Posters

Abstract

A perfect surface of rotor blades is a precondition of harvesting wind in the most efficient way. Blade damage causes immediate energy losses of up to five per cent. Reduced aerodynamic performance diminishes power generation. One of the significant impacts on rotor blades is rain, characterized by erosion, material degradation, surface roughening, and loss of coating. Reduced efficiency, higher maintenance costs and safety concerns can be addressed in the most efficient way by optimizing routine inspection procedures and a comprehensive, state of the art artificial intelligence (AI) based analysis. Given the importance of "Energiewende", the transition of the current energy system from fossil to renewable energies, Germany's senior scientific and technical Federal institute, Bundesanstalt fuer Materialforschung und -pruefung (BAM), has launched an R+D project for the development of an innovative inspection and blade damage assessment methodology. Novel ground-based thermographic and visual photographs are done simultaneously to create a full wind turbine inspection. To avoid turbine downtime, and to get better results, the inspection is done while the turbine is spinning and generating energy. This efficient, combined, comprehensive approach reveals the location and depths of blade damages. In addition, thermographic turbulence patterns can also hint at damages below the blade surface, which are not (yet) visible in conventional images. State of the art AI models and methodologies are used to detect, localize, and match visual damages and thermographic turbulence patterns. The quality of an automated detection tool is outstanding compared to conventional manually assessed visual inspections. The ranking and categorization of rain erosion damages as well as aerodynamic effects on thermographic inspection images of rotor blades with AI allows the roll-out of innovative, advanced, and cost-efficient inspection methodologies. This increases wind turbine efficiency and at the same time reduces costs of inspections and rotor blade maintenance. Wind park operators and inspection companies are in full control of their data when using this newly developed method. Inspection visuals can be uploaded to a web-platform to access the advance AI algorithms, which conduct a speedy, all-inclusive analysis of blade rain erosion damages and their impact on overall energy efficiency and performance. The underlying AI model, methodology and system are no black box. The R+D project was structured in three phases such as the creation of high-quality data for training & testing, development of training models, and review and reiteration of results. The AI based solution developed by LATODA in conjunction with BAM and Romotioncam will reduce inspection time significantly, avoid turbine downtime and generate analytical insights for turbine operators, blade inspection and repair companies, with the ultimate goal to increase turbine efficiency and thus, production of renewable energy. The first inspections with visual and thermographic images are already done and presentable.