Application of pattern-recognition methods for the investigation of turbulent flow structure

The applicability of pattern recognition techniques to the analysis of turbulent flow structure is investigated theoretically and experimentally. A flow structure algorithm is developed based on the fundamental equations of pattern recognition theory. The algorithm is used to derive the structural formations (large-scale eddies secondary flows, and intermittent or periodic processes) of intermittent flow in a tube. Values obtained with the algorithm were compared with anemometer measurements obtained with a single-filament probe. Histograms derived from the experimental data are found to be in good agreement with calculations based on the algorithm.