Experimental study of thermal conductivity coefficient of GNSs-WO3/LP107160 hybrid nanofluid and development of a practical ANN modeling for estimating thermal conductivity

Abstract

In the present study, the effects of nanoparticles, mass fraction percentage and temperature on the conductive heat transfer coefficient of Graphene nanosheets- Tungsten oxide/Liquid paraffin 107160 hybrid nanofluid was investigated. For this purpose, four different mass fractions were used in the range of 0.005%–5% in a number of examinations. The results illustrated that the thermal conductivity coefficient was increased with the increment of the mass fraction percentage and the temperature of Graphene nanosheets- Tungsten oxide nanomaterials in the base fluid. Then, a feed-forward artificial neural network was used to model the thermal conductivity coefficient. In general, with the increase in temperature and concentration of nanofluid, the value of thermal conductivity increases. The optimum value of thermal conductivity for this experiment was observed in the volume fraction of 5% and at the temperature of 70 °C. The results of this modeling indicated that the fault of the data estimated for the coefficient of thermal conductivity in the Graphene nanosheets- Tungsten oxide/Liquid paraffin 107160 nanofluid, as a function of mass fraction percentage and temperature, was less than 3%, as compared to the experimental data.

Keywords

• Thermal conductivity coefficient
• Hybrid nanofluid
• Graphene nanosheets
• Tungsten nanoparticles
• Liquid paraffin
• Feed-forward neural network