AI-driven oscillatory heat transfer efficiency in MnZnFe2O4-water nanofluid flow over solar collector plate with radiative source and exothermic reaction effects

Oscillatory and steady performance of heat and mass transfer in MnZnFe 2 O 4 -water nanofluid flow over solar-collector plate using exothermic reaction and solar radiation is significant novelty of this analysis. The artificial neural network based Levenberg–Marquardt scheme is applied on fluid temperature to predict the accuracy and convergence of network. The influence of magnetic field, activation energy and mixed convection is applied to predict the performance of fluid temperature and heat-mass transport. The transient model is transformed into primitive based real, imaginary and steady parts using stokes conditions and complex variables. The numerical and graphical results of unknown physical quantities are illustrated using implicit finite difference and Gaussian elimination techniques. The influence of exothermic reaction K R , solar radiation R d , activation energy E A , Schmidt number Sc and Richardson number R iT on steady and oscillatory results is computed. The stronger heat and mass transfer rate is depicted as buoyancy number Br increases. The lowest mean square error, performance, gradient, and damping factor of training network is noticed at 1000 epochs. The amplitude behavior of fluid temperature enhances as exothermic reaction enhances. The stronger isothermal contour values are displayed at higher volume fraction Γ = 0.05. The higher peak in oscillating streamlines and isotherms is depicted due to buoyancy-driven nanoparticle migration. The volume fractions significantly affect the increasing and decreasing oscillation in heat and mass transfer rate.