48‐1: A Data‐Driven Intelligent Stress Monitoring for a Robust Manufacturing of a Phone Display with the Extremely Narrow Bottom Bezel

Abstract

A data‐driven intelligent stress monitoring methodology is proposed for a robust manufacturing of a phone display with the extremely narrow bottom bezel. For the narrow bottom bezel, a bottom of a display panel should be bent 180 degrees with asymmetric bended protrusion by tight attachment. The mechanical stress of the bended panel increases rapidly depending on the amount attached after bending 180 degrees. For stable operation of assembly process with high risk process method, the virtual stress monitoring are applied for total product samples during the process. The intelligent stress sensor produces the virtual stress/strain data using a surrogated model, which is learned with the dataset of measured data and simulated data of the tightly bending attachment using it as input. The virtual stress/strain data which matches 99.4% of the simulation value with regression error of less than 1% can be obtained in real time from all products during assembly process with bending of the display panel. The crack risks of the bending 180 degrees are monitored in real time for anomaly detection using the virtual stress data produced by the proposed virtual sensor. With these approaches, we believe that the manufacturing processes having high risks could be stably controlled for a real infinite display creating an incredibly immersive viewing experience.