Electroosmotically Driven Casson Blood Flow in Stenosed Arteries Using the Levenberg–Marquardt Algorithm

ABSTRACT

To develop a novel mathematical model of Casson blood flow through a stenosed artery under electroosmotic and thermal transport effects, this study explores a CFD‐based artificial intelligence (AI) solver that employs the Levenberg–Marquardt algorithm within a backpropagated neural network (LMA‐BPNN). This investigation examines the influence of Casson rheology, electroosmosis, and stenosis severity on blood flow through a constricted artery. A mild‐stenosis assumption is used to simplify the governing equations, and the system is solved analytically using the DSolve command in Mathematica. For the proposed LMA‐BPNN model, the dataset is generated using analytical solutions and comprises nine scenarios with various flow properties, including electric field, axial velocity, and heat transfer. of data is utilized for training, for validation, and for testing. Error histograms, performance evaluation, correlation indices, and regression analysis are used to assess the precision and effectiveness of LMA‐BPNN, with MSE ranging from to . ANN‐predicted findings are further presented using graphs and tables for axial velocity and heat transfer. With applications in microfluidics, drug delivery, and biomedical device optimization, results demonstrate the potential of CFD‐AI‐driven solvers as a fast, reliable, and accurate tool for modeling complicated biofluidic systems.