Machine learning-assisted computational analysis of radiated viscoelastic tri-hybrid nanofluid flow past a Riga plate with Cattaneo-Christov model

The thermal performance of the solar collectors is enhanced by the addition of nanomaterial to the radiative surface, which increase its stability and lifetime. Due to its numerous applications, this study utilizes machine learning framework to simulate the energy and hydrodynamic behavior of a time-dependent Prandtl tri-hybrid nanofluid (HNF) over a radiated Riga surface. The Cattaneo–Christov heat flux model, mixed convection, radiant heat, dissipating, and magnetic effects are considered. A Crank-Nicolson computational method (CNM) is used to solve the controlling pressure-driven, unsteady flow equations, and the generated dataset is used to assess and train the artificial neural network (ANN) model. he generated solution data are used to train a feed-forward artificial neural network (ANN) with tan-sigmoid hidden activation, and Levenberg–Marquardt training, using a dataset of 320 samples. The results show that increasing the thermal relaxation parameter reduces the peak temperature by up to about 14%, mitigating thermal stress, whereas stronger thermal radiation and higher nanoparticle concentration raise the temperature by up to about 26%. A higher modified Hartmann number lowers the velocity by roughly 10–17% while slightly increasing the temperature, and a larger Grashof number enhances the velocity by nearly 22%. The ANN attains minimum validation mean-squared-error values of \(1.35\times 10^{-9}\) for velocity and \(1.25\times 10^{-9}\) for temperature, overall prediction errors within ±3.8%, and regression coefficients exceeding 0.9996.