Physics-informed neural networks for simulating nonlinear peristaltic flow dynamics: A comparative analysis

Abstract

The objective of this study is to examine the peristaltic transport of a Prandtl fluid under complex waveforms, incorporating the effects of viscous dissipation, elastic wall motion, and magnetic field interactions. By applying the long-wavelength approximation, the governing equations are simplified to explore how these physical factors influence velocity distribution, pressure rise, and trapping phenomena. To address this, a Physics-Informed Neural Network scheme is developed to obtain mesh-free solutions, which are validated against the classical BVP4c solver to ensure accuracy and reliability. The study ultimately aims to establish a robust and efficient modeling approach that not only enhances understanding of complex peristaltic flows but also provides practical insights for biomedical applications such as blood transport, drug delivery, and gastrointestinal fluid dynamics, as well as for industrial processes including micro-pumps, implants, and microfluidic devices where non-Newtonian effects and intricate geometries play a critical role.