SYSTEM AND PROCEDURE FOR COMBINING DIFFERENTIAL PARTIAL DIFFERENTIAL EQUATION SOLVERS AND NEURAL GRAPH NETWORKS FOR FLUID FLOW PREDICTION

摘要

A computer-implemented method includes receiving a coarse grid input that includes a first set of nodes, the coarse grid input into a computerized fluid dynamics solver with physical parameters to obtain a coarse grid solution, receiving a fine grid input from a second set of nodes where the second set of nodes includes more nodes than the first set of nodes, concatenating the fine mesh input with the physical parameters and performing the concatenation through a graphene convolution layer to obtain a hidden fine mesh layer, increasing the resolution (upsampling) of the coarse mesh solution to obtain a coarse-mesh upsampling including the same number of nodes as the second set of nodes and outputting a prediction in response to at least the coarse-mesh upsampling.