THE DEVELOPMENT OF AN AUTOMATED MEASUREMENT METHOD FOR THREE-DIMENSIONAL MEAN FLOW FIELDS (MIXING, IMAGE PROCESSING).

摘要

A better characterization of the mean flow systems occurring in the discharge flows of rotating impellers can advance the knowledge of mixing mechanisms in agitated vessels. Since the flow field is rather complex, the resolution of various flow systems and accurate descriptions of these are desired. The method of flow visualization which uses a stereoscopic motion picture technique has been employed to obtain qualitative information regarding coherent flow structures and their interactions. Sheu et al. showed that quantitative data could be obtained by manually tracking small neutrally buoyant tracer particles entrained in the flow. Three dimensional velocity profiles, which are needed for flow pattern studies and are functions of time also, can be evaluated from the discrete sets of particle locations in stereo scenes.;This automated measurement technique which combines the stereoscopic motion pictures and the image processing is capable of providing sufficient data with high resolution to characterize the flow field. Data such as this have been taken only in limited quantities before due to the tediousness of the work. Such an automated analysis system will be highly useful in the understanding of the mean flows in agitated vessels and other flow systems.;Instead of tracking particles manually as in the works of Komasawa et al., Schwartzberg and Treybal, and Sheu et al., a new automated data acquisition method has been developed. This newly developed measurement method applies the image processing techniques to extract quantitative data from a sequence of films. The steps are: (1) visual images are transformed into numerical form by digitization so that the images are amenable to computer analysis, (2) a software package utilizing 'THRESHOLDING' and 'PATTERN MATCHING' is then developed to perform the image identification, tracking, and matching, and (3) an evaluation of instantaneous flow velocities in a three dimensional space. Important attributes for flow characterization such as the velocity, acceleration, turbulent intensity, and other properties can be collected for analysis at a much faster rate than ever before. Furthermore, considerable improvement in accuracy over that of manually tracked method is also achieved.