Performance of joined artificial neural network and genetic algorithm to study the effect of temperature and mass fraction of nanoparticles dispersed in ethanol

Abstract

There are many experimental results in the field of finding nanofluids in optimal conditions, which are useful and effective in using artificial neural network methods for better analysis of these results. In this study, by using an artificial neural network, this paper checked the effect of temperature and concentration of different types of nanoparticles, such as copper oxide or Multi Wall Carbon Nano Tubes (MWCNT), on thermal conductivity and the interaction of nanofluid particles. The temperature and mass fraction of the nanoparticles are considered as the inputs of the neural network, and the thermal conductivity and molecular interaction as the outputs of this network are considered. The experimental data extracted for this study change in four temperatures of 25°C, 40°C, 55°C, and 70°C, as well as the mass fraction of nanoparticles in the range of 0.007%–7%. This paper considered the best network with a hidden layer, 15 neurons, and the Levenberg–Marquardt training method. The results showed that at a concentration between 2% and 4% of nanoparticles and at higher temperatures than 60°C, maximum thermal conductivity and molecular interaction occurred. Nevertheless, considering that the maximum heat conductivity combined with minimum fluid interaction with nanofluid (for more stability) is our desire, so this paper used a genetic algorithm to optimize the results obtained.

Develop artificial neural network numerical modeling to study fluid flow and heat transfer of dispersed nanoparticles through base liquid

“…Then, the 3D graph of heat transfer was plotted. A curve was fitted on this graph with the “Orthogonal Distance Regression (ODR)” algorithm and ANN model was reported, Nguyen et al (2020a, 2020b). For the TC and the TCR 3 D-graphs, these models were calculated for the both nanofluids.…”

“…Carbon nanotube (CNT) can be used in industrial applications. The CNT has remarkable mechanical and thermo-electrical features (Tlili, 2019; Afridi et al , 2019; Nguyen et al , 2020a, 2020b; Goodarzi et al , 2015; Akbari et al , 2016). Predict the nanofluid properties, first needs to the preparation and stabilization a homogeneous mixture; then the suitable exact devices would be needed to predict the thermo-physical properties (Goshayeshi et al , 2016a, 2016b; Alrashed et al , 2018; Goodarzi et al , 2016; Xu et al , 2020; Jiang et al , 2019; Safaei et al , 2019; Ranjbarzadeh et al , 2017; Tlili, 2019; Du et al , 2020).…”