SVAE‐GRU‐based degradation generation and prediction for small samples

The degradation process analysis is the basis of the system reliability assessment. Too few degradation samples will affect the accuracy of the reliability analysis. So, it is necessary to expand the degradation data and construct the degradation prediction model. This paper proposes an SVAE‐GRU‐bas...

Full description

Saved in:
Bibliographic Details
Published in:Quality and reliability engineering international Vol. 39; no. 7; pp. 2851 - 2868
Main Authors: Shangguan, Anqi, Feng, Nan, Mu, Lingxia, Fei, Rong, Hei, Xinhong, Xie, Guo
Format: Journal Article
Language:English
Published: Bognor Regis Wiley Subscription Services, Inc 01.11.2023
Subjects:
Degradation
Effectiveness
Heterogeneity
Prediction models
Reliability analysis
System reliability
ISSN:0748-8017, 1099-1638
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The degradation process analysis is the basis of the system reliability assessment. Too few degradation samples will affect the accuracy of the reliability analysis. So, it is necessary to expand the degradation data and construct the degradation prediction model. This paper proposes an SVAE‐GRU‐based degradation generation and prediction model to handle the issues. New degradation data are generated by combining the Stacked Variational Autoencoder (SVAE) and Gated Recurrent Units (GRU) in each time step, which can improve the learning effectiveness of the degradation features. The GRU and the Multi‐layer perceptron calculate the mean and the variance of the degradation distribution for each timestamp, the deeply heterogeneity and temporal feature of the original degraded samples are learned. Then the degradation amount of different samples are predicted. The optimal hyper‐parameters of the SVAE‐GRU are obtained by the grid search method. To verify the effectiveness of the proposed method, three datasets and four methods are used for experiment comparison in this paper. The actual example indicate that the SVAE‐GRU method has the best effect on the degradation generation and prediction.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0748-8017
1099-1638
DOI:10.1002/qre.3394