Back to Search Start Over

Path classification and estimation model based prognosis of pneumatic cylinder lifetime

Authors :
Xiaoye Qi
Wei Wu
Juan Chen
Chunwen Song
Source :
Chinese Journal of Mechanical Engineering. 25:392-397
Publication Year :
2012
Publisher :
Chinese Journal of Mechanical Engineering, 2012.

Abstract

Prognosis is a key technology to improve reliability, safety and maintainability of products, a lot of researchers have been devoted to this technology. But to improve the predict accuracy of remaining life of products has been difficult. To predict the lifetime specification of pneumatic cylinders with high reliability and long lifetime and small specimen, this paper put forward the prognosis algorithm based on the path classification and estimation (PACE) model. PACE model is based entirely on failure data instead of failure threshold. Pneumatic cylinders normally characterize with failure mechanism wear and tear. Since the minimum working pressure increases with the number of working cycles, the minimum working pressure is chosen as degradation signal. PACE model is fundamentally composed of two operations: path classification and remaining useful life (RUL) estimation. Path classification is to classify a current degradation path as belonging to one or more of previously collected exemplary degradation paths. RUL estimation is to use the resulting memberships to estimate the remaining useful life. In order for verification and validation of PACE prognostic method, six pneumatic cylinders are tested. The test data is analyzed by PACE prognostics. It is found that the PACE based prognosis method has higher prediction accuracy and smaller variance and PACE model is significantly outperform population based prognostics especially for small specimen condition. PACE model based method solved the problem of prediction accuracy for small specimen pneumatic cylinders’ prognosis.

Details

ISSN :
10009345
Volume :
25
Database :
OpenAIRE
Journal :
Chinese Journal of Mechanical Engineering
Accession number :
edsair.doi...........d42b746d731af611040596f1386163a6