Joint pretraining improves degradation detection across physical systems
CORD: Learning Reusable Degradation Representations Across Heterogeneous Physical Systems
Machine Learning
Summary
Dealing with the wear and tear of machines like bearings, batteries, and cutting tools can be tricky because each system behaves differently. The authors created CORD, a computer approach that learns shared patterns of damage from many types of machines at once. This method helps predict the health of new machines more accurately, even if they weren't part of the initial learning. By training across different devices, CORD finds common signs of degradation that work well in various settings.
What this means in practice
- •For maintenance engineers: Improve prediction of equipment failure by using shared damage patterns learned from multiple machine types during diagnostics.
- •For industrial automation teams: Integrate preread models to better adapt predictive maintenance systems for new machine types without extensive retraining.
Authors
Haibo Li, Zhiguo Zeng
Abstract
Can heterogeneous physical degradation systems benefit from joint pretraining and move beyond system-specific prognostics toward reusable cross-system representation learning? CORD combines type-specific observation interfaces with a shared degradation backbone. Its two self-supervised objectives learn at complementary scales: Intra-Observation Structure Modeling (ISM) captures structure within observations, while Inter-Observation Dynamics Modeling (IDM) captures latent degradation evolution across observation histories. We evaluate CORD under two transfer boundaries: Pretraining-Included System Types, where downstream datasets and held-out units are unseen but their system types are represented during source pretraining, and Pretraining-Excluded System Types, where the entire turbofan-engine type is absent from pretraining. Across bearings, batteries, and cutting tools, CORD (Multi-domain) consistently improves over CORD (Single-domain) under Frozen adaptation, provides further gains under Full FT in most settings, and remains competitive with representative external baselines. Source-pretrained initialization also improves low-label adaptation to the pretraining-excluded engine type. Frozen-representation analysis further shows improved cross-unit lifecycle consistency after multi-domain pretraining. Joint pretraining across heterogeneous physical systems thus produces degradation representations reusable across devices, datasets, and system types.