A hybrid artificial neural network-genetic algorithm modeling approach for viscosity estimation of graphene nanoplatelets nanofluid using experimental data

Predicting the viscosity of graphene nanoplatelets nanofluid with the help of multi-layered perceptron artificial neural network and genetic algorithm was the main aim of this study. In order to achieve the experimental results nanofluid which contains graphene nanoplatelets and deionized water at 20 to 60 °C and 0.025, 0.05, 0.075, and 0.1 wt% is used. Furthermore, genetic algorithm in artificial neural network is used to improve the learning process. In other words, different weights have been chosen for neurons' relations. Also, the bias preoccupation is based on improvements by genetic algorithm. On the other hand, for analyzing the accuracy of the presented model which gives us the nanofluid viscosity predictions MAPE, RMSE, R2, and MBE indexes were used. The values of the presented indexes are 0.777, 0.086, 0.985, − 0.0009 respectively. In case of comparison the results show that the presented model which is the combination of genetic algorithm and artificial neural network is compatible with experimental work.