AI-Based Self-Learning Ventilation System introduced in the Metro of Madrid
Inspired by foraging behaviour of bees, it reduces cost, CO2 emissions and ensures high-quality air
- Τετάρτη 27 Φεβρουάριος 2019 21:30:00 EET, 831 impressions
- Aseniya Dimitrova
Latest innovation from the public transportation system of Madrid: Metro de Madrid, together with Accenture Applied have implemented self-learning Artificial intelligence ventilation system which helps to minimize costs, CO2 emissions with the upshot of increased comfort of passengers, thanks to air with higher quality. So far, the new system is accountable for 25 percent decrease in ventilation costs and 1800 tons less carbon dioxide emissions annually.
The network of the Metro of Madrid is close to 300 kilometres long with 301 stations, serving roughly 2.3 million of passengers daily. This, naturally, requires well-functioning conditioning systems (the existing one currently using almost 900 fens, producing 80 gigawatt hours of energy every year), which are intrinsically linked to high costs, particularly elevated during the hot summer season.
For this project, experts from the Metro of Madrid and Accenture Applied Intelligence have joined forces and managed to deploy an optimized algorithm, based on a large data collection, combining air temperature, station architecture, train frequency, passenger load and electricity price throughout the day. Using machine learning, the new algorithm improves the performance of the air conditioning network over time. What is particularly interesting about the new solution is that it has been inspired by bees and their coordinated foraging behaviour.
Discover more about the project from Accenture.