A Hypergraph Neural Network-Based Surrogate Model for Predicting the Durability of OLED Displays under Drop Impact

This study presents a hypergraph neural network–based surrogate model to predict the durability of layered OLED displays under drop impact. Finite element explicit dynamics simulations are used to generate training data, capturing time-dependent deformation and stress during impact. By modeling interactions between two colliding bodies using a hypergraph structure, the proposed model accurately predicts the full impact process using only initial conditions. The results demonstrate high accuracy compared to a commercial finite element solver while significantly reducing computational cost, enabling efficient and real-time durability evaluation.

Korean Society of Mechanical Engineers (KSME 2024)