Environmental innovation to protect biodiversity
Interview with Carole Le Goc, department head – ALTEN in France
An alarming finding.
Scientific studies, notably those of the WWF, reveal a dramatic decline in the vertebrate population, with a loss of 70% in just 50 years. Birds, particularly rare and protected species, are seriously threatened by wind turbines, whose blades can reach speeds of 300 km/h, making them difficult for birds to detect. In France, around 10,000 wind turbines are spread across 2,300 wind farms, causing the death of between 84,000 and 126,000 birds every year.
Dbird, an innovative solution.
The Dbird project, developed by a team of experts from ALTEN’s Grenoble laboratory, is at the forefront of technological solutions to protect biodiversity. This ambitious and innovative project tackles a major environmental problem: bird mortality due to collisions with wind turbine blades.
Can you tell us about the Dbird project and its main objectives?
The Dbird project aims to protect biodiversity around wind farms by reducing bird mortality caused by collisions with turbine blades. We have developed an agile low environmental impact solution that uses advanced technologies to detect and react to the presence of birds, the environment in the vicinity of wind turbines, etc.
What are the main innovations you have implemented in this project?
We have incorporated environmental innovations. We use Edge architecture within the wind turbine itself to process data close to the sensors, reducing data transmission and carbon impact. In addition, our Artificial Intelligence models are designed and optimised for embedded devices with limited resources, thereby reducing energy consumption.
Can you tell us more about the Edge architecture and its importance to the project?
The Edge architecture is important for our project because it brings data processing closer to the sensors. This reduces communication to the Cloud, improves data security and, in the case of wildlife detection, reduces latency and improves results.
How have you optimised the Artificial Intelligence models to make them more sustainable?
We optimised lightweight AI models with techniques such as weight quantisation and pruning, enabling them to run on resource-limited microcontrollers to reduce power consumption. For example, we use STM32 microcontrollers to detect farm machinery and Jetson Nano cards to detect birds.
What countermeasures have been put in place to protect birds?
We have developed a system that assesses the risk of bird collisions according to various parameters. Depending on the risk, we can slow down or stop the wind turbines, emit sounds or turn on lights to scare the birds away, and make detection more accurate thanks to AI.
How do you assess the carbon impact of your solution?
We assess the carbon impact of our solution at every stage using a life cycle analysis, including during the use phase, with the support of another Grenoble Lab project. This helps us to identify the elements that have the greatest impact and to propose improvements. For example, the industrial PC used for data aggregation has a significant impact, and we are considering replacing it with a more environmentally-friendly option. Our integrated AI solution has also been found to meet its environmental impact commitments.