Flow Pattern Identification of Horizontal Two-Phase Refrigerant Flow Using Neural Networks
Mechanical and Thermal Systems Branch, Power and Control Division Wright-Patterson Air Force Base United States
Pagination or Media Count:
In this work, electrical capacitance tomography ECT and neural networks were used to automatically identifytwo-phase flow patterns for refrigerant R-134a flowing in a horizontal tube. In laboratory experiments, highspeedimages were recorded for human classification of liquidvapor flow patterns. The correspondingpermittivity data obtained from tomograms was then used to train feedforward neural networks to recognizeflow patterns. An objective was to determine which subsets of data derived from tomograms could be used asinput data by a neural network to classify nine liquidvapor flow patterns. Another objective was to determinewhich subsets of input data provide high identification successwhen analyzed by a neural network. Transitionalflow patterns associatedwith common horizontal flow patternswere considered. A unique feature of the currentwork was the use of the vertical center of mass coordinate in pattern classification. The highest classificationsuccess rates occurred using neural network input which included the probability density functions in timefor both spatially averaged permittivity and center of mass location in addition to the four statistical momentsin time for spatially averaged permittivity data. The combination of these input data resulted in an averagesuccess rate of 98.1 for nine flow patterns. In addition, 99 of the experimental runs were either correctlyclassified or misclassified by only one flow pattern.