Neural network modeling of flow hydrodynamics in a nozzle column

• Main Authors: Kondrateva, M. I., Bronskaya, V. V., Mukhametzyanova, A. G., Ignashina, T. V., Khairullina, L. E., Balzamov, D. S., Bashkirov, D. V., Garifullina, E. V., Kharitonova, O. S.
• Format: Conference Proceeding
• Language: English
• Subjects: Chemical engineering; Devices; Distillation; Dust collectors; Emissions; Fluid flow; Gas mixtures; Low pressure; Neural networks; Nozzles; Pressure drop; Roughness
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0182086

Nozzle columns are used in modern chemical, petrochemical, oil and gas and other industries, as well as for processes of cooling and separation of liquid and gas mixtures. A characteristic feature of devices with a nozzle with regular roughness is their versatility, simple design, low pressure drops and high efficiency, these devices are equally good to use in the processes of absorption, distillation, condensation, contact cooling, and as gas dust collectors. The main application of these devices is found in chemical engineering, where large volumes of gas emissions and gas cleaning processes require a large amount of liquid pumping. The purpose of the work is to study the hydrodynamic characteristics of devices with a nozzle with the regular roughness, to obtain dependencies for calculating the coefficients of hydraulic resistance. In this paper, the approach using a neural network is effective and reliable that is proved by numerical experiment in the comparison. Neural networks are developed and modeled using Wolfram Mathematica to predict the resistance of the nozzle column.