Optimizing Stochastic Grid Dynamics at Exascale
The unprecedented power grid evolution poses a monumental challenge in optimizing grid planning. In the next 10-15 years, uncertain renewable sources, electric vehicles, and smart loads will vastly change the power grid behavior and will result in new stochastics and transients that the grid has never seen nor been designed for. The enormous number of variables, the decade-long time horizon, and the inherent dynamics of the power grid make this a problem that is solvable only with exascale computers.
The project goal to develop computational models and simulation solutions related to the power grid. The researchers will formulate the optimization planning problem with uncertainty input and transient constraints, including multitime periods for look-ahead capabilities. In developing the exascale software, they will leverage several synergistic activities, including a high-performance computing library for power grid applications and work on data analytics for quantifying renewable uncertainty and correlation.
Solving this exascale problem under uncertainty will represent a significant step toward achieving the nation's goal in clean energy and reducing electricity cost for consumers.
The project is funded as a "seed" activity under DOE's Exascale Computing Project. It is led by Pacific Northwest National Laboratory. Argonne is participating together with the National Renewable Energy Laboratory.