Artificial neural network and sensitivity analysis of heat transfer in Carreau hybrid nanofluid using response surface methodology

The present study integrates the prediction and optimisation of engineering parameters using an artificial neural network (ANN) and response surface methodology (RSM). As hybrid nanofluids have emerged as a powerful tool to enhance thermal performance in industries, automobiles, and energy systems, this study aims to examine the drag force, energy transmission rate, and mass transport of non-Newtonian Carreau Al2O3+MWCNT/ C2H6O2 fluid flow over a wedge. The numerical simulation includes external factors such as thermophoresis, Brownian motion, thermal radiation, activation energy, and chemical reactions to compute Cfx, Nux, Shx, ShAandShB of Al2O3+MWCNT/ C2H6O2 and Al2O3/ C2H6O2. Using the Levenberg–Marquardt backpropagation algorithm, the ANN predicted the engineering factors precisely, with R≅1 and absolute error ranging 10−4−10−3. The findings demonstrate the superiority of hybrid nanofluids over conventional nanofluids, achieving a 12.57% higher energy transmission rate. From the numerical simulation, it is found that increasing the Weissenberg number (We) exhibited higher values of Cfx. Heightening the Brownian motion Nb and thermophoresis (Nt) parameters enhances heat transfer. The activation energy parameter E is boosted in the Shx with decrease in the Lewis number Le. The Schmidt number Sc increases in ShA and decreases in ShB. The Response Surface Methodology is implemented to optimize the Nusselt number using the input parameters Rd,φnp1,np2andNt. The sensitivity analysis for optimization shows that the Nusselt number is highly sensitive to the thermophoresis parameter than to thermal radiation and nanoparticle volume fraction parameters. The RSM model effectively predicts the Nusselt number with an error less than 10−3.