ABSTRACT
Nanofluids exhibit superior thermal conductivity and present higher specific surface area compared to traditional colloidal suspensions. Moreover, nanofluids are easily transportable in flow paths and have better long term stability. Experimental determination of viscosity of nanofluids usually is a difficult, time consuming, and expensive task. Many theoretical models have been developed for prediction of viscosity of nanofluids. Due to disability of the theoretical model in prediction of viscosity of nanofluids, researchers have developed many empirical correlations based on their own experimental data. Modeling viscosity of nanofluids is a complicated task because of two main reasons. Firstly, viscosity of nanofluids is dependent on many variables, and, secondly, the relationships between these variables are very complex. The most accurate and general model for estimation of viscosity of nanofluids was developed by A. Hemmati-Sarapardeh et al. based on committee machine intelligent system. An artificial neural network based on genetic algorithm was proposed by A. Karimi et al. to estimate viscosity of nanofluids.
