University of the Aegean
Dependability modelling and environmental-based maintenance planning for renewable energy systems
Abstract
dc:descriptionOffshore renewable energy systems are increasingly deployed to support decarbonization goals, yet their operation in harsh marine environments creates persistent challenges for reliability, maintainability and cost-effective asset management. In offshore settings, environmental stressors (e.g., wind-speed variability, wave impacts, corrosion and access constraints) increase failure rates and can delay interventions because maintenance is feasible only within safe “weather windows”, thereby amplifying downtime and energy losses. These conditions expose limitations of conventional dependability approaches that assume stationary conditions or apply fixed-interval maintenance policies developed for onshore assets and they are further compounded by limited publicly available offshore failure data that forces practitioners to rely on imperfect proxies. This thesis develops an integrated stochastic modelling and maintenance-planning framework that explicitly couples environmental variability, component degradation and maintenance decision-making for renewable energy systems, with emphasis on offshore and hybrid platforms. The methodological core combines Markov modelling, using Homogeneous Continuous-Time Markov Chains (HCTMC) as a foundation and advancing to Cyclic Non-Homogeneous Markov Chains (Cyclic NHMC) to represent periodic and seasonally varying transition dynamics, with multi-state system performance evaluation via the Universal Generating Function (UGF) technique. The Cyclic NHMC formulation is leveraged to encode predictable seasonal patterns (e.g., winter/spring/summer/autumn regimes) through period-specific transition structures, enabling a realistic representation of non-stationary offshore conditions and their impact on failure and maintenance/repair activities and energy conversion performance.Building on these models, the thesis proposes and evaluates adaptive maintenance concepts that integrate environmental and operational constraints into scheduling decisions. These include Maintenance Wind Speed Window-Based Opportunistic Maintenance (MWSW-Based OM) for offshore wind turbine rotor blades, targeting low-wind periods that naturally reduce production opportunity cost and Seasonal Energy Demand-Based Maintenance (SEDBM), which adjusts preventive maintenance (PM) intensity in anticipation of seasonal demand and stressor severity to improve resilience during critical periods. The framework is demonstrated across multiple engineering systems, spanning wind energy assets (from wind farm dependability with minimal repair assumptions to offshore rotor blade maintenance modelling), hybrid wind-wave power generation configurations, wind-powered reverse osmosis (RO) desalination systems and broader offshore hybrid renewable energy systems (HRES) integrating wind turbines, wave energy converters and solar photovoltaics. Real-world meteorological inputs and seasonally varying parameters are incorporated to ground the analysis in realistic operating conditions, and the resulting numerical investigations illustrate how environment-aware modelling and maintenance alignment can enhance dependability measures (e.g., availability and energy-related measures) while supporting more economically efficient maintenance planning for offshore renewable infrastructures. A novel MWSW-Based OM strategy is formalized within an HCTMC framework that explicitly integrates wind speed class dynamics into maintenance state transitions for offshore rotor blades. Rather than scheduling interventions at fixed intervals, the model continuously monitors the evolution of wind conditions and triggers maintenance actions during naturally occurring low-wind windows, when the turbine is environmentally constrained from producing power. This direct coupling of environmental state information with operational and maintenance decisions transforms classical opportunistic maintenance into a proactive, wind-aware scheduling mechanism. A multi-state model for a hybrid wind-wave power generation system is developed using HCTMC, jointly embedding wind turbine and wave energy converter operational dynamics within a unified stochastic framework. The model captures how the natural complementarity between wind and wave resources stabilizes system output under both regular and intense offshore weather scenarios, reflecting the harsh intermittency regimes characteristic of offshore environments. Extending beyond homogeneous modelling, a Cyclic NHMC is applied to a multi-technology offshore hybrid platform. By structuring season-specific transitions that repeat annually, the framework embeds environmental periodicity directly into the system’s operational and degradation dynamics. Complementing this modelling advancement, a SEDBM strategy is introduced, which aligns maintenance intensity with the interplay between anticipated seasonal energy demand and prevailing environmental stressor severity. By intensifying maintenance before high-stress, high-demand periods, SEDBM dynamically adapts the system’s health management posture to the environmental cycle itself, offering a practically deployable alternative.
Degree
thesis:*- Grantor dc:publisher
- University of the Aegean
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Psomas, Panagiotis
- Ψωμάς, Παναγιώτης
Subjects
dc:subject × 18- Μοντελοποίηση αξιοπιστίας
- Συστήματα ανανεώσιμων πηγών ενέργειας
- Σχεδιασμός συντήρησης
- Dependability modelling
- Renewable energy systems
- Maintenance planning
- Άλλες Επιστήμες Μηχανικού και Τεχνολογίες
- Επιστήμες Μηχανικού και Τεχνολογία
- Μηχανική, διεπιστημονική προσέγγιση
- Μηχανική, γενικά
- Περιβαλλοντική Μηχανική
- Ανανεώσιμη ενέργεια, Αειφορία και Περιβάλλον
- Other Engineering and Technologies
- Engineering and Technology
- Engineering, interdisciplinary
- Engineering, general
- Environmental Engineering
- Renewable Energy, Sustainability and the Environment
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/62457