Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Numerical and machine learning analyses of entropy generation in an unsteady squeezing flow of copper/aluminum oxide/water hybrid nanofluid
Several studies have been carried out on the squeezing flow of mono‐nanofluids. However, this study investigates numerical and machine learning analyses of entropy generated in an unsteady squeezing flow of a copper–aluminum oxide/water hybrid nanofluid. Numerical analysis and the flow model simulation are done using the hybridization of Chebyshev pseudospectral and quasilinearization methods. The results show that the magnetic force is the most significant in the entropy generation number. A support vector machine learning model is introduced to calculate the average entropy generation number. The machine learning model's accuracy is measured using known performance metrics of regression models. The root mean square error is obtained as 4.1184, the mean absolute error as 1.8776, and the coefficient of determination () as 0.995. Furthermore, the Hartman and Eckert numbers are identified to be highly positively correlated to the entropy generation number. However, for increasing values of the Hartman number, the temperature distribution between the two parallel plates decreases. We found that the temperature of copper/aluminum oxide/water nanofluid is greater than that of copper/water nanofluid, and the percentage difference between the temperature at the lower plate is estimated to be between 6% and 9% for .
Numerical and machine learning analyses of entropy generation in an unsteady squeezing flow of copper/aluminum oxide/water hybrid nanofluid
Several studies have been carried out on the squeezing flow of mono‐nanofluids. However, this study investigates numerical and machine learning analyses of entropy generated in an unsteady squeezing flow of a copper–aluminum oxide/water hybrid nanofluid. Numerical analysis and the flow model simulation are done using the hybridization of Chebyshev pseudospectral and quasilinearization methods. The results show that the magnetic force is the most significant in the entropy generation number. A support vector machine learning model is introduced to calculate the average entropy generation number. The machine learning model's accuracy is measured using known performance metrics of regression models. The root mean square error is obtained as 4.1184, the mean absolute error as 1.8776, and the coefficient of determination () as 0.995. Furthermore, the Hartman and Eckert numbers are identified to be highly positively correlated to the entropy generation number. However, for increasing values of the Hartman number, the temperature distribution between the two parallel plates decreases. We found that the temperature of copper/aluminum oxide/water nanofluid is greater than that of copper/water nanofluid, and the percentage difference between the temperature at the lower plate is estimated to be between 6% and 9% for .
Numerical and machine learning analyses of entropy generation in an unsteady squeezing flow of copper/aluminum oxide/water hybrid nanofluid
Oloniiju, Shina D. (Autor:in) / Oyelakin, Ibukun S. (Autor:in) / Goqo, Sicelo P. (Autor:in) / Sibanda, Precious (Autor:in)
Heat Transfer ; 50 ; 3822-3841
01.06.2021
20 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch