Wind Turbine Blade Maintenance Automation: 85% repair recommendation accuracy and transparent costs

Maintenance for wind-power facilities has traditionally relied on the personal experience and judgment of maintenance personnel. As a result, the timing, method, and scope of repairs can vary when defects occur, while different cost-estimation standards across providers make it difficult to ensure transparency. Wind-turbine blades in particular require consideration of many variables, including weather conditions and defect history, yet the relevant data has been fragmented, limiting precise maintenance planning. These inefficient decisions lead to wasted maintenance effort and higher operating costs.

To address these field challenges, a low-code AIoT platform and industrial data-structuring technology were applied to the wind-power operations and maintenance domain. Fragmented defect information, repair history, and weather data were combined and structured for AI learning. An LLM-based repair-plan generation module then recommends the optimal repair process and automates cost-estimation logic for labor and materials. A compact AIDC infrastructure in Jeju further enhanced the model, creating a fast and transparent decision environment that automatically produces complex maintenance reports within five to ten minutes.

The solution achieved repair-process recommendation accuracy above 85% while keeping maintenance cost-estimation error within 20%. It also reached data consistency above 95%, improving confidence in maintenance records and sharply reducing repetitive administrative work. Ultimately, it transformed experience-led reactive maintenance into a data-centered intelligent predictive-maintenance system, maximizing cost efficiency for operators.