Navigating viscosity of ferrofluid using response surface methodology and artificial neural network

Abstract

The main purpose of this study is to investigate the capabilities of artificial neural network (ANN) and response surface methodology (RSM) in estimating the viscosity of Fe3O4 nanofluid. Nanoparticles increase the resistance to motion and thus boost the viscosity. Initially, the rheological behavior of the base fluid and nanofluid was investigated and it was found that both fluids are not particularly sensitive to the shear rate, which indicates the Newtonian behavior. Input parameters of temperature and volume fraction and output parameter, nanofluid viscosity were introduced to both techniques to find the best correlation in which the viscosity can be predictable. Comparison of R-square in ANN (0.999) and RSM (0.996) techniques showed that both techniques can navigate the viscosity well. Also the margin of deviation (MOD) and mean square error (MSE) for ANN were 4.22% and 0.0000741 which were lower than the corresponding values in RSM one (MOD = 5.52%, MSE = 0.00027).