Cross-attention-based bipartite graph neural network for coupled nodal and elemental field prediction in large-deformation sheet material forming
Researchers have developed a novel cross-attention-based bipartite graph neural network (CAtt-BiGNN) to accelerate simulations in large-deformation sheet material forming. This model treats mesh nodes and elements as distinct but interconnected entities, enabling more accurate prediction of nodal displacement and elemental thinning. A hierarchical extension, CAtt-BiUGNN, further enhances performance on larger meshes by incorporating graph downsampling and upsampling techniques. AI
IMPACT This new model could significantly speed up complex engineering simulations, enabling faster design iterations and material development.