A Gamma Process-Based Prognostics Method for CCT Shift of High-Power White LEDs

作者：Wu, Zeyu; Wang, Zili; Feng, Qiang^*; Sun, Bo; Qian, Cheng; Ren, Yi; Jiang, Xiaopeng

来源：IEEE Transactions on Electron Devices, 2018, 65(7): 2909-2916.

DOI：10.1109/TED.2018.2835651

• Correlated color temperature (CCT)
• gamma distribution
• life estimation
• light-emitting diodes (LEDs)
• prediction methods
• reliability
• stochastic processes

The correlated color temperature (CCT) is one of the performance indicators of color shift failure of the high-power white light-emitting diodes (HPWLEDs). The prognostics for CCT shift of HPWLEDs using deterministic or statistical models usually generate a considerable amount of uncertainties by assuming the degradation model of CCT is a deterministic exponential function, which may not be accurate. To overcome this problem, a stochastic process model based on gamma process is established to present the distribution of CCT maintenance lifetime with the probability density function and the cumulative distribution function. The test data reported from LUXEON was used as an example to demonstrate the utilization of the gamma process model. The mean time to failure of which was proved to be more conservative than the L70 lifetime predicted by the conventional method. The CCT shift appears to have a period of nonmonotonic seasoning stage, whereas the gamma process is only applicable to the monotonic degradation. Aiming at this problem, a unified analytical method was established based on generalized Eyring model to estimate seasoning stage under various conditions. The results of which were proved to be reliable.

• 出版日期2018-7
• 单位北京航空航天大学