Artificial neural network based predictive modeling of viscosity of an oil based hybrid nanofluid

In many industries like micro-electronics, mechanical engines, transportation, manufacturing and nuclear reactors cooling characteristics which is a crucial challenge can optimize using nanofluid. Nanofluids are also used for better energy transfer and storage in many devices. In the current study the viscosity variation studied by the impact of input variables solid volume fraction and temperature. Sol-gel auto-combustion was used to prepare nickel ferrite (NiFe₂O₄) nano-particles. The structure of the NiFe₂O₄ was obtained using an X-ray diffraction method. The Fourier transform infrared spectrum provided evidence of NiFe₂O₄ presence. In order to better understand the surface morphology of nickel ferrite nano-particles, scanning electron microscopy (SEM) was employed. AT the various concentrations of nanofluid 0, 0.25, 0.50, 0.75, and 1 % were prepared. The viscosity was calculated at temperatures ranges 40 to 80 °C. The experimental results showed that the viscosity of nanofluid (µ_nf) falls as the volume fraction(φ) increases and as the temperature rises. RSM and ANN predicted data have R² values of 0.9987 and 0.9999 respectively.. ANN accuracy can also be seen by MSE value. ANN shows less MSE value (0.00001) as compare to RSM that has MSE 0.0002. Which shows the ANN accuracy over RSM. Predictive modelling for viscosity of Oil-Based Hybrid Nanofluid. Artificial Neural Network and Response surface methodology methods were used for correlation. Results were correlated with experimental data.