TY - JOUR

T1 - Computer Vision Algorithm for Characterization of a Turbulent Gas–Liquid Jet

AU - Starodumov, Ilya

AU - Sokolov, Sergey

AU - Mikushin, Pavel

AU - Nikishina, Margarita

AU - Mityashin, Timofey

AU - Makhaeva, Ksenia

AU - Blyakhman, Felix

AU - Chernushkin, Dmitrii

AU - Nizovtseva, Irina

N1 - The research funding from the Ministry of Science and Higher Education of the Russian Federation (Ural Federal University Program of Development within the Priority-2030 Program) is gratefully acknowledged.

PY - 2024

Y1 - 2024

N2 - A computer vision algorithm to determine the parameters of a two-phase turbulent jet of a water-gas mixture traveling at a velocity in the range of 5–10 m/s was developed in order to evaluate the hydrodynamic efficiency of mass exchange apparatuses in real time, as well as to predict the gas exchange rate. The algorithm is based on threshold segmentation, the active contours method, the regression of principal components method, and the comparison of feature overlays, which allows the stable determination of jet boundaries and is a more efficient method when working with low-quality data than traditional implementations of the Canny method. Based on high-speed video recordings of jets, the proposed algorithm allows the calculation of key characteristics of jets: the velocity, angle of incidence, structural density, etc. Both the algorithm’s description and a test application based on video recordings of a real jet created on an experimental prototype of a jet bioreactor are discussed. The results are compared with computational fluid dynamics modeling and theoretical predictions, and good agreement is demonstrated. The presented algorithm itself represents the basis for a real-time control system for aerator operation in jet bioreactors, as well as being used in laboratory jet stream installations for the accumulation of big data on the structure and dynamic properties of jets.

AB - A computer vision algorithm to determine the parameters of a two-phase turbulent jet of a water-gas mixture traveling at a velocity in the range of 5–10 m/s was developed in order to evaluate the hydrodynamic efficiency of mass exchange apparatuses in real time, as well as to predict the gas exchange rate. The algorithm is based on threshold segmentation, the active contours method, the regression of principal components method, and the comparison of feature overlays, which allows the stable determination of jet boundaries and is a more efficient method when working with low-quality data than traditional implementations of the Canny method. Based on high-speed video recordings of jets, the proposed algorithm allows the calculation of key characteristics of jets: the velocity, angle of incidence, structural density, etc. Both the algorithm’s description and a test application based on video recordings of a real jet created on an experimental prototype of a jet bioreactor are discussed. The results are compared with computational fluid dynamics modeling and theoretical predictions, and good agreement is demonstrated. The presented algorithm itself represents the basis for a real-time control system for aerator operation in jet bioreactors, as well as being used in laboratory jet stream installations for the accumulation of big data on the structure and dynamic properties of jets.

UR - http://www.scopus.com/inward/record.url?partnerID=8YFLogxK&scp=85185910409

UR - https://gateway.webofknowledge.com/gateway/Gateway.cgi?GWVersion=2&SrcAuth=tsmetrics&SrcApp=tsm_test&DestApp=WOS_CPL&DestLinkType=FullRecord&KeyUT=001171991300001

U2 - 10.3390/inventions9010009

DO - 10.3390/inventions9010009

M3 - Article

VL - 9

JO - Inventions

JF - Inventions

SN - 2411-5134

IS - 1

M1 - 9

ER -