Title :
Searching Failure Distribution of Sample Based on Slope-Mode Distance
Author :
Hang Luo ; Hou-Jun Wang ; Bing Long ; Jian-Guo Huang
Author_Institution :
Autom. Eng. Coll., Univ. of Electron. Sci. & Technol. of China, Chengdu
Abstract :
In reliability engineering, how to determine failure distribution of object according to failure data is an important and practical work. In this paper, the shortcoming of determining failure distribution in intuition was pointed out, which was based on leaf-figure maneuver, and an idea of quantitating similarity between theoretical curve and testing curve to determine distribution of sample was presented. Then, a method called weighted slope-mode distance was used to define similarity between two curves. At the same time, how to analysis similarity and realize it with method of weighed slope-mode distance were analyzed in detail. At last, several examples were taken to test validity to presented method. The result of experiment showed that the similarity method based on weighed slope-mode distance could quantitate shape difference between two curves, and offer a quantitative standard for determining distribution of failure data.
Keywords :
failure analysis; reliability; failure data; failure distribution; leaf-figure maneuver; reliability engineering; weighted slope-mode distance; Automation; Data mining; Discrete wavelet transforms; Educational institutions; Failure analysis; Optimization methods; Reliability engineering; Shape; Testing; Yarn;
Conference_Titel :
Testing and Diagnosis, 2009. ICTD 2009. IEEE Circuits and Systems International Conference on
Conference_Location :
Chengdu
Print_ISBN :
978-1-4244-2587-7
DOI :
10.1109/CAS-ICTD.2009.4960869