Peter Zeiler’s research while affiliated with Esslingen University of Applied Sciences and other places

Data-driven methods have been shown to be suitable for the diagnosis and prognosis of the health of engineering systems. However, the training of data-driven methods usually requires a large amount of data, which is rarely available in industry. In addition, the prediction accuracy often degrades when operating or environmental conditions change or when there are similar systems with different technical characteristics. Transfer learning offers the possibility to transfer knowledge about the degradation behavior between such systems. However, there is a risk that the degradation behavior differs too much, which leads to a so-called negative transfer. Therefore, the authors argue that a similarity assessment of degradation behavior is essential. An assessment based on the operational data of systems seems particularly appropriate. In this paper, the suitability of time series similarity measures for such data-based similarity assessments is investigated. The current state of research is presented. Thereby, no studies on the similarity comparison of degradation curves of engineering systems using time series similarity measures are found. Furthermore, measures for assessing the similarity of degradation curves are identified and categorized. In a case study on filter clogging curves, similarity tests are performed using these measures to find the most similar time series. Various approaches are proposed for evaluation, two of which are used in this paper. The results show that mostly a good selection is made, with some measures performing particularly well. The work presented in this paper represents valuable groundwork for the use of time series similarity measures in transfer learning.

Being able to predict the remaining useful life (RUL) of an engineering system is an important task in prognostics and health management. Recently, data-driven approaches to RUL predictions are becoming prevalent over model-based approaches since no underlying physical knowledge of the engineering system is required. Yet, this just replaces required expertise of the underlying physics with machine learning (ML) expertise, which is often also not available. Automated machine learning (AutoML) promises to build end-to-end ML pipelines automatically enabling domain experts without ML expertise to create their own models. This paper introduces AutoRUL, an AutoML-driven end-to-end approach for automatic RUL predictions. AutoRUL combines fine-tuned standard regression methods to an ensemble with high predictive power. By evaluating the proposed method on eight real-world and synthetic datasets against state-of-the-art hand-crafted models, we show that AutoML provides a viable alternative to hand-crafted data-driven RUL predictions. Consequently, creating RUL predictions can be made more accessible for domain experts using AutoML by eliminating ML expertise from data-driven model construction.

Vorwort Sehr geehrte Damen und Herren, durch die zunehmende Digitalisierung, Autonomisierung und Vernetzung von technischen Systemen, beispielsweise in einer Smart Factory im Kontext von Industrie 4.0 oder in autonomen Fahrzeugen, werden hohe Anforderungen an die Zuverlässigkeit, die Verfügbarkeit und die Sicherheit dieser Systeme gestellt. Dies erfordert den konsequenten Einsatz und die ständige Weiterentwicklung von Methoden und Modellen der Zuverlässigkeitstechnik entlang des gesamten Lebenszyklus zur Planung, Entwicklung und Absicherung der Zuverlässigkeit. Die zunehmende Digitalisierung bietet durch die steigende Zugänglichkeit und Verfügbarkeit von relevanten Daten gleichzeitig enorme Chancen und neue Möglichkeiten für die Anwendung dieser Methoden und Modelle für Zuverlässigkeitsanalysen und -prognosen. Der vorliegende Tagungsband enthält die Manuskripte der Referenten und Referentinnen und der Posterreferenten. Der Tagungsleiter und die Mitglieder des Programmausschusses danken allen, die zum Gelingen der Veranstaltung mitwirken. Die VDI Wissensforum GmbH und die VDI-Gesellschaft Produkt- und Prozessgestaltung (GPP) führen die Tagung Technische Zuverlä...

Approaches for diagnosis and prognosis of the health of engineering systems are divided into data-driven, model-based, and hybrid methods. Data-driven methods depend on the availability of data. Model-based methods require knowledge of the degradation process. A great effort for data generation along with the high complexity of degradation processes often limits both approaches. To mitigate these limitations, the combination of data and knowledge through hybrid methods is examined in this paper. This approach is compared to the alternative approach of reducing the effort for generating training data, as both are gaining importance in diagnostics and prognostics. A new categorization of hybrid prognostic methods for combining data-driven and physics-based models is presented, along with references to existing realizations of these methods. Based on the categorization, a case study on the hybrid remaining useful life prediction of a filtration process is conducted. Several hybrid methods are implemented and tested in this study. Through the combination of models, an improvement in predictive accuracy is achieved. In addition, the paper examines systematic attributes of the individual hybrid methods. Statements on the influence of data scarcity on the predictive accuracy, data-driven models with high variance, and the computational efficiency of the hybrid methods are made. It is shown that these statements are supported by the case study’s results.

Vorwort Sehr geehrte Damen und Herren, durch die zunehmende Digitalisierung, Autonomisierung und Vernetzung von technischen Systemen, beispielsweise in einer Smart Factory im Kontext von Industrie 4.0 oder in autonomen Fahrzeugen, werden hohe Anforderungen an die Zuverlässigkeit, die Verfügbarkeit und die Sicherheit dieser Systeme gestellt. Dies erfordert den konsequenten Einsatz und die ständige Weiterentwicklung von Methoden und Modellen der Zuverlässigkeitstechnik entlang des gesamten Lebenszyklus zur Planung, Entwicklung und Absicherung der Zuverlässigkeit. Die zunehmende Digitalisierung bietet durch die steigende Zugänglichkeit und Verfügbarkeit von relevanten Daten gleichzeitig enorme Chancen und neue Möglichkeiten für die Anwendung dieser Methoden und Modelle für Zuverlässigkeitsanalysen und -prognosen. Der vorliegende Tagungsband enthält die Manuskripte der Referenten und Referentinnen und der Posterreferenten. Der Tagungsleiter und die Mitglieder des Programmausschusses danken allen, die zum Gelingen der Veranstaltung mitwirken. Die VDI Wissensforum GmbH und die VDI-Gesellschaft Produkt- und Prozessgestaltung (GPP) führen die Tagung Technische Zuverlä...

In Prognostics and Health Management, there are three main approaches for implementing diagnostic and prognostic applications. These approaches are data-driven methods, physical model-based methods, and combinations of them, in the form of hybrid methods. Each of them has specific advantages but also limitations for their purposeful implementation. In the case of data-driven methods, one of the main limitations is the availability of sufficient training data that adequately cover the relevant state space. For model-based methods, on the other hand, it is often the case that the degradation process of the considered technical system is of significant complexity. In such a scenario physics-based modeling requires great effort or is not possible at all. Combinations of data-driven and model-based approaches in form of hybrid approaches offer the possibility to partially mitigate the shortcomings of the other two approaches, however, require a sufficiently detailed data-driven and physics-based model. This paper addresses the transitional field between data-driven and hybrid approaches. Despite the issues of formulating a physics-based model that provides a representation of the degradation process, basic knowledge of the considered system and of the laws governing its degradation process is usually available. Integration of such knowledge into a machine learning process is part of a research field that is either called theory-guided data science, (physics) informed machine learning, physics-based learning or physics guided machine learning. First, the state of research in Prognostics and Health Management on methods of this field is presented and existing research gaps are outlined. Then, a concept is introduced for incorporating fundamental knowledge, such as monotonicity constraints, into data-driven diagnostic and prognostic applications using approaches from theory-guided data science. A special aspect of this concept is its cross-application usability through the consideration of knowledge that repeatedly occurs in diagnostics and prognostics. This is, for example, knowledge about physically justified boundaries whose compliance makes a prediction of the data-driven model plausible in the first place.

Prognostics and health management (PHM) is an engineering approach dealing with the diagnosis, prognosis, and management of the health state of engineering systems. Methods that can analyze system behavior, fault conditions, and degradation are crucial for PHM applications, as they create the basis for determining, predicting, and monitoring the health of engineering systems. Data-driven methods have been proven to be suitable for automated diagnosis or prognosis due to their pattern recognition and anomaly detection abilities. Moreover, they do not require knowledge of the underlying degradation process. However, training data-driven methods usually requires a large amount of data, whose collection, cleansing, organization, and preparation are generally very time-consuming and costly. There are usually little or no run-to-failure data available at market launch, especially for new systems such as new machine generations. Nevertheless, related systems, hereinafter referred to as similar systems, often already exist, differing only in some technical characteristics. In this paper, the similar system problem is defined, and explanations of the difficulties that arise when using data from similar systems are presented. Furthermore, it is discussed why the usage of these data offers great potential for condition diagnosis and prognosis of engineering systems. An overview of data-driven methods that can be used to utilize data from similar systems is provided, and the methods that such systems already consider are highlighted. Two related research areas are identified, namely, fleet learning and transfer learning. In the paper, it is shown that similar system approaches will become an important branch of research in PHM. However, some difficulties must be overcome.