180 Chapter 8. Discussion 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. Data availability for deterioration modelling of sewer mains: the source of (many) challenges The goal of deterioration modelling is to create a mathematical abstraction that adequately links context with the system’s/component’s degradation behaviour. Data-driven approaches’ success relies heavily on data quality, which is particularly challenging in sewer asset management. High-quality inspection datasets collected from sewer mains (if any) are rare let alone public, hindering methods comparison and analyses (see Table II.1). To address this gap, we have made our case study available, sharing relevant damage codes and cohorts (Section II.4.3). A high-level comparison with other case studies shows that our dataset, featuring sewer covariates and historical inspections is fairly typical. Though this is a step forward, it is far from closing the gap. We now must question whether this information alone su"ces to build robust deterioration models for long-term sewer condition assessment. Likely the answer is no, for several reasons. For starters, a sewer network, in its simplest form—ignoring dependencies—is a population of pipes with covariates such as material, content, and age, and conditions like damage severity—data gathered from inspections. Data-driven models for long-term condition assessment require ‘rich’ and ‘su"cient’ data, which is di"cult to obtain from sewer inspection datasets. This is mainly because inspections in sewer mains are performed to assess pipe conditions to support maintenance or replacement actions and planning, not to feed degradation models. Sewer mains often represent heterogeneous populations. As shown in Figures II.2 and II.3, the datasets are highly unbalanced, missing key contextual data like soil properties and road usage. This leads to poor statistics for some covariates,
RkJQdWJsaXNoZXIy MjY0ODMw