An improved similarity-based prognostic approach for estimating remaining useful life using gated recurrent units, migratory linear regression, and greedy matching strategy

View at publisher Summarise with VIRAT
Authors: G H Li
Cited by: 0
References: 14

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Abstract Accurate estimation of remaining useful life is essential for condition-based maintenance and system reliability. This paper proposes a data-driven similarity-based prognostic framework for predicting the remaining useful life of aero engines. A gated recurrent unit network is first used to extract degradation features from multi-sensor run-to-failure data and compress them into a one-dimensional health index that represents the degradation trend. Migratory linear regression models are then constructed to preserve health index trajectories and enable similarity comparison between test and training engines. An optimized greedy matching strategy is introduced to improve trajectory alignment and ensure stable remaining useful life estimation. The proposed method is validated on the C MAPSS benchmark dataset. Experimental results demonstrate that the proposed GRU MLR GM approach achieves higher prediction accuracy and lower error can be juxtaposed with modern methods, showing its effectiveness and robustness for remaining useful life prediction.

Fields of Study & Subjects

Machine Fault Diagnosis Techniques Reliability and Maintenance Optimization Technical Engine Diagnostics and Monitoring

Publication details

Journal / source
Engineering Research Express
Volume
8
Issue
7
Pages
075506–075506
Publication date
18 March 2026
Language
EN