Sensor System and Signal Processing for Automated Blade Collision Detection on Wind Turbines

摘要

This paper presents the design and implementation of an automated blade collision detection system for use on wind turbines, toward the goal of supporting quantitative assessment of wind energy impacts on wildlife. A wireless, multi-sensor module mounted at the blade root measures surface vibrations, and a blade-mounted camera provides image capture of colliding objects. Using sensor data recorded on an operational wind turbine, an automated detection algorithm was developed using a machine learning approach. The sensor system design is presented, along with algorithm design, implementation, training, and testing results. This demonstrates a new approach for automated, on- blade collision detection for wind turbines, with broad utility across structural health monitoring (SHM) applications.