Prognostics-based qualification of high-power white LEDs using Lévy process approach

Due to their versatility in a variety of applications and the growing market demand, high-power white light-emitting diodes (LEDs) have attracted considerable attention. Reliability qualification testing is an essential part of the product development process to ensure the reliability of a new LED product before its release. However, the widely used IES-TM-21 method does not provide comprehensive reliability information. For more accurate and effective qualification, this paper presents a novel method based on prognostics techniques. Prognostics is an engineering technology predicting the future reliability or determining the remaining useful lifetime (RUL) of a product by assessing the extent of deviation or degradation from its expected normal operating conditions. A Lévy subordinator of a mixed Gamma and compound Poisson process is used to describe the actual degradation process of LEDs characterized by random sporadic small jumps of degradation degree, and the reliability function is derived for qualification with different distribution forms of jump sizes. The IES LM-80 test results reported by different LED vendors are used to develop and validate the qualification methodology. This study will be helpful for LED manufacturers to reduce the total test time and cost required to qualify the reliability of an LED product.