Journal of Flow Visualization and Image Processing

ISSN for PRINT: 1065-3090

For Online Access

Best Paper Award Selection - Editorial Board Site

2001, Volume8

Issue 2-3

110 pages

Issue price - $280.00

Murthy Balu
University of Saskatchewan, Collage of Engineering, 57 Campus Drive, Saskatoon, S7N 5A9, Canada

Ram Balachandar
University of Saskatchewan, Collage of Engineering, 57 Campus Drive, Saskatoon, S7N 5A9, Canada

Hugh Wood
University of Saskatchewan, Collage of Engineering, 57 Campus Drive, Saskatoon, S7N 5A9, Canada

Knowledge of the variation of scalar quantities like concentration is required to understand the mixing and dilution characteristics of environmentally important flow fields. The present study deals with the development of a nonintrusive flow visualization technique using neural networks to study the topology of dye concentration distribution in two-dimensional flow fields. Aflow fast a bluff body in a shallow open-channel flow is used to illustrate the technique developed. The present study aims to eliminate problems associated with previous image to concentration conversion techniques. The flow field is captured using a video camera. The captured images are then converted into concentration data with the aid of neural network. To this end, the use of four different networks with varying input data was investigated. The estimations were validated using concentration measurements conducted by a light absorption probe.

Download article, 19 pages

Article price - $35.00

<< Previous article   Next article >>