Posters

Abstract

As an OEM for wind turbines your ambition is to produce maximum possible quantities at low-cost base in a perfect quality, so customers get their turbine up and running due to their time and financial planning. Especially quality and time are metrics which are not easy to handle for large composite structures such as wind turbine rotor blades and are sometimes risky for OEMs to handle. The shear dimensions and the relatively low-automization rate / high manual work make rotor blade production a perfect area in which a real-time feedback during the material lay-up, on of the crucial process steps, will lower risk for flaws or costly re-work. An AI-based system can give the workers direct feedback via a lasers and correct position before the blade is casted. Technically this is done with the help of several cameras distributed overhead of the moulds, streaming visual information and distribute them to an edge computing system which hosts the models and decision-making part and send feedback signal to the laser system which projects feedback directly onto the mould. Major features are: * State-of-the-art machine learning models in the field visual deep learning 1. Edge detection 2. Human detection * Modular software solution at an edge architecture * Near real-time feedback * Intuitive user interface (UI) Our features serve the needs of the workers at the mould and they appreciate the ease of using the UI and even more: it helps new colleagues to speed up faster. On the other hand, it protects the privacy of each worker by anonymizing the visual data, while also complying with other high GDPR-standards in SGRE. We are constantly investigating in other manufacturing and/or business areas for continues improvement. We call it this product “Optical Blade” and are proudly in production since 2022.