A fuzzy-driven neural network framework for predicting heat transfer in triangular fuzzy hybrid nanofluids

This work considers the analysis of the fuzzy flow phenomenon under the influence of nanoparticles on the heat transmission of hybrid nanofluid (Al2O3+Cu/EO) streaming over a shrinking/stretching Riga wedge via artificial neural networks (ANNs). The key factors contributing to the flow include heat source, nonlinear thermal radiation, and stagnation point. The two nanomaterials (Al2O3 and Cu) are embedded in host fluid engine oil. Since the fluid flow phenomenon frequently involves imprecise measurements and uncertainties, fuzzy set theory (FST) accounts for this vagueness. To make comparison of hybrid nanofluids (Al2O3+Cu/EO) and nanofluids Cu/EO, (Al2O3/EO) by engaging membership function (MF), the nanoparticle concentration is adopted as a triangular fuzzy number (TFN) 0,0.1,0.2. The σ-cut and membership functions are bounded TFN in a range [0,1]. This fuzzy analysis highlights that hybrid nanofluid (Al2O3+Cu/EO) offers more heat transfer in comparison to nanofluids Cu/EO (Al2O3/EO). To analyze the phenomenon, an artificial neural network approach based on the Bayesian Regularization scheme (BRS) and Levenberg–Marquardt scheme (LMS) is integrated comparatively to evaluate initial data retrieved for quantities of physical importance. The tabular and graphical data present an outstanding match between the predicted ANN values and the targeted values.