AI-augmented flow of hybrid nanofluid with magnetic and activation energy effects via various neural networks: an ANN–fuzzy logic integration

In this study, an analysis of the impacts of MHD, Joule heating, and activation energy on Blasius-Sakiadis flow with hybrid nanofluid (Ag–Cu/EG) is carried out. The transformation of nonlinear differential equations is done through the process of similarity transformation of nonlinear governing equations. Furthermore, for the consideration of uncertainty that occurs due to the variation in nanoparticle volume fraction, triangular fuzzy number of fuzzy logic is adopted. Further, ANN models are designed using Levenberg–Marquardt, Bayesian regularization, and scaled conjugate gradient algorithms. From this analysis, it is observed that an increase in magnetic parameter reduces the velocity due to the presence of Lorentz force but increases the temperature due to the impact of Joule heating effect. The accuracy of ANN predictions is very high as indicated by R values approximating 0.999 and MSE values between $${10}^{-6}$$ and $${10}^{-17}$$ . The Bayesian Regularization algorithm yields superior results in terms of generalization ability. The Bayesian approach gives the most consistent results. The artificial neural network-based fuzzy model offers a systematic procedure in dealing with hybrid nanofluids with uncertainties and is applicable to other nonlinear flow processes as well.