New PNNL Computing Project to Tackle Grid Modernization at Warp Speed
Researchers work to optimize grid planning through "exascale" supercomputer applications
One billion calculations per second. Is that even possible? In the world of high-performance computing, the answer is yes.
But one quintillion—what even IS that? It's one followed by eighteen zeroes. And why would you ever need that kind of computing speed?
Because to solve today's complex problems—such as clean energy, cyber security, and climate change—it becomes increasingly important to connect and evaluate numerous variables quickly and accurately.
Under DOE's Exascale Computing Project, PNNL was recently awarded $3.8 million over three years for a 'seed' project related to computational modeling of the power grid. Led by PNNL power engineer Henry Huang, the project aims to develop advanced modeling and simulation solutions for the massive calculations grid planners will need to deal with uncertain renewable power sources, electric vehicles and smart loads that will introduce random changes to the grid.
"The random changes together with the inherent dynamics in the power grid form an extremely complex problem,” said Huang. “Further coupled by a time horizon of decades, the number of equations we need to solve for long-term power system planning rapidly increases. Only exascale computers would have a chance to solve this problem. Once solved, there could be hundreds of millions of dollars in savings or deferred investment."
Juggling a multitude of variables can only be solved with high-performance exascale computers capable of handling at least a quintillion calculations per second.
The PNNL-led project, Optimizing Stochastic Grid Dynamics at Exascale, is one of 22 selected for funding in a suite of nearly $40 million worth of new projects to develop applications for future exascale computer systems. ‘Seed’ projects like Huang’s target new areas of computational science, and are intended to develop into full-blown applications for exascale systems that will be 50-100 times more powerful than today's fastest supercomputers.
For more information on the Exascale Computing Project awards, please see the DOE news release.