A Deep Learning Approach to Predict the Flow Field and Thermal ‎Patterns of Nonencapsulated Phase Change Materials ‎Suspensions in an Enclosure‎

The flow and heat transfer of a novel type of functional phase change nanofluids, nano-encapsulated phase change suspensions, is investigated in the present study using a deep neural networks framework. A deep neural network was used to learn the natural convection flow and heat transfer of the phase change nanofluid in an enclosure. A dataset of flow and heat transfer samples containing 3290 samples of the flow field and temperature distributions was used to train the deep neural network. The design variables were fusion temperature of nanoparticles, Stefan number, and Rayleigh number. The results showed that the proposed combination of a feed-forward neural network and a convolutional neural network as a deep neural network could robustly learn the complex physics of flow and heat transfer of phase change nanofluids. The trained neural network could estimate the flow and heat transfer without iterative and costly numerical computations. The present neural network framework is a promising tool for the design and prediction of complex physical systems.