Title :
A pre-processing method for degradation parameter
Author :
Wang, Lizhi ; Jiang, Tongmin ; Li, Xiaoyang ; Wang, Xiaohong
Author_Institution :
Dept. of Syst. Eng. of Eng. Technol., Beihang Univ., Beijing, China
Abstract :
In the application of lifetime prediction, sometimes the degradation parameter would be disturbed due to the unstable environment. This phenomenon is negative to the prediction of the lifetime and reliability. So we take Super-luminescent Diode (SLD) as an example, to research the method of degradation parameter pre-processing method under the temperature stress. Firstly, analyze the relationship between degradation variable and unstable environment variable of the parameter. Secondly, remove the unstable environment variable by modeling method based on Support Vector Machines (SVM) and filtering method based on wavelet analysis. Finally, the pre-processing of the degradation parameter is finished, and then the accuracy of the lifetime prediction is improved.
Keywords :
diodes; filtering theory; luminescent devices; reliability; support vector machines; wavelet transforms; SLD; SVM; degradation parameter; filtering method; lifetime prediction; preprocessing method; reliability; super-luminescent diode; support vector machines; temperature stress; wavelet analysis; Data models; Kernel; Optical fibers; Optical filters; Optical sensors; Superluminescent diodes; Super-luminescent Diode; pre-processing; support vector machine; wavelet analysis;
Conference_Titel :
Prognostics and System Health Management (PHM), 2012 IEEE Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4577-1909-7
Electronic_ISBN :
2166-563X
DOI :
10.1109/PHM.2012.6228780