A group of scientists has developed a new AI-based technology that can help lock up greenhouse gases like carbon dioxide in porous rock formations faster and more precisely than ever before.
Carbon capture and sequestration are climate change mitigation method that capture carbon dioxide emitted from power plants and then store it underground.
This technology is expected to help industries such as refining, cement, and steel to reach decarbonization and reduce emissions.
More than a hundred carbon capture and storage facilities are now being built throughout the world.
When injecting carbon dioxide into rock formations, scientists must avoid excessive pressure build-up that could damage geological formations and leak carbon into aquifers above the site or even into the atmosphere.
U-FNO, a new neural operator architecture, simulates pressure levels during carbon storage in a fraction of a second while tripling accuracy on specific tasks, assisting scientists in determining the best injection rates and locations.
Carbon storage simulations are used by scientists to determine the best injection sites and rates, limit pressure building, enhance storage efficiency, and guarantee that the injection activity does not fracture the rock formation. Understanding the carbon dioxide plume — the spread of CO2 through the earth — is also critical for a successful storage scheme.
In comparison, traditional carbon sequestration models are time-consuming and computationally costly. With machine learning models, the new method can achieve similar levels of accuracy while requiring far less time and money.
U-FNO is a gas saturation and pressure buildup prediction system based on the U-Net neural network and the Fourier neural operator architecture, also known as FNO. U-FNO is twice as accurate as a state-of-the-art convolutional neural network for the job while consuming only a third of the training data.
Scientists can use U-FNO to simulate how pressure levels would rise and where carbon dioxide will spread over the course of 30 years of injection. In addition, researchers can simulate a large number of injection sites quickly using GPU-accelerated machine learning.
Farah Hariri, a collaborator on U-FNO, commented that the research has showed how AI can help accelerate the process of climate change mitigation.