An ADT data evaluation method of SLD based on Bayesian theory

Abstract

Super luminescent diode (SLD) is a typical product with long lifetime and high reliability which has great advantages and wide application prospects in many areas. Accelerated degradation testing (ADT) is used to obtain performance parameter data in a short time and extrapolate the lifetime and reliability of the products under normal operation conditions. However, in the application of Super luminescent diode (SLD) in ADT, on one hand, sometimes the sample size may be small for the limited cost, the evaluation accuracy of the lifetime and reliability would be low for the lack of information. On the other hand, with the developing of the SLD, usually a fixed number of samples are put into various tests besides ADT and extra information can be obtained in that process. So if all the available information can be used together in a feasible way, the evaluation accuracy of the lifetime and reliability would be improved. In this paper, Bayesian theory is introduced to solve the problems above for its great capability of data fusion. In this method, feasible degradation data of the SLD from different sources are collected, then processed and tested to make them have the same data format and physical meaning, finally fused to estimate the lifetime and reliability with the established degradation statistical model and reliability evaluation model of SLD. This paper provides a new ADT data evaluation method and the method is also suitable for other similar products.