CNN-Based Pressure Prediction for a Hybrid Pneumatic–Electromagnetic Bonding System in the OLED Lamination Process

As OLED adoption for large displays continues to grow, there is a need for compact equipment designs that can apply uniform pressure during display and cover glass adhesion. In this study, we propose a multiphysics design methodology using electromagnetic and pneumatic actuators, which offers advantage in terms of actuator size and fast responsiveness. Through experiments, we confirmed that this system improves pressure uniformity by up to 53.7%p. Furthermore, we built a CNN-based image generation model that can help us predict the pressure distribution during the process without high-cost experiments or time-consuming simulations. We first compared the experimental data with the FEM simulation and found that the error is approximately 4.6%. After that, we confirmed that there is an error of less than 6% between the simulation and the CNN model’s predictions.