Modeling of viscosity of composite of TiO2–Al2O3 and ethylene glycol nanofluid by artificial neural network: experimental correlation

Luke O. Ajuka, Moradeyo K. Odunfa, Miracle O. Oyewola, Omolayo M. Ikumapayi*, Stephen A. Akinlabi, Esther T. Akinlabi

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

2 Citations (Scopus)

Abstract

The viscosity of TiO2-Al2O3/EG composite nanofluid was examined using a correlation fitted from experiment and an artificial neural network approach. Using a 15 and 13 nm nominal surface-area weighted diameters sized TiO2 and Al2O3 nanoparticle, respectively, the two-step method was utilized to formulate TiO2-Al2O3/EG nanofluid at 0.04, 0.06, 0.07, 0.08, 0.13, 0.19 and 0.24% volume fractions. The viscosity of the composite as well as the individual nanofluids was experimentally examined. Also, at 0.03% volume fraction, the viscosities of the nanofluids were determined at varied temperature range of 303 to 373 K. Thereafter, theoretical correlation centered on experimental data was developed. Also, a multilayer perceptron neural network was used in predicting the composite nanofluid viscosity as a dependence on nanoparticle volume fraction and temperature. Experimental outcomes show that the nanofluids viscosity increase with increase in nanoparticle volume fraction and decrease with increase in temperature. The mean coefficient of determination relative error for the Levenberg–Marquardt algorithm and the proposed correlation were 0.1% and 4.8%, respectively when compared to the experimental result. This study reveals that the proposed correlation and the artificial neural network have a high predicting ability for the nano-composite viscosity with minimal relative error.

Original languageEnglish
Pages (from-to)1969–1978
Number of pages10
JournalInternational Journal on Interactive Design and Manufacturing
Volume18
Issue number4
Early online date29 May 2022
DOIs
Publication statusPublished - 1 May 2024
Externally publishedYes

Cite this