A key enabler of intelligent maintenance systems is the ability to predict the remaining useful lifetime (RUL) of its components, i.e., prognostics. The development of data-driven prognostics models requires datasets with run-to-failure trajectories. However, large representative run-to-failure datasets are often unavailable in real applications because failures are rare in many safety-critical systems. To foster the development of prognostics methods, we develop a new realistic dataset of run-to-failure trajectories for a fleet of aircraft engines under real flight conditions. The dataset was generated with the Commercial Modular Aero-Propulsion System Simulation (CMAPSS) model developed at NASA. The damage propagation modelling used in this dataset builds on the modelling strategy from previous work and incorporates two new levels of fidelity. First, it considers real flight conditions as recorded on board of a commercial jet. Second, it extends the degradation modelling by relating the degradation process to its operation history. This dataset also provides the health, respectively, fault class. Therefore, besides its applicability to prognostics problems, the dataset can be used for fault diagnostics.

The problem of Uncertainty Quantification (UQ) is of paramount importance when Machine Learning (ML) and Deep Learning (DL) models are deployed in the real world. In the context of safety-critical applications, such as the prediction of the Remaining Useful-Life (RUL) of infrastructure and industrial assets, the relevance of effective UQ approaches is even higher, given the potentially catastrophic consequences or substantial costs associated with maintenance decisions that are performed either too late or too early. Particularly in such safety critical application contexts, transparency and reliability are essential requirements and any ML-based solution not providing meaningful uncertainty estimates would not fully satisfy such desiderata. However, most ML and DL techniques used for RUL estimation are often not designed to perform UQ, thus limiting their effective applicability in real-world scenarios. To address this limitation, in this paper we investigate the performance of a recently proposed class of algorithms, Deep Gaussian Process (DGPs), for predicting the remaining useful lifetime (RUL). DGPs provide uncertainty estimates associated with their predictions, yet retaining the expressive power and learning capabilities of modern DL methods. DGPs are able to scale to very large datasets, which was a limitation of some of the previous algorithms but has increasingly become an essential requirement for many industrial applications and, in particular, for RUL prediction tasks for which UQ is of key importance. The main contribution of this paper is a thorough evaluation and comparison of several variants of DGPs applied to the problem of RUL prediction. The performance of the DGP algorithms is evaluated on the N-CMAPSS (New Commercial Modular Aero-Propulsion System Simulation) dataset from NASA for aircraft engines. The results demonstrate that DGPs are able to provide very accurate RUL predictions along with sensible uncertainty estimates, providing more reliable solutions for (safety-critical) real-life industrial applications.

Lithium-ion (Li-I) batteries have recently become pervasive and are used in many physical assets. For the effective management of the batteries, reliable predictions of the end-of-discharge (EOD) and end-of-life (EOL) are essential. Many detailed electrochemical models have been developed for the batteries. Their parameters are calibrated before they are taken into operation and are typically not re-calibrated during operation. However, the degradation of batteries increases the reality gap between the computational models and the physical systems and leads to inaccurate predictions of EOD/EOL. The current calibration approaches are either computationally expensive (model-based calibration) or require large amounts of ground truth data for degradation parameters (supervised data-driven calibration). This is often infeasible for many practical applications. In this paper, we introduce a reinforcement learning-based framework for reliably inferring calibration parameters of battery models in real time. Most importantly, the proposed methodology does not need any labeled data samples of observations and the ground truth parameters. The experimental results demonstrate that our framework is capable of inferring the model parameters in real time with better accuracy compared to approaches based on unscented Kalman filters. Furthermore, our results show better generalizability than supervised learning approaches even though our methodology does not rely on ground truth information during training.

With the increased availability of condition monitoring data on the one hand and the increased complexity of explicit system physics-based models on the other hand, the application of data-driven approaches for fault detection and isolation has recently grown. While detection accuracy of such approaches is generally very good, their performance on fault isolation often suffers from the fact that fault conditions affect a large portion of the measured signals thereby masking the fault source. To overcome this limitation, we propose a hybrid approach combining physical performance models with deep learning algorithms. Unobserved process variables are inferred with a physics-based performance model to enhance the input space of a data-driven diagnostics model. The resulting increased input space gains representation power enabling more accurate fault detection and isolation. To validate the effectiveness of the proposed method, we generate a condition monitoring dataset of an advanced gas turbine during flight conditions under healthy and four faulty operative conditions based on the Aero-Propulsion System Simulation (C-MAPSS) dynamical model. We evaluate the performance of the proposed hybrid methodology in combination with two different deep learning algorithms: deep feed forward neural networks and Variational Autoencoders, both of which demonstrate a significant improvement when applied within the hybrid fault detection and diagnostics framework. The proposed method is able to outperform pure data-driven solutions, particularly for systems with a high variability of operating conditions. It provides superior results both for fault detection as well as for fault isolation. For the fault isolation task, it overcomes the smearing effect that is commonly observed in pure data-driven approaches and enables a precise isolation of the affected signal. We also demonstrate that deep learning algorithms provide a better performance on the fault detection task compared to the traditional machine learning algorithms.

Understanding the current health condition of complex systems and their temporal evolution is an important step in prognostics and health management (PHM). However, when managing a fleet of complex systems, variations arising from manufacturing, environmental factors, mission profiles, and maintenance practices result in diverse health index (HI) trajectories. Therefore, in PHM, it is essential not only to identify common fleet-wide trends but also to account for individual asset-level variations when inferring HIs. While several data-driven approaches exist for inferring individual asset-level HIs from unsupervised run-to-failure degradation data, little research has been devoted to deriving analytical probabilistic representations of HIs that encompass both fleet-level trends and individual asset-level fluctuations. This paper aims to bridge this gap by addressing the research question of how to obtain an analytical representation of probability distributions for the time to reach intermediate degradation levels, using run-to-failure data or incomplete degradation trajectories from a fleet of complex systems. In this work, it is assumed that suitable, asset-specific HI curves have been inferred through a fusion of deep learning techniques and prior expert knowledge of degradation physics . Given this context, we derive an analytical probabilistic description of the health index (HI) that reflects both fleet-wide trends and asset-specific conditions in the cases of Gamma or Weibull time-to-failure (TTF) distributions. Our approach involves defining HIs with a power law function, enabling the modeling of TTF and time to reach intermediate degradation levels. Moreover, we also detail the procedure for estimating the power law exponent from field data through regression analysis and conduct a sensitivity analysis regarding this exponent. To illustrate our methodology, we present two case studies based on the N-CMAPPS dataset of turbofan engines and Li-ion batteries, validating the aforementioned assumptions and illustrating our methodology steps