Heterogeneous Sensor Fusion for Autonomous Detection of Wildlife Collisions with Wind Turbines

Mortality of endangered or protected avian and bat species resulting from interactions with onshore and offshore wind turbines is a major conservation concern. Wildlife interaction monitoring systems can support turbines... [ view full abstract ]

Mortality of endangered or protected avian and bat species resulting from interactions with onshore and offshore wind turbines is a major conservation concern. Wildlife interaction monitoring systems can support turbines installation decisions and effective siting verification. Wind farm operation procedures for wildlife damage control and active or passive wildlife deterrent measures must be applied. In any of the above options, an autonomous monitoring system for strike detection and taxonomic verification must be applied for verification, validation and eventual siting permission process. Such autonomous and efficient system currently does not exist.
A novel multi-sensor system designed on the wake of an existing proof-of-concept sensor array developed and field-tested under a recent US Department of Energy grant for removing market barriers to offshore wind development is presented. The existing system, consisting of an integrated sensor package developed around five fundamental sensor types: 1) accelerometers 2) contact microphones, 3) visual cameras, 4) infrared cameras and 5) bioacoustics microphones was tested at wind turbine sites at the North American Wind Research and Training Center (NAWRTC) at Mesalands Community College in Tucumcari, NM and the NREL-National Wind Technology Center (NWTC) in Boulder, CO. Field tests results on vibrations node and bioacoustics will be presented as well as lessons learned for system design and components with inferences on the sensors types and number required for best performance versus cost ratio.
The future version of the platform, currently under development, prioritizes board-level integration to significantly decrease the size, weight, and power consumption of the sensor unit. The new research platform, designed for extremely small size, will integrate a 3-axis MEMS accelerometer, 3-axis gyro, low-power CMOS imager, and contact microphone with on-board computation to enable local processing of sensor signals and detection of strikes in real time. Heterogeneous sensor fusion will allow removal of blade rotation motion artifacts and generator-induced vibrations to lower the overall noise floor. Importantly, localized computation also enables wireless transmission of only detected events in place of continuously streaming raw data, which dramatically reduces power consumption. The platform will be powered through a battery with possible integration from rotational or vibrational energy harvesting and small solar panels. This core sensor platform can be adapted for on-blade use or modified for permanent, embedded installation during blade fabrication. The unit will integrate both Bluetooth Low Energy (BLE) and WiFi modules for investigation of appropriate node-node and node-nacelle communication links, and it may include a 3G uplink for cloud-based data logging. The system can also be integrated with appropriate deterrent systems.