NVIDIA – a multinational technology company which is a dominant supplier of artificial intelligence hardware and software, has presented its new approach to carbon capture and storage, readily available for real-world applications using NVIDIA Modulus and NVIDIA Omniverse.
The approach is accelerating carbon capture and storage modeling using Fourier Neural Operators (FNO) architecture. NVIDIA is using FNO, NVIDIA Modulus and AI to accelerate simulations by 700,000 times which helps engineers quickly select the best geographical locations, determine the optimal depth and spacing of wells, and select the optimal injection rate and pressure for the CO2.
Via NVIDIA Omniverse, engineers can visualize and optimize the full inspection process. The accelerated speed of simulations by 700,000x enables many critical tasks for carbon capture and storage decision-making that were prohibitively expensive.
For example, it saves time and resources – a rigorous probabilistic assessment for maximum pressure buildup and CO2 plume footprint would take nearly 2 years with numerical simulators while using Nested FNO, it can take just 2.8 seconds.
The video released by NVIDIA shows the process of storing CO2 in an underground reservoir consisting of saline aquifer. The CO2 arrives at the surface facility, from which it is distributed to four wells and pumped to the underground reservoir. Inside the reservoir, pressure buildup and evolution of CO2 saturation are shown over a 30-year period.
NVIDIA technology has been used as a basis to achieve an AI-powered carbon capture and storage with interactivity at scale. With the NVIDIA platform, engineers can predict, visualize what will happen, and interact with the model to ensure the safety and reliability of the storage. According to NVIDIA, this helps ensure a safe storage process, reducing the amount of CO2 released into the atmosphere.