Artificial neural network optimized by Bayesian-regularization algorithm to solve the convective flow of trihybrid nanofluid with LTNECs and thermal radiation

The aim of this study is to examine an artificial intelligence neural network based on the Intelligent Bayesian-Regularization optimizer to evaluate the effect of thermal radiation on the trihybrid nanofluid inside a stenotic artery with LTNECs (local thermal non-equilibrium conditions). The built network used the log-sigmoid and radial basis activation functions. Artificial intelligence neural networks are trained on 80 % of the data, with the remaining 20 % being used for testing. Blood serves as both a base fluid and a trihybrid nanofluid that contains gold, silver, and titanium oxide nanoparticles. Simulating blood flow in damaged arteries can aid in the diagnosis and treatment of cardiovascular conditions. Additionally, this classic has the possible to enhance cancer treatments and targeted drug delivery, specifically hyperthermia therapy, and improve the design of biomedical devices that depend on heat and mass transmission in complex fluids. The appropriate similarity variable method has been used to convert PDEs into dimensionless ODEs. After that, the ODEs are solved both numerically and visually with MATLAB's built-in bvp4c solver. We chose the Bayesian-Regularization optimizer because it enhances generalization, decreases overfitting, and maintains consistent convergence even with sparse or noisy data, which is crucial for biological applications.