Develop a Fuzzy System with LM/ODR Artificial Neural Network Model to Predict the Trend of Thermal Conductivity and Rheological Behavior for Graphene Aqueous-Based Nanofluid

Abstract

Graphene is a flexible and transparent conductor which can be used in varied material-apparatus applications, counting solar cells, phones, touch panels, and light-emitting diodes (LED). In this study, to begin with, Graphene aqueous-based nanofluid is processed at separate mass fractions of 0.1 − 0.45 W%. Hence, thermal conductivity of these units was detected at separate temperatures of 25 − 50°C aside the KD2-Pro appliance. Similarly, Rheological behavior at noticed temperatures, for 12.23 and 122.3 S-1 shear rates, was detected aside the DV2EXTRA-Pro appliance. To shorten the expense of research, neural network designs and fuzzy system were trained to discover addition thermal conductivities and viscosities for unalike temperatures and mass fractions. Purpose of this study is to broaden Fuzzy system and ANN algorithms to predict the TC/VIS therefore it predicts the targeted-input dataset as factual as practicable. Hence, the numerical research was accomplished and related aside Levenberg Marquardt and Orthogonal Distance Regression models of Artificial Neural Networks, and Recursive Least Squares Fuzzy system. To train, 14400 data were placed. To test, 2160 ones, to train-control 2160 ones, and to train-output 10080 ones. Conclusions of comparison between algorithms and Fuzzy, exhibited Fuzzy system was fitted on the three-dimensional data more corrected than LM/ODR designs that leads to a better prediction.