In order to predict the behavior of bubbly flows, it is necessary to know the three dimensional profiles of the bubbles present in the flow. With advancements in the field of flow visualization, accurate reconstruction of the bubble shape has become necessary. The PIV and the SIV techniques, used to acquire images of particles and bubbles, have been found to be extremely useful in this regard. The study, development, implementation, applications and limitations of a unique reconstruction technique applied to various regular and irregular bubble shapes, using the two orthogonal projections of the three-dimensional bubble profiles as captured by the SIV cameras are presented here. The technique is a blend of neural networks, combinatorial optimization and advanced computer aided design methods. The technique involves the robustness and ruggedness of the neural network approach and the flexibility and reliability of advanced computer aided design methods. The technique uses a well-known problem in neural networks and combinatorial optimization known as the Traveling Salesman Problem approach to identify the bubble boundaries on the images. An optimization solution technique known as the Simulated Annealing technique is employed to solve the Traveling Salesman Problem and obtain the bubble profiles. These results are employed to reconstruct bubble shapes using NURBS computer aided design software.
Two main applications of this technique are demonstrated and the results are found to be promising. The first application included the calculation of the void fraction at a particular depth of the channel/ pipe and at a particular radius of the channel. The second application was Lagrangian tracking of bubbles, wherein the centroids of the bubbles were tracked between image frames to determine the linear and transverse velocities of the bubbles.
This technique has shown scope for development including the development as integrated bubble surface reconstruction software and advanced modifications at various levels for efficient and accurate reconstruction.
| Identifer | oai:union.ndltd.org:tamu.edu/oai:repository.tamu.edu:1969.1/384 |
| Date | 30 September 2004 |
| Creators | Vasudevamurthy, Gokul |
| Contributors | Hassan, Yassin A. |
| Publisher | Texas A&M University |
| Source Sets | Texas A and M University |
| Language | en_US |
| Detected Language | English |
| Type | Book, Thesis, Electronic Thesis, text |
| Format | 1405596 bytes, 81060 bytes, electronic, application/pdf, text/plain, born digital |
Page generated in 0.0062 seconds