AI for Energy & Climate

Through enhanced monitoring and optimization, AI can aid in predictive, proactive, and reactive action against climate change. AI is used to monitor and mitigate the current impacts of climate change more efficiently, and experts predict that AI could precipitate scientific breakthroughs to help combat climate change and global warming. AI can help optimize solar and wind farms, simulate climate and weather, and advance carbon capture and power fusion breakthroughs. AI is also helping to decrease auto emissions by increasing efficiency and optimizing the performance of shared, electric, and autonomous transportation.

Reducing Auto Emissions

Google Project Greenlight

Google’s Project Greenlight uses AI to optimize traffic lights and reduce vehicle emissions in cities.  By analyzing traffic patterns and developing recommendations for traffic light timing, Green Light is helping to reduce stop-and-go traffic, lower greenhouse gas emissions, and improve urban mobility.  The system, which can analyze thousands of intersections simultaneously, collects data from Google Maps driving trends and provides actionable recommendations for city traffic engineers to implement without the need for additional hardware.  Green Light is currently deployed in 13 U.S. cities.

 

Optimizing Offshore Wind Output

Siemens Gamesa and NVIDIA

Renewable energy company Siemens Gamesa is working with NVIDIA to apply AI surrogate models to optimize its offshore wind farms to output maximum power at minimal cost. Together, the companies are using physics-informed super-resolution AI models to generate high-resolution simulation data, orders of magnitude faster, enabling more accurate engineering wake models. Optimizing the configuration of each wind turbine during the construction of new wind farms is crucial to maximizing output. This process typically takes up to 40 days using traditional methods. Using AI models, the process now takes just 15 minutes.

 

Reducing Aviation Emissions

American Airlines, Google Research, and Breakthrough Energy

American Airlines, Google Research, and Breakthrough Energy teamed up on a project to reduce the global warming effects of contrails using AI. Clouds created by contrails account for roughly 35% of aviation’s global warming impact, over half the impact of the world’s jet fuel. Together, the companies combined huge amounts of data — like satellite imagery, weather and flight path data — and used AI to develop contrail forecast maps to test if pilots can choose routes that avoid creating contrails.

A group of pilots at American flew 70 test flights over six months while using Google’s AI-based predictions, cross-referenced with Breakthrough Energy’s open-source contrail models, to avoid altitudes that are likely to create contrails. After these test flights, they analyzed satellite imagery and found that the pilots were able to reduce contrails by 54%. This was the first proof point that commercial flights can verifiably avoid contrails and thereby reduce their climate impact.

 

Decarbonizing Buildings

Mortar IO

Buildings are responsible for 40% of carbon emissions globally, and 80% of the buildings that will be standing in 2050 are already built. Catalyzing retrofits for building decarbonization using AI could greatly impact global emission reductions.

Mortar IO, a virtual modeling and simulation platform, makes decarbonization retrofit planning more accurate and cheaper to implement. Mortar IO uses AI to digitize and quickly plan carbon reduction for thousands of buildings. These automated digital audits help organizations understand how to achieve net zero for entire real estate portfolios in minutes rather than months, potentially unlocking an accelerated path to net zero for every building.

 

Eliminating Pipeline Leaks

Orbital Sidekick

Orbital Sidekick (OSK) uses hyperspectral imagery, satellite and aerial asset monitoring, and AI to detect hundreds of suspected gas and hydrocarbon leaks across thousands of miles of pipeline across the globe. Using AI technology, OSK continues to drive improvements in community safety, environmental performance, and overall operating efficiency for critical pipeline infrastructure.

 

Democratizing and Accelerating Climate Research

NASA, IBM, and Hugging Face

Access to the latest data remains a significant challenge in climate science where environmental conditions change almost daily. Despite growing amounts of data — estimates from NASA suggest that by 2024, scientists will have 250,000 terabytes of data from new missions — scientists and researchers still face obstacles in analyzing these large datasets.

In collaboration with NASA, IBM built an AI foundation model for geospatial data to accelerate climate-related discoveries by removing obstacles that scientists and researchers face in analyzing large datasets. The new model can analyze geospatial data up to four times faster than state-of-the-art deep-learning models, with half as much labeled data. By making this large geospatial foundation model available via Hugging Face — a recognized leader in open-source and a well-known repository for all transformer models — IBM and NASA have democratized access to leverage the power of collaboration to help advance new innovations in climate and Earth Science that will improve our planet.

 

Building Better Batteries

Carnegie Mellon’s Clio

Scientists at Carnegie Mellon recently used robots and AI to design faster-charging lithium-ion batteries. The team of researchers was looking for electrolytes that would allow for batteries to charge faster — which is one of the biggest problems in battery technology today and a major barrier to widespread electric vehicle adoption. The team used an automated arrangement of pumps, valves, vessels, and other lab equipment that they dubbed “Clio” to mix together different ratios of three potential solvents and one salt. As their paper points out, “battery innovations can take years to deliver” in part because there are so many potential chemicals that can be used in various ratios that optimizing them is “time-consuming and laborious” — at least for people. By feeding Clio’s results into a machine-learning system called “Dragonfly,” Clio was able to run experiments significantly faster, yielding six solutions that out-performed existing commercially available electrolyte solutions.