01.12.2025

InInspekt delivers a step change in wind turbine blade inspection. It combines autonomous crawler robotics, multimodal sensing, and AI-driven data fusion to enable precise, objective detection of internal and subsurface damage directly inside rotor blades. This approach reduces inspection risk, downtime, and uncertainty while providing standardized, high-quality diagnostic data that supports condition-based maintenance and extends blade service life.

Objectives

The InInspekt project aims to fundamentally advance the internal inspection of wind turbine rotor blades by developing an autonomous, robotic inspection system capable of operating inside blade structures. The core objective is to enable objective, repeatable, and high-precision detection of internal defects—such as delaminations, cracks, bonding failures, and lightning damage—using a combination of robotics, multimodal sensing, and artificial intelligence. The system is designed to replace subjective, manual inspections with a data-driven, standardized approach suitable for both onshore and offshore wind turbines.

Approach

• InInspekt follows a robot-centric, multisensor, AI-assisted inspection concept tailored to the confined and complex geometry of rotor blade interiors:

• A compact mobile crawler robot navigates autonomously inside the blade using LiDAR, IMU, and advanced 3D localization methods.

• A multimodal sensor payload integrates:

  • RGB cameras for visual inspection,

  • passive infrared thermography for rapid full-coverage screening,

  • active thermography with controlled heating to reveal subsurface defects,

  • and high-resolution 3D laser scanning for precise geometric context.

• Sensor data are spatially calibrated and fused in real time, enabling thermographic and visual data to be mapped directly onto 3D blade geometries.

• The inspection process is two-stage:

  • Primary inspection using passive thermography and visual sensing to identify suspicious areas.

  • Targeted high-precision inspection using active thermography and refined sensor positioning.

• AI models (including deep-learning-based object detection and 3D point-cloud analysis) analyze data both in real time and offline, supporting anomaly detection, damage classification, and visualization.

• The project also develops field-deployable calibration methods and produces a standardized, open benchmark dataset to support validation and future research.

Benefits for Industry

• Detection of internal blade damages, which cannot be detected by conventional inspection methods.

• Improved damage detection quality, including early-stage and subsurface defects not accessible through visual inspection alone.

• Reduced turbine downtime, as inspections are faster and more targeted compared to conventional methods.

• Higher inspection safety, by minimizing or eliminating human entry into confined blade interiors.

• Standardized and reproducible inspection results, enabling better comparability across turbines, sites, and operators.

• Lifecycle cost reduction, through earlier intervention, extended blade lifetime, and avoidance of catastrophic failures.

• A technology platform that is scalable and transferable to other inspection tasks in energy and infrastructure systems.

​

Partners’ Roles

• LATODA: Responsible for AI and data analytics, including real-time anomaly detection, damage classification, sensor data fusion, and development of machine-learning pipelines based on extensive wind-energy domain expertise.

• EduArt Robotik GmbH (Coordinator): Leads the development of the mobile robotic platform, including mechanical design, locomotion, robustness, power supply, and system integration using ROS-based architectures.

• Bundesanstalt für Materialforschung und -prüfung (BAM): Contributes expertise in non-destructive testing and thermographic methods, including passive and active thermography, validation strategies, and transfer of results into standards and regulatory frameworks.

• Justus-Maximilians-Universität Würzburg (JMU): Provides advanced research in robotics, SLAM, 3D laser scanning, sensor calibration, and data fusion, ensuring precise spatial alignment and robust autonomous navigation inside complex blade geometries.

This project is funded by BMFTR