Materials-oriented computational analysis of fractional magnetohydrodynamic Casson blood–Titanium dioxide nanofluid transport in inclined stenosed arteries

Abstract

Arterial stenosis can lead to severe cardiovascular disease due to the obstruction of normal blood flow. In this study, a fractional-order mathematical model is formulated to investigate magnetohydrodynamic (MHD) Casson blood flow containing TiO $$_2$$ nanoparticles in an inclined stenosed artery using Caputo-Fabrizio derivative (CFD). Flow governing equations for momentum, energy, and concentration are formulated by taking thermal radiation and chemical reactions. Exact analytical solutions of the proposed problem are achieved using Laplace and Hankel transforms (HTs). The impact of main physical parameters such as Hartmann number, fractional order parameter, Casson parameter, thermal radiation parameter, and chemical reaction parameter on the velocity, temperature, and concentration fields is analyzed through graphical and tabular representations. The obtained result shows that the fractional order parameter offers an enhanced fluid flow. With a stronger magnetic field, the velocity distribution is high in the flow, whereas higher values of thermal radiation raises the temperature distribution. In contrast, an enhancement in the amount of chemical reaction parameter reduces the concentration distribution. An artificial neural network (ANN) using the Levenberg-Marquardt (LM) backpropagation algorithm with a 70%, 15%, and 15% data split for training, validation, and testing, respectively, is developed. The ANN model demonstrates a strong computational capability and predictive performance. The developed model offers useful insights to know the transport of nanoparticle-infused blood under different physiological conditions. The current findings would be significantly useful in the design of directed drug delivery systems, magnetic nanoparticle-assisted therapies, hyperthermia treatment for cancer, and also for effective hemodynamic management of patients suffering from stenosed arteries and other cardiovascular problems. The combination of fractional-order modeling and machine learning offers an important approach for future biomedical diagnosis and therapeutic applications.