LED Lifetime Prediction Under Thermal-Electrical Stress
Kai Zhe Tan, See-Keong Lee, H.G. Low
Abstract
The lifetime of light-emitting diodes (LEDs) is highly dependent on usage conditions such as operating temperature and drive current. It is costly and impractical to test the LEDs on every usage condition. Previous studies used the Arrhenius equation and Black's model to investigate the relationship of operating temperature, drive current and lumen maintenance life. However, they are found to be less accurate for certain conditions. The study aims to improve the prediction of lumen maintenance life under different thermal-electrical conditions. The Eyring model is proposed in this study. The model parameters are determined by regression approach, which provides the goodness of fit of the prediction model as well as the prediction interval. Furthermore, a method to predict lumen depreciation trend for different operating conditions based on the Eyring model and regression approach is established. As a result, the Eyring model produced higher prediction accuracy compared to Arrhenius equation and Black's model.