A Comparison of Residual-based Methods on Fault Detection


Loading...

Date

2023-10-26

Publication Type

Conference Paper

ETH Bibliography

yes

Citations

Altmetric

Data

Abstract

An important initial step in fault detection for complex indus trial systems is gaining an understanding of their health con dition. Subsequently, continuous monitoring of this health condition becomes crucial to observe its evolution, track changes over time, and isolate faults. As faults are typically rare occurrences, it is essential to perform this monitoring in an unsupervised manner. Various approaches have been proposed not only to detect faults in an unsupervised manner but also to distinguish between different potential fault types. In this study, we perform a comprehensive comparison be tween two residual-based approaches: autoencoders, and the input-output models that establish a mapping between oper ating conditions and sensor readings. We explore the sensor wise residuals and aggregated residuals for the entire sys tem in both methods. The performance evaluation focuses on three tasks: health indicator construction, fault detection, and health indicator interpretation. To perform the comparison, we utilize the Commercial Modular Aero-Propulsion System Simulation (C-MAPSS) dynamical model, specifically a sub set of the turbofan engine dataset containing three different fault types. All models are trained exclusively on healthy data. Fault detection is achieved by applying a threshold that is determined based on the healthy condition. The detection results reveal that both models are capable of detecting faults with an average delay of around 20 cycles and maintain a low false positive rate. While the fault detection performance is similar for both models, the input-output model provides better interpretability regarding potential fault types and the possible faulty components.

Publication status

published

Book title

Proceedings of the Annual Conference of the PHM Society 2023

Journal / series

Volume

15 (1)

Pages / Article No.

Publisher

PHM Society

Event

15th Annual Conference of the Prognostics and Health Management Society (PHM 2023)

Edition / version

Methods

Software

Geographic location

Date collected

Date created

Subject

condition monitoring; run-to-failure; anomaly detection; n-cmapss; fault detection

Organisational unit

03869 - Franck, Christian / Franck, Christian check_circle
02632 - Inst. f. El. Energieübertragung u. Hoch. / Power Systems and High Voltage Lab.

Notes

Funding

Related publications and datasets

Is new version of: 10.48550/arXiv.2309.02274