1 1.7. Main contributions 23 Our findings suggest that using MOEAs for the inference of FT models generally has a positive impact in terms of robustness, scalability, and convergence speed. Contributions on Markov Process-based Prognostics: Multistate deterioration modelling In Part II, we used Markov chains to model Multi-State Deterioration (MSD) in sewer mains. Our contributions are three-fold: 1. We present a real-world case study from the Netherlands (Section II.4.3). Part of the data is publicly available at zenodo.org/record/6535853. 2. We evaluate two types of Markov chain structures (Chapter 5) typically used for MSDM in sewer mains, discussing their benefits and drawbacks. Additionally, we extend and propose a Markov chain structure (Chapter 6) that accounts for functional failure states. 3. We compare the assumptions of homogeneous and inhomogeneous time Markov chains (Chapter 6), identifying inhomogeneous-time Markov chains as more versatile for long-lived assets like sewer mains. Data and implementations are available at https://gitlab.utwente.nl/fmt/ degradation-models/ihctmc. Our findings suggest that using Markov chains for MSDM in sewer mains has the potential to describe and predict the severity level across populations of sewer mains. Contributions on Maintenance Optimisation: Maintenance optimisation of multi-state components In Part III, we used Multi-State Deterioration Model (MSDM) and Deep Reinforcement Learning to devise component-level strategic maintenance planning with applications in sewer mains. Our contributions are two-fold: 1. In Chapter 7, we propose a DRL framework for devising maintenance policies at the pipe level, considering MSDM. We detail model calibration and have made our models and dataset publicly available in the repository: zenodo.org/records/11258904. 2. We evaluate the influence of homogeneous and inhomogeneous MSDM on devising strategic maintenance, comparing agent behaviours against wellknown maintenance policy heuristics. Our findings suggest that DRL o!ers a flexible framework with untapped potential for maintenance strategies, and it is crucial to integrate degradation model assumptions, as they significantly influence policy behaviour.
RkJQdWJsaXNoZXIy MjY0ODMw