Designing a neuron network topology for forecasting streaming patterns of highly magnetized couple stress trihybrid nano-blood in a constricted/dilated artery by means of electric double layer principle

Abstract

Exploring electrokinetic phenomena in the context of biofluids blended with nanoparticles opens up possibilities for innovative biomedical applications. It may open pathways to novel treatments involving electromagnetic fields and nanoparticle-infused blood tailored to target specific medical issues. This research is centred on investigating streaming patterns in highly magnetized couple stress blood infused with single-walled carbon nanotubes (SWCNTs), titania, and alumina within a constricted/dilated arterial channel influenced by electroosmosis. The model is designed with the imposition of an intense external magnetic field oriented perpendicularly to the channel, giving rise to phenomena such as Hall currents, ion-slip currents, and Joule heating. The electric potential within the electric double layer (EDL) is calculated through the solution of the Poisson-Boltzmann equation. Computation of the proposed blood flow model is accomplished by harnessing the Runge-Kutta-Fehlberg (RKF45) shooting scheme via the bvp4c solver in Mathematica. The study is carried out by employing informative graphs and tables to comprehensively understand and elucidate the physical sequels of crucial factors on flow dynamics and physical quantities. Notable outcomes include the observation that blood temperature regenerates with higher values of Hall and ion-slip parameters, while skin friction on the artery’s upper wall abates as the electroosmosis parameter amplifies. An increase in nanoparticles’ volume fraction (NVF) leads to higher Nusselt numbers on the upper wall. Furthermore, an artificial neural network (ANN) model is developed using reference datasets obtained from numerical outcomes. This ANN model accurately predicts crucial flow quantities, with error rates of only 0.07% for the skin friction coefficient and 0.05% for the Nusselt number. The findings of this study hold potential implications for the design of more effective drug delivery systems, biomedical devices, improved diagnostic accuracy and informed treatment decisions, among other applications.