CORD: Learning Reusable Degradation Representations Across Heterogeneous Physical Systems

2026-09-30

AI Summary

This study addresses the challenge that degradation knowledge in heterogeneous physical systems is difficult to share across domains, confining predictive models to specific systems. To overcome this limitation, we propose a multi-domain joint pretraining framework. Methodologically, we design type-specific observation interfaces coupled with a shared degradation backbone network, and introduce dual self-supervised objectives: intra-observation structural modeling (ISM) to extract internal features, and inter-observation dynamic modeling (IDM) to capture latent degradation evolution patterns, thereby achieving cross-system reusable degradation representation learning. Experimental results demonstrate that the proposed framework significantly outperforms single-domain models in bearing and battery scenarios, effectively enhances frozen adaptation performance, and enables zero-shot transfer to unseen engine types.

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.

Problem

• heterogeneous physical systems
• degradation representation
• cross-system transfer
• prognostics
• self-supervised learning

Innovation

• Cross-system Representation Learning
• Self-supervised Pretraining
• Heterogeneous Physical Systems
• Prognostics
• Domain Adaptation