PrimaVera Project
No more train delays, power outages, or failure of production machines? The PrimaVera project, funded by the Dutch National Research Agenda (NWA), represents a major step towards this goal. With predictive maintenance, or just-in-time maintenance (maintenance just before a system breaks down), the reliability of infrastructure and production resources can be increased and the costs of maintenance can be reduced.
Existing predictive maintenance techniques only work for small-scale systems and are difficult to scale up. Choices made in one place in the chain have an important influence on other processes in the chain. The choice of a certain type of sensors and measurements influences the type of predictions that can be made, and therefore also the quality of the predictions. That is why cross-level optimization methods are being developed within PrimaVera.
2025 |
Nubia Nale Alves da Silveira; Annemieke Angelique Meghoe; Tiedo Tinga: Quantifying the suitability and feasibility of predictive maintenance approaches. In: Computers & industrial engineering, vol. 194, iss. 110342, 2025. (Type: Journal Article | Abstract | Links | BibTeX)@article{daSilveira2025, Predictive maintenance is a promising concept for maintenance optimization which requires reliable and accurate predictions of a system's lifetime. Despite the many studies on this topic, selecting the best approach is still a topic of debate. In general, predictive maintenance approaches can be roughly classified as knowledge-based, data analytics and physics-based models. However, this classification does not provide maintainers clear guidance on how to select the most suitable approach for a specific case. This work, therefore, presents a selection method for this process. For that purpose, a list of selection criteria was established, and six predictive maintenance approaches were analysed. The proposed selection process is based on two main groups of criteria: the suitability criteria check the match with the desired ambition level of predictive maintenance, while the feasibility criteria identify whether this can be realized, given the labour, models and data available. Finally, three case studies are presented, demonstrating that the tool effectively guides to an optimal approach. |
Luc Stefan Keizers: Hybrid Prognostics For Predictive Maintenance. Combining Physics-Based And Data-Driven Methods To Overcome Prognostic Challenges. 2025, ISBN: 978-90-365-6574-5. (Type: PhD Thesis | Abstract | Links | BibTeX)@phdthesis{Keizers2025, Predictive maintenance is a growing research field, aiming to perform maintenance only when it is required. Prognostic algorithms are essential to achieve this, as predictions of upcoming failures help to increase availability of systems, utilize the full life time of equipment and improve maintenance logistics. Since sensors are getting cheaper and data storage and processing have become cheaper and more efficient during the fourth industrial revolution, there is a lot of interest in data-driven prognostic algorithms. However, high data requirements limit applicability is many practical applications. Physics-based prognostic models yield quantitative relations between system usage and degradation independent from historical data, but the development of such physics-of-failure models is complex and expensive. Because both data-driven and physics-based models have their advantages and limitations, combinations of both types of methods have the potential to get rid of the limitations and profit from the benefits. When both loads and the condition of a component can be monitored, physics-of-failure models can be updated in real-time using Bayesian filtering algorithms. This results in updated quantitative relations between loads and degradation, calibrated for a specific component. The first part of this thesis describes how such methods can be applied for components in variable operating conditions and implements it in a generic prognostic framework. These Bayesian filters preferably receive a direct measure of degradation, such as crack length or the amount of removed material. In many practical applications it is only possible to measure indirect consequences of degradation, such as increased vibration levels, elevated temperatures or acoustic emissions. Therefore, the second part of this thesis focuses on improving quantitative diagnostics to act as input for prognostic algorithms. Because of their modularity and applicability in multiple physical domains, bond graphs are proposed to simulate faults to enhance quantitative diagnostics. A combined diagnostic and prognostics framework is developed which is suitable for prognostics under varying operating conditions, when only limited historical run-to-failure data are available. One of the biggest challenges remains validation of the methods on a real-world case study, as the lack of real-world data is one of the biggest challenges from which this research partly originated. |
Thom Badings: Robust Verification of Stochastic Systems. 2025, ISSN: 2950-2772. (Type: PhD Thesis | Abstract | Links | BibTeX)@phdthesis{Badings2025, Verifying that systems are safe and reliable is crucial in today’s world. For example, we want to prove that an autonomous drone will safely reach its target, or that a manufacturing system will not break down. Classical algorithms for verifying such properties often rely on a precise mathematical model of the system, for example, in the form of a Markov chain or a Markov decision process (MDP). Such Markov models are probabilistic transition systems and are ubiquitous in many areas, including control theory, artificial intelligence (AI), formal methods, and operations research. However, as systems become increasingly complex with more cyber-physical and AI components, uncertainty about the system’s behavior is inevitable. As a result, transition probabilities in Markov models are subject to uncertainty, rendering many existing analysis algorithms inapplicable. In this thesis, we fill this gap by developing novel verification methods for Markov models in the presence of uncertainty. To capture this uncertainty, we use parametric Markov models, where probabilities are described as functions over parameters. We study two perspectives on rendering verification methods robust against uncertainty: (1) set-bounded uncertainty, where only the set of possible parameter values is known and one can be robust in a worst-case sense, and (2) stochastic uncertainty, where the parameter values are described by a probability distribution and one can be robust in a probabilistic sense. By combining techniques from formal methods, AI, and control theory, our contributions span the following general problem settings: 1. We develop robust abstraction techniques for solving control problems for Markov models with continuous state and action spaces, and with set-bounded uncertain parameters. Based on the notion of probabilistic simulation relations, we show that such continuous control problems can be solved by only analyzing a finite-state abstraction, formalized as an MDP with sets of transition probabilities. 2. We present novel and scalable verification techniques for parametric Markov models with a prior distribution over the parameters. Our approaches are sampling-based, do not require any assumptions on the parameter distribution, and provide probably approximately correct (PAC) guarantees on the verification results. 3. We show that parametric models can be used to improve the sample efficiency of data-driven learning methods. We leverage tools from convex optimization to perform a sensitivity analysis, where we measure sensitivity in terms of partial derivatives of the polynomial function that describes a measure of interest. 4. We study continuous-time Markov chains where the initial state must be inferred from a set of (possibly uncertain) state observations. This setting is particularly relevant in runtime monitoring. We compute upper and lower bounds on reachability probabilities, conditioned on these observations. Our approach is based on a robust abstraction into an MDP with intervals of transition probabilities. In conclusion, we develop verification methods that reason over uncertainty without sacrificing guarantees. While dealing with uncertainty can be computationally expensive, providing such guarantees is crucial for designing safe and reliable systems with cyber-physical and AI components. The aim of this thesis is to contribute to a better understanding of dealing with uncertainty in stochastic verification problems. |
Lisandro Arturo Jimenez-Roa: Reliability and Maintenance for Engineering Systems: Fault Trees, Degradation Modelling and Maintenance Optimisation. 2025, ISBN: 978-90-365-6407-6. (Type: PhD Thesis | Abstract | Links | BibTeX)@phdthesis{Jimenez-Roa2025, Modern infrastructures, machines, and manufacturing processes require effective management through sustainable policies under constrained resources, where deter- mining when and how to intervene becomes crucial. The Prognostics and Health Management (PHM) paradigm provides a systematic framework for leveraging data collection and computational models, supporting the management of virtually any engineering component or system. This dissertation delves into three key aspects of PHM: Reliability Modelling, Markov Process-based Prognostics, and Maintenance Optimisation. Data-driven techniques are crucial in these areas, enhancing the automation of model development and deployment. Part I centres on Reliability Modelling, specifically the automatic inference of Fault Tree (FT) models. Traditionally, graph-based models like FTs are manually constructed through iterative collaboration between system experts and FT modellers. However, this manual approach is prone to human error and may result in incomplete models. With the increasing availability of data, methodologies that attempt to automate this process, discover patterns and reduce dependency on manual intervention have gained significant relevance. Thus, in Part I of this dissertation, we focus on how to obtain efficient and compact Fault Tree models from failure datasets in a robust and scalable manner. For this matter, we propose, for the first time, using Multi-Objective Evolutionary Algorithms (MOEAs) to automatically infer FT models and cast the optimisation as a multi-objective task. This resulted in the FT-MOEA algorithm (Chapter 2), focusing on three optimisation metrics, including FT size and accuracy-related metrics. FT-MOEA consistently produced compact FT structures, but faced scalability issues. To address this, we developed the SymLearn toolchain (Chapter 3), which uses a ‘divide-and-conquer’ approach by identifying modules and symmetries in the failure dataset, breaking the inference problem into smaller tasks. Additionally, to improve robustness and scalability, the FT-MOEA-CM extension (Chapter 4) includes additional metrics from the confusion matrix. Our approaches in Part I of this dissertation contribute to automating FT model construction, revealing compact structures. These consistent structures can help uncover relationships between basic and intermediate events, providing valuable insights for asset managers to improve reliability modelling. Part II focuses on Markov Process-based Prognostics, specifically the stochastic deterioration modelling of sewer mains. Sewer systems are critical to social welfare but pose significant challenges due to their extensive scale, slow degradation, and limited capacity to monitor the entire network. Accurate modelling of the deterioration profile is crucial for optimising inspections and maintenance, thereby enhancing the reliability and availability of the networks. Various deterioration models are discussed in the literature, ranging from physics-based to data-driven approaches, each with distinct advantages and limitations. In Part II of this dissertation, we address how and to what extent it is possible to accurately model Multi-State Deterioration with applications in sewer mains. For this, we focus on Markov chains, widely used to model stochastic sequences through states and transitions. Since the 1990s, they have been applied to represent damage severity levels in sewer mains using inspection data from Closed Circuit Television cameras. Nonetheless, further evaluation of their assumptions and properties is required. We present a case study of a Dutch sewer network (Chapter 5), starting with Discrete-Time Markov Chains for deterioration modelling and examining two Markov chain structures. Given challenges such as interval-censored data, advanced analysis was necessary. In Chapter 6, we use the Turnbull estimator for non-parametric analysis to establish a ground truth. Although both homogeneous and inhomogeneous-time Markov chains are employed for sewer mains deterioration, no prior studies have compared their performance on the same dataset. Chapter 6 addresses this by demonstrating that inhomogeneous-time Markov chains are more versatile at capturing non-linear stochastic behaviour, while also highlighting issues like overfitting that reduce predictive accuracy. Part II provides a real-world case study, emphasising the need to critically evaluate modelling assumptions to enhance deterioration modelling of sewer mains using Markov chains. Finally, Part III focuses on Maintenance Optimisation of sewer mains, where obtaining optimal maintenance policies for such large-scale systems is a complex task. This complexity arises, among others, from the system’s scale and simplifications in the deterioration model. Among the various techniques available, Reinforcement Learning (RL) approaches remain largely unexplored for devising strategic main- tenance actions in sewer mains. Thus, in Part III of this dissertation, we focus on how to devise optimal maintenance strategies for components with Multi-State Deterioration such as sewer mains using Deep Reinforcement Learning.In Chapter 7, we frame the sequential decision-making problem using Deep Reinforcement Learning (DRL) for component-level maintenance of sewer mains. This framework considers damage severity levels, testing different deterioration model assumptions and evaluating their impact on maintenance policy. Our results show that agent-based policies outperformed heuristics by learning optimal sequences of maintenance actions. Part III provides evidence that DRL-based techniques o!er a flexible framework with the potential to improve heuristics and support maintenance decision-making for sewer mains. However, training these models to achieve the desired behaviour remains a challenging task. |
Bram Ton; Niek Tempert; Danny Plass: Immersive visualisation of point cloud data of railway environments. In: 2024 10th International Conference on Virtual Reality (ICVR), IEEE, 2025, ISBN: 979-8-3503-6423-1. (Type: Proceedings Article | Links | BibTeX)@inproceedings{Ton2025, |