Phase change materials in fluid–structure interaction within microchannel: A numerical and deep learning approach

Microchannels have revolutionized fields such as chemical analysis, drug delivery, and diagnostics, by offering enhanced control over fluid flows at microscale dimensions. The behavior of fluids in microchannels, governed by principles like viscous and capillary forces, differs significantly from macroscopic flows, and these differences have implications for fluid–structure interaction as well. Microchannel geometries often involve thin walls, making them susceptible to deformation under fluid-induced stresses. This interplay between fluid flow and structural mechanics introduces complex phenomena such as fluid-induced vibrations, instabilities, and alterations in flow patterns. This study evaluates fluid–structure interaction involving nano-encapsulated phase change materials in the existence of a flexible fine structure at the base of a microchannel. The boundaries of microchannel are supposed rigid, while the fin is elastic. The versatile finite element method is implemented to numerically evaluate the partial differential equation of phase change materials fluid flow and heat transfer. A grid refinement and validation test are conducted using a convolutional neural network, and a convolutional neural network with bidirectional gated recurrent unit assures the accuracy of the numerical results. Moreover, the results demonstrate that suspending the viscosity of nano-encapsulated phase change materials in water significantly enhances heat transfer during fluid–structure interaction within the microchannel.