General Electric (GE) will use IBM’s Summit, one of the globe’s most powerful supercomputers, to operate two new research projects aimed at increasing the supply of cleaner energy. Last month, the United States Department of Energy (DoE), which runs Summit at Oak Ridge National Laboratory, awarded 20 research teams a sum of more than seven million node hours on a supercomputer, including two from GE Research.

Behind Japan’s Fugaku supercomputer, the Summit supercomputing machine is the world’s second most powerful. Summit, which IBM built, has the processing capacity of 70 million iPhone 11s, allowing scientists to do huge computations such as modelling system behaviour or solving challenging physics problems. The two projects chosen by the Department of Energy to operate on Summit have now been revealed by GE. They are going to both address challenges in renewable energy generation.

One team, led by GE expert Jing Li, got 240,000 node hours to boost offshore wind power research. Li plans to use the Summit supercomputer to undertake complicated simulations to investigate novel ways of regulating and running offshore turbines to maximize wind production.

Li’s team will focus on a wind phenomenon termed the low-level coastal jets, which can influence the reliability and performance of offshore wind turbines and occur along many coastlines. The researchers will use high-fidelity computational models to study interactions between wind farms and low-level coastal aircraft to advise future, more effective wind farm designs. The results will also be utilized to assist the Department of Energy’s ExaWind project, which aims to speed up the development of the onshore and offshore wind farms in the United States.

To do so, you’ll need a thorough understanding of how natural wind events interact with constructed infrastructure. Due to the numerous components at play, simulating these interactions has a high computational cost. Most research efforts can only anticipate the behaviour of the small number of turbines at this time.

ExaWind aims to create predictive models of wind farms featuring tens of wind turbines that are megawatt-scale, distributed over many square kilometres with complicated terrain. This computation might entail up to 100 billion grid points. As a result, Summit’s massive compute power offered to Li’s team is a promising start toward completing the ExaWind challenge.

General Electric (GE) will use IBM’s Summit, one of the globe’s most powerful supercomputers, to operate two new research projects aimed at increasing the supply of cleaner energy. Last month, the United States Department of Energy (DoE), which runs Summit at Oak Ridge National Laboratory, awarded 20 research teams a sum of more than seven million node hours on a supercomputer, including two from GE Research.

Behind Japan’s Fugaku supercomputer, the Summit supercomputing machine is the world’s second most powerful. Summit, which IBM built, has the processing capacity of 70 million iPhone 11s, allowing scientists to do huge computations such as modelling system behaviour or solving challenging physics problems. The two projects chosen by the Department of Energy to operate on Summit have now been revealed by GE. They are going to both address challenges in renewable energy generation.

One team, led by GE expert Jing Li, got 240,000 node hours to boost offshore wind power research. Li plans to use the Summit supercomputer to undertake complicated simulations to investigate novel ways of regulating and running offshore turbines to maximize wind production.

Li’s team will focus on a wind phenomenon termed the low-level coastal jets, which can influence the reliability and performance of offshore wind turbines and occur along many coastlines. The researchers will use high-fidelity computational models to study interactions between wind farms and low-level coastal aircraft to advise future, more effective wind farm designs. The results will also be utilized to assist the Department of Energy’s ExaWind project, which aims to speed up the development of the onshore and offshore wind farms in the United States.

To do so, you’ll need a thorough understanding of how natural wind events interact with constructed infrastructure. Due to the numerous components at play, simulating these interactions has a high computational cost. Most research efforts can only anticipate the behaviour of the small number of turbines at this time.

ExaWind aims to create predictive models of wind farms featuring tens of wind turbines that are megawatt-scale, distributed over many square kilometres with complicated terrain. This computation might entail up to 100 billion grid points. As a result, Summit’s massive compute power offered to Li’s team is a promising start toward completing the ExaWind challenge.

Michal Osusky, a GE researcher, was also given additional 256,000 node hours on the Summit for a second research project focusing on using machine-learning approaches to enhance the design of physical devices such as aircraft engines and power production turbines. Osusky’s team swiftly simulated real-world engines and executed virtual testing to validate designs faster than they could with traditional methods by combining machine learning with simulation.

Leave a Reply

Your email address will not be published. Required fields are marked *