Prediction of Micropolar Fluid Flow Characteristics in a Stenosed Bifurcated Artery Using Feed-forward Neural Networks Trained by the Levenberg Marquardt Algorithm

The present article aims to investigates the characteristics of flow of blood through a porous bifurcated artery by assuming blood as a micropolar fluid. The effects of porous wall and branching of artery are assimilated to account for the physiological conditions includes permeability and resistance to flow. Artery under consideration is assumed to be symmetric about it’s own axis and circular cylinders of restricted length. Partial differential equations governing the flow are formulated with reference to micropolar fluid framework in cylindrical polar co-ordinate system. These equations are made non-dimensional and coordinate transformation is used to convert the irregular geometry to a regular. The transformed equations along with concern boundary conditions are initially solved by finite difference scheme. A feedforward neural network is trained utilizing the Levenberg - Marquardt backpropagation method to model responses for various parameter alterations on a reference dataset produced by the numerical solver. Present study concentrates on the influence of Darcy number on physical parameters. These results shows enhanced permeability and lower resistance, resulting in higher velocity, increased flow rate, more stable shear stress distribution, and decreased impedance, thereby resembling healthy arterial flow behavior. Findings of this work provide potential implications for understanding pathological conditions such as atherosclerosis and stenosis in porous arterial segments.