A hybrid prognostic & health management framework across multi-level engineering systems with scalable convolution neural networks and adjustable functional regression models

Kaigan Zhang, Tangbin Xia, Yuhui Xu, Yutong Ding, Yong Zhao, Nagi Gebraeel, Lifeng Xi

Abstract: Highlights•An efficient data-driven structure with scalable time sequences and kernel sizes is developed.•A novel model-based regression model is formulated to provide adjustable capabilities.•A hybrid PHM framework that integrates data-driven RUL and model-based TTF is proposed.•We comprehensively improve hybrid PHM frameworks to adapt to various system hierarchies.