Evaluation of MWCNTs-ZnO/5W50 nanolubricant by design of an artificial neural network for predicting viscosity and its optimization

• Published in: Journal of molecular liquids 2019-03, Vol.277, p.921-931
• Main Authors: Hemmat Esfe, Mohammad, Goodarzi, Mohsen, Reiszadeh, Mahdi, Afrand, Masoud
• Format: Article
• Language: English
• Subjects: Artificial neural network (ANN); Mathematical correlation; MWCNTs-ZnO(10%–90%)/5W50; Nanolubricant viscosity; Optimization
• ISSN: 0167-7322; 1873-3166
• DOI: 10.1016/j.molliq.2018.08.047

This research presents the design of an artificial neural network (ANN) and experimental evaluation of MWCNTs-ZnO(10%–90%)/5W50 nanolubricant at different temperatures and shear rates, and presentation of a mathematical correlation to predict viscosity and its optimization. The nanofluid experimental evaluation was carried out at the solid volume fractions of 0.05, 0.1, 0.25, 0.5, 0.75 and 1% and the temperature range of 5 to 55 °C. Nanofluid viscosity optimization was performed with respect to temperature, volume fraction, and shear rates. A point at the temperature of 54.29 °C, solid volume fraction of 0.1%, and shear rate of 1029.89 (1/s) had the optimal minimum viscosity of 38.1654 mPa·s. The ANN designed for the nanofluid included two hidden layers with an optimal structure with 3 neurons in the first layer and 3 neurons in the second layer. The value of R for this neural network was 0.9998057. In the final stage, ANN data have an error lower than 7%. This research reports the ANN model parameters. •Designing an artificial neural network with errors lower than 7%•Proposing a three variable correlation with R-squared of 0.9467•One factor and two factor analysis to investigate effect of parameters on viscosity•Temperature was determined as the most effective parameters on viscosity.