Collaborative System Failure Prognostics via Federated Longitudinal-Survival Modeling

2026-07-28

AI Summary

This study addresses the challenge of system failure prediction in distributed settings, where data privacy and proprietary constraints hinder the integration of multi-source sensor measurements and time-to-failure information. To overcome this limitation, the authors propose a federated vertical-survival modeling framework that, for the first time, incorporates a separable discrete-time hazard function into federated learning. By combining local temporal representation learning on client devices with global collaborative optimization, the method enables cross-institutional reliability modeling and remaining useful life (RUL) estimation without sharing raw data. This approach circumvents the intractability of directly optimizing traditional Cox models under federated settings. Experiments on the four C-MAPSS turbofan engine datasets demonstrate that the proposed method significantly outperforms locally trained models and achieves performance approaching that of centralized training, even under heterogeneous operating conditions and diverse failure modes.

Abstract

Time-to-event modeling provides a systematic framework for estimating time-dependent failure risk, reliability, and remaining useful life (RUL) from longitudinal condition monitoring data. However, applying these models to distributed prognostics remains challenging because sensor trajectories and failure-time records are often stored across organizations or operational sites and cannot be centrally pooled due to privacy or proprietary constraints. Moreover, the classical Cox proportional hazards model relies on a nonseparable partial likelihood involving global risk sets, making direct optimization difficult under standard federated learning protocols. This paper presents a federated longitudinal-survival modeling framework for collaborative system failure prognostics. The proposed framework combines longitudinal sensor representation learning with a client-separable discrete-time hazard objective, enabling multiple clients to collaboratively train a prognostic model without sharing raw sensor measurements or individual failure records. Time-dependent representations extracted from multivariate sensor histories are used to estimate interval-specific failure hazards, reliability curves, and system RUL. Experiments on the four C-MAPSS turbofan engine degradation subsets under simulated decentralized settings demonstrate that the proposed framework consistently improves prognostic performance over isolated local training while maintaining performance comparable to centralized training across heterogeneous operating conditions and failure modes. These results demonstrate the potential of federated longitudinal-survival modeling for collaborative, data-aware condition monitoring and system failure prognostics.

Problem

• federated learning
• survival analysis
• system failure prognostics
• longitudinal data
• privacy-preserving

Innovation

• federated learning
• longitudinal-survival modeling
• failure prognostics
• remaining useful life
• discrete-time hazard