Predicting long-term lumen maintenance life of LED light sources using a particle filter-based prognostic approach

作者：

Highlights：

• Destructive life test is time-consuming and expensive to estimate the LED’s life.

• TM-21 standard with least-squares regression has weakness in predicting LED’s life.

• A dynamic recursive method of PF is developed to model the lumen degradation data.

• An SMC method is proposed to predict RUL distribution with a confidence interval.

• PF has higher accuracy than TM-21 standard in the LED’s long-term life prediction.

摘要

•Destructive life test is time-consuming and expensive to estimate the LED’s life.•TM-21 standard with least-squares regression has weakness in predicting LED’s life.•A dynamic recursive method of PF is developed to model the lumen degradation data.•An SMC method is proposed to predict RUL distribution with a confidence interval.•PF has higher accuracy than TM-21 standard in the LED’s long-term life prediction.

论文关键词：LED light sources,Lumen maintenance,Remaining useful life,Prognostics,TM-21 standard,Particle filter

论文评审过程：Available online 18 October 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.10.021