnology. Generally, bubbly flows are operated at relatively high gas volume fractions, resulting in strong interactions between bubbles and complex unsteady flow structures, as for example in a bubble column. A detailed experimental analysis of these phenomena can however only be performed using optical measurement techniques. For this purpose methods of pulse-light velocimetry (PLV) were extended for reliable applications in bubbly flows. Two approaches were developed namely a large scale PIV (particle image velocimetry) which may be applied for velocity measurements of both phases over an entire cross-section of a bubble column and a small scale PLV which provides detailed information on the flow structure in a bubble swarm and additionally yields the bubble size. The PIV system for the large scale flow analysis consists of a pulsed Nd-YAG Laser to produce a light sheet over the entire cross-section of the bubble column (diameter 140 mm). In order to determine the velocity of the fluid phase fluorescing tracer particles were added. For recording the scattered light of the bubbles and the fluorescing light of the tracer two CCD-cameras in a non-perpendicular arrangement were used, each having an appropriate optical filter. Hence, simultaneous velocity measurement of both phases were possible with a reliable discrimination between the phases. Using a fully automated traversing system the time-averaged flow field in the entire column was recorded (Bro¨der & Sommerfeld 2002 a). With this optical arrangement it was not possible to determine the bubble size, since in a light sheet the bubbles appear on the image only through their glair points. However, by employing a pulsed background illumination using a LED array, shadow images of the bubbles can be recorded. In order to realise this approach, only one CCD-camera was required, which however was equipped with a macro-lens yielding a small depth of focus (in this case about 4 mm). Hence, the image plane was not produced by a light sheet, but determined by the depth of focus of the macro-lens. For evaluating the bubble phase properties only sharply depicted bubbles were considered by using an edge detecting Sobel-filter as illustrated in Fig. 1. This method allowed to determine an equivalent bubble diameter based on their cross-section. Furthermore, other relevant parameters such as bubble orientation and bubble aspect ratio were evaluated. The bubble velocity was obtained using particle tracking velocimetry (PTV). A typical result of the aspect ratio of the bubbles as a function of their size is shown in Fig. 2. Bubbles below 1 mm are spherical as expected. A further increase of bubble size is coupled with a linear increase of the aspect ratio up to bubble sizes of about 1.8 mm. This result coincides with the findings of Duineveld (1994). For bubble sizes between 2 and 5 mm a slight decrease of the aspect ratio is found. With the same approach also images of tracer particles could be recorded simultaneously in the considered image plane (see Fig. 3a)), produced by back-lightning and a macro lens with a small depth of focus. For separating out-of-focus tracer particles from the image a filter called Laplacian of Gaussian (LoG) was used (Bro¨der & Sommerfeld 2002 b). The velocity field of the tracer was obtained with the PIV-technique using a hybrid approach combining the fast MAD (maximum absolute difference) and the accurate MQD (maximum quadratic difference) approaches. The later was developed by Gui & Merzkirch (1996, 2000). Hence, the flow structure within a bubble swarm (Fig. 3b)) could be determined and analysed. The above described method was also applied for measurements in a special laboratory loop reactor for analysing bubble coalescence.