ANN implementation in predicting the entropy optimized heat and mass transfer of an EMHD trihybrid nanofluid past a stretching sheet with variable viscosity, viscous dissipation and convective boundary conditions

Conventional heat transfer fluids often fail to meet the thermal demands of advanced industrial systems. This study investigates electromagnetohydrodynamic (EMHD) flow of a trihybrid nanofluid containing copper, gold and silver nanoparticles dispersed in water over a porous stretching surface. The model incorporates temperature-dependent viscosity, convective heating, velocity slip, suction, viscous and thermal dissipation, buoyancy, electromagnetic forces, internal heat generation, mass diffusion and chemical reactions. The governing nonlinear partial differential equations are transformed to ordinary differential equations using a suitable similarity transformation and then solved using the fourth-order Runge-Kutta shooting method. Entropy generation and Bejan number analyses to show thermodynamic performance. An Artificial Neural Network (ANN) employing the Scaled Conjugate Gradient (SCG) algorithm is used to predict skin friction, Nusselt number and Sherwood number. A dataset of 200 points is used and divided into 70% training, 15% testing and 15% validation. The ANN demonstrates high predictive accuracy with overall correlation coefficients of 0.97209, 0.97951 and 0.97080 for skin friction, Nusselt number and Sherwood number, respectively. Results show that radiation has little effect on entropy generation under variable viscosity conditions but decreases the Bejan number. The investigation confirms the potential of trihybrid nanofluids better choice for nanotechnology and energy applications.