Optimization of wind turbine blade maintenance for a Danish utility company using defect detection and life-cycle assessment

Thesis Abstract

Wind turbine blade integrity and lifecycle performance are critical to the reliability and cost-effectiveness of offshore and onshore wind energy assets operated by Danish utilities. The study addresses the persistent challenge of escalating maintenance costs and unplanned downtime caused by blade surface damage, delamination, and moisture ingress, which compromise aerodynamic efficiency and structural safety. The aim is to develop an integrated maintenance optimization framework that combines defect detection, condition assessment, and life-cycle assessment (LCA) to minimize total ownership cost while sustaining power output. Specific objectives include (i) identifying and validating non-destructive evaluation (NDE) techniques for early defect detection in rotor blades, (ii) quantifying defect propagation under operational loads using Paris law-inspired fracture mechanics and Bayesian updating, (iii) developing a predictive maintenance policy via stochastic optimization that integrates defect states with blade repair and replacement decisions, (iv) performing a cradle-to-grave LCA for blade maintenance scenarios to compare environmental and economic impacts, and (v) crafting decision-support guidelines for asset managers at the Danish utility. The methodology adopts a mixed-methods, case-study design anchored in theory-driven decision modelling. The population comprises 150 utility-grade wind turbines operated by the Danish company across offshore and onshore sites. A stratified sample of 60 turbines, representative of turbine models, hub heights, and geographic exposure, will be monitored over 36 months. Data collection integrates (i) defect data from quarterly NDE inspections using ultrasound, thermography, and digital image correlation, (ii) operational data including wind speed, rotor load, pitch angle, and temperature from SCADA databases, (iii) maintenance logs detailing repair type, cost, and downtime, and (iv) environmental impact data required for LCA from supplier and lifecycle inventories. Instrument validity will be ensured through calibration against baseline blade samples and corroboration with manufacturer inspection reports; reliability will be tested via test-retest on a subset of 10 turbines. The analytical framework comprises (i) defect kinetics modelling using a hierarchical Bayesian approach to update failure probability as a function of detected defects and loading history, (ii) regression analysis to identify key predictors of defect growth, (iii) a stochastic dynamic programming model for maintenance scheduling, and (iv) LCA using the ISO 14040/44 framework to compare scenarios in terms of global warming potential, cumulative energy demand, and cost. The model specification will integrate a defect-affected reliability function with a maintenance decision rule that minimizes expected total cost of ownership under budget constraints. Sensitivity analyses will examine variations in discount rate, inspection frequency, and repair lead times. Ethical considerations include data privacy, supplier confidentiality, and adherence to Danish safety regulations. Expected findings indicate that targeted, data-driven inspection intervals based on defect growth forecasts can reduce unplanned outages by 15–25% while lowering blade-related maintenance costs by 10–20% over the study horizon. It is anticipated that the combination of ultrasound-based defect metrics with thermographic indicators will yield a robust early-warning system with an estimated false-positive rate below 8%. The LCA is expected to reveal that optimized maintenance, though potentially increasing short-term inspection costs, will lower long-term environmental burdens and energy payback time due to higher turbine availability and reduced replacement frequency. The study will contribute to knowledge by advancing an integrative maintenance paradigm that links defect physics, probabilistic forecasting, and lifecycle sustainability assessment within a real-world Danish utility context, thereby offering a replicable framework for other Northern European operators. The main conclusion is that a defect-informed, lifecycle-aware maintenance strategy significantly enhances asset reliability and cost efficiency without compromising environmental performance. Recommendations include adopting a tiered inspection regime guided by defect growth projections, integrating predictive maintenance into enterprise asset management platforms, standardizing data collection for cross-site comparability, and extending the framework to consider blade tip repair options and end-of-life recycling pathways to further reduce total cost of ownership and environmental impact.