A Gaussian Process Based Method for Data- Efficient Remaining Useful Life Estimation

Maximilian Benker,Michael F Zaeh,Artem Bliznyuk

doi:10.1109/access.2021.3116813

Maximilian Benker, Michael F Zaeh + Show 1 more

Open Access

https://doi.org/10.1109/access.2021.3116813

Copy DOI

Abstract

The task of remaining useful life (RUL) estimation is a major challenge within the field of prognostics and health management (PHM). The quality of the RUL estimates determines the economical feasibility of the application of predictive maintenance strategies, that rely on accurate predictions. Hence, many effective methods for RUL estimation have been developed in the recent years. Especially deep learning methods have been among the best performing ones setting new record accuracies on bench mark data sets. However, those approaches often rely on numerous and representative run-to-failure sequences of the components under investigation. In real-world use cases, this kind of data (i.e. run-to-failure sequences and RUL labels) is hardly ever present. Therefore, this paper proposes a new, data-efficient method, which is based on Gaussian process classification to derive abstract health indicator (HI) values in a first step, and warped, monotonic Gaussian process regression for indirect RUL estimation in a second step. The proposed approach does neither rely on entire run-to-failure sequences nor on any RUL labels and was tested on the benchmark C-MAPSS turbo fan and FEMTO bearing data sets, achieving comparable results to the state-of-the art whilst using only a small fraction of the available training data. Hence, the proposed approach allows RUL estimation in use cases, in which gathering enough failure data for the application of deep learning models is infeasible.

Highlights

With increasing digitization of production processes and the introduction of Industry 4.0 into factories all over the world, access to additional economic profit by further automating and flexibilizing the industrial processes is expected [1], [2]
This work aims at addressing the shortcomings in the state of the art by proposing a novel approach, which can be trained with only a small amount of healthy and degraded observations, making run-to-failure sequences needless, and which applies a warped, monotonic GPR model to extrapolate sensible health indicator (HI) values within the interval [0, 1] in order to accurately estimate remaining useful life (RUL)
In addition to the HI produced by the GPC model, results produced by simple linear regression and logistic regression models are shown

Summary

Introduction

With increasing digitization of production processes and the introduction of Industry 4.0 into factories all over the world, access to additional economic profit by further automating and flexibilizing the industrial processes is expected [1], [2]. The topic of prognostics and health management (PHM), which forms the basis of deploying predictive maintenance strategies [3], is a prominent use case of Industry 4.0, since it offers the avoidance of unnecessary and unplanned failures of industrial assets This leads to higher equipment efficiencies, less downtime and lower costs due to disturbances in global supply chains [4]. Many authors have addressed the task of RUL estimation in many different ways, which can be categorized into physicsbased approaches, statistical approaches and artificial intelligence (AI) approaches [6] The latter two have been subject of extensive investigations within the last few years.

Objectives

Methods

Results

Conclusion