Domain Adaptation via Alignment of Operation Profile for Remaining Useful Lifetime Prediction

TL;DR

This paper introduces two novel domain adaptation methods that improve remaining useful life predictions by aligning operation profile phases separately, addressing domain shifts in different operating conditions.

Contribution

The paper proposes phase-aware adversarial domain adaptation techniques for RUL prediction, enhancing accuracy across diverse operational domains.

Findings

Improved RUL prediction accuracy over state-of-the-art methods.

Effective alignment of operation phases in source and target domains.

Validated on N-CMAPSS dataset with different flight classes.

Abstract

Effective Prognostics and Health Management (PHM) relies on accurate prediction of the Remaining Useful Life (RUL). Data-driven RUL prediction techniques rely heavily on the representativeness of the available time-to-failure trajectories. Therefore, these methods may not perform well when applied to data from new units of a fleet that follow different operating conditions than those they were trained on. This is also known as domain shifts. Domain adaptation (DA) methods aim to address the domain shift problem by extracting domain invariant features. However, DA methods do not distinguish between the different phases of operation, such as steady states or transient phases. This can result in misalignment due to under- or over-representation of different operation phases. This paper proposes two novel DA approaches for RUL prediction based on an adversarial domain adaptation framework that considers the different phases of the operation profiles separately. The proposed methodologies align the marginal distributions of each phase of the operation profile in the source domain with its counterpart in the target domain. The effectiveness of the proposed methods is evaluated using the New Commercial Modular Aero-Propulsion System (N-CMAPSS) dataset, where sub-fleets of turbofan engines operating in one of the three different flight classes (short, medium, and long) are treated as separate domains. The experimental results show that the proposed methods improve the accuracy of RUL predictions compared to current state-of-the-art DA methods.