Intelligent machinery health prognostics under variable operation conditions with limited and variable-length data. (August 2022)

• Record Type: Journal Article
• Authors: Ding, Peng; Jia, Minping; Ding, Yifei; Cao, Yudong; Zhao, Xiaoli
• Is Part Of: Advanced engineering informatics. Volume 53(2022)
• Publication Date: 2022-08
• Subjects: Limited data driven health prognostics -- Encoder-decoder -- Attention mechanism -- Meta learning -- Rotating machinery; Computer-aided engineering -- Periodicals; Engineering -- Data processing -- Periodicals; 620.00285
• DOI: 10.1016/j.aei.2022.101691
• Languages: English
• ISSNs: 1474-0346

Abstract

Highlights: A general variable-length limited data driven prediction method is proposed. Encoder-decoder with attention mechanism framework is deeply integrated with meta -learning. It provides strong generalizations even when the collected data is reduced to one-twentieth. Public and real-measured experiments prove the effectiveness of the established prediction method. Abstract: High accuracy health prognostics are significant to machinery intelligent operation and maintenance. Current data-driven prognostics achieve great success that benefits from amply learning samples. In fact, data scarcity challenge widely exists in machinery prognostics and health management, especially for high-end equipment. This study aims to solve this dilemma and proposes a novel meta learning algorithm reconstructed by classic variable-length prediction mode and attention mechanism, namely meta attention recurrent neural network (MARNN). Specifically, we first develop the encoder-decoder with attention mechanism (EDA) cell to perform episodic learning for the subtask-level upgrade. Then multiple subtasks with EDA as prediction models are aggregated to accomplish meta -level upgrade, thus mining the general degradation knowledge from historical datasets. Finally, cross-domain prognostics tasks can be easily realized through fine-tuning tricks, and three rotating machinery run-to-failed experiments are conducted to prove the generalizations of MARNN, which can obtain desired results even when the on-site adaptation data is reduced to one-twentieth.