Cova-PINN: Cross-Domain Conservation Physics-Informed Neural Network for Fluid-Solid Conjugate Heat Transfer in Complex Geometries
Cova-PINN is a multi-domain Physics-Informed Neural Network designed to solve complex fluid–solid conjugate heat transfer problems in intricate geometries. It addresses conservation gaps by coupling local energy balances around heat transfer walls with global constraints across the entire device scale. This approach significantly reduces errors in outlet temperature and overall energy transfer compared to standard methods.