Lifetime estimation of LED lamp using gamma process model

• The degradation path models for LED were presented schematically and compared to each other by examples.
• The lifetime was estimated as statistical distribution expressing the sampling and temporal uncertainty of degradation.
• By using gamma process model, the warranty life can be suggested from optimum replacement policy.
• The gamma process model can also be a method to estimate the lifetime of LED for chromaticity shift.

The data for the light flux degradation and chromaticity shift of LED lamps exhibits a considerable amount of scatter due to intrinsic and extrinsic factors. In this study, some degradation models, such as the gamma process model, were reviewed in terms of uncertainties associated with the continuous, gradual, and monotonic nature of degradation. Statistically varying light flux degradation and chromaticity shift data in a test report from Lumileds were used as an example to demonstrate the use of the gamma process model. This model can describe the condition and lifetime as statistical distribution curves whose shapes vary with operation time. The service life was estimated as a median value, while the warranty life was estimated as a B2.5 life for lumen degradation and B25 life for chromaticity shift from consideration of the optimal replacement life and percentile life.