EAPSI: Using an Unmanned Aerial Vehicle to Detect Wind Turbine Blade Defects in Offshore Wind Parks

Zhang Huiyi, Ames IA

Investigators

Abstract

Current wind turbine blades (WTB) maintenance requires a crane, a highly trained expert who can diagnose issues, and a large amount of safety equipment. Performing these inspections is enormously costly and time intensive, and in some cases nearly impossible to perform in a safe and consistent manner. This project aims to reduce inspection costs and improve worker safety by developing a prototype of an unmanned aerial vehicle that will be able to perform inspections that can identify hairline thickness cracks on turbine blades in even the most challenging environments. This research will be conducted at Shanghai Jiao Tong University in China in collaboration with Dr. Long Chengnian. China is an especially good site for studying semi-automatic blade inspections because China has a larger fraction of wind energy capacity than the United States and the air pollution in China complicates blade inspections because dust and particles can reduce the visibility of site employees and inspection devices. This project will develop a cost-effective semi-automatic device that can identify and classify WTB surface defects using advanced image processing techniques and especially will focus on offshore WTBs because offshore sites represent a worst-case scenario of obstacles to performing safe and consistent blade inspections. Offshore WTBs are especially large in size, are far from land on which a crane may be anchored, and are prone to defects and poor inspection lighting caused by dust, salt, and water vapor in the air. If the wind energy is to further develop its offshore capabilities, it is critical to develop a semi-automatic blade inspection system that can detect early defects and minimize the costs that remain an obstacle to wind energy's adoption as a replacement for burning fossil fuels. This NSF EAPSI award is funded in collaboration with the Chinese Ministry of Science and Technology.