Machine learning-integrated computational analysis on a blood-based ternary nanofluid flow in a stenosed artery with the artificial neural networks and modified finite difference approach

Heart failure and stroke continue to be the most common cause of global death, with atherosclerosis artery stenosis becoming a significant contributor. Although prior research has progressed in comprehending blood circulation behaviour, however, the incorporation of artificial intelligence (AI) and machine learning (ML) in the examination of ternary nanofluids for stenosed arterial diseases signifies a notable breakthrough in this domain. We proposes an integration of ML technique with computational fluid dynamics (CFD) to analyse the non-linear dynamics of thermal characterization in blood-based tri-hybrid nano-fluid flow, influenced by thermal radiation, variable heat sources and sinks, and aligned magnetic field effects within a blood artery exhibiting cosine stenosis. The investigation is based on AI approach, the Levenberg–Marquardt algorithm (LMA), with back propagation Artificial Neural Network (ANN-BP). The mathematical model is developed in the form of partial differentia equations and transformed into ordinary differential equation by similarity scaling, and then numerically evaluated by a modified finite difference approach, the Keller-Box method. The suggested ANN-LMA accuracy is compared to the ML solution for boundary layer flow. Regression values indicate an excellent fit between the predictions and the real data. It is observed that the inclined magnetic angle affects the drag force and heat transfer rate. There is a 27.9% increase in the heat transfer rate for ternary nano-fluid. Conclusively, the non-linear interaction between the magnetic field and nanofluid flow may significantly enhance heat transfer rates, which could have potential applications in biomedical sciences.