THERMAL TEMPERATURE ESTIMATION BY MACHINE LEARNING METHODS OF COUNTERFLOW RANQUE-HILSCH VORTEX TUBE USING DIFFERENT FLUIDS
-
Published:2023
Issue:12
Volume:54
Page:61-79
-
ISSN:1064-2285
-
Container-title:Heat Transfer Research
-
language:en
-
Short-container-title:Heat Trans Res
Author:
Korkmaz Murat,Doğan Ayhan,Kırmacı Volkan
Abstract
In the counterflow Ranque-Hilsch vortex tube (RHVT), the output control valve on the hot fluid side is left entirely open. The data were obtained using polyamide and brass materials and nozzles at 50 kPa intervals from 150 kPa to 700 kPa inlet pressure. In counterflow RHVT, the difference (ΔT) between the temperature of the cold outflow and the temperature of the outgoing hot flow was found, and the RHVT was modeled. The deficiency in the literature was tried to be eliminated. In this study, we planned the modeling of a counterflow RHVT using compressed air, oxygen, and nitrogen gas with machine learning models to predict the thermal temperature. Linear regression (LR), support vector machines (SVM), Gaussian process regression (GPR), regression trees (RT), and ensemble of trees (ET) machine learning methods were preferred in this study. While each of the machine learning methods in the study was analyzed, 75% of all data was used as training data, 25% as a test, 65% as training data, and 35% as testing data. As a result of the analysis, when the temperatures of air, oxygen, and nitrogen gases (ΔT) were compared, the Gaussian process regression method, which is one of the machine learning models, gave the best result with 0.99 in two different test intervals, 75-25%, and 65-35%. In the ΔT estimations made in all fluids, much better results were obtained in the machine learning models estimations of nitrogen gas when compared to other gases.
Subject
Fluid Flow and Transfer Processes,Mechanical Engineering,Condensed Matter Physics
Reference36 articles.
1. Aslan, Y., Yavasca, S., and Yasar, C., Long Term Electric Peak Load Forecasting of Kutahya Using Different Approaches, Int. J. Tech. Phys. Probl. Eng., vol. 3, no. 2, pp. 87-91, 2011. 2. Behera, U., Paul, P.J., Kasthurirengan, S., Karunanithi, R., Ram, S.N., Dinesh, K., and Jacob, S., CFD Analysis and Experimental Investigations towards Optimizing the Parameters of Ranque-Hilsch Vortex Tube, Int. J. Heat Mass Transf., vol. 48, no. 10, pp. 1961-73, 2005. DOI: 10.1016/j.ijheatmasstransfer.2004.12.046 3. Chen, J., Gao, X., Shao, S., Hu, H., Xie, J., Li, N., and Gao, N., Numerical Investigation of the Vortex Tube Performance in Novel Precooling Methods in the Hydrogen Fueling Station, Int. J. Hydrogen Energy, vol. 46, no. 7, pp. 5548-5555, 2021. DOI: 10.1016/j.ijhydene.2020.11.070 4. Cristianini, N. and Shawe-Taylor, J., An Introduction to Support Vector Machines and Other Kernel-Based Learning Methods, Cambridge, UK: Cambridge University Press, 2000.
|
|