A similarity-based method for remaining useful life prediction based on operational reliability

• Published in: Applied intelligence (Dordrecht, Netherlands) Vol. 48; no. 9; pp. 2983 - 2995
• Main Authors: Zeming, Liang, Jianmin, Gao, Hongquan, Jiang, Xu, Gao, Zhiyong, Gao, Rongxi, Wang
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.09.2018; Springer Nature B.V
• Subjects: Artificial Intelligence; Computer Science; Degradation; Life prediction; Machines; Manufacturing; Mechanical Engineering; Processes; Reliability; Similarity; Useful life; Remaining useful life prediction; Similarity; Operational reliability; Degradation degree; Degradation index
• ISSN: 0924-669X; 1573-7497
• DOI: 10.1007/s10489-017-1128-4

Remaining useful life (RUL) prediction is an effective method to ensure system safety and avoid catastrophe in applied intelligent perspective. A similarity-based method for RUL prediction based on operational reliability has been proposed in this paper to overcome the shortcomings existing in the conventional similarity-based RUL prediction method of devices. The first shortcoming is the construction difficulty of degradation index (DI) and the second shortcoming is the low accuracy of similarity measurement. First, a novel operational reliability fused index (ORI) was proposed by calculating offset distance and offset angle between current state and normal state of devices, and the index was used as the DI for life prediction. Then, a multi-factor similarity measurement method was proposed to select the most similarity samples and a novel similarity measurement and weight assigned method considering degradation degree was proposed based on the previous similarity measurement. Finally, the RUL was weighted predicted. The effectiveness of the proposed method was validated by implementing two cases.