Sandia Develops Stochastic Production Cost Model Simulator for Electric Power Systems

Sandia has developed a software toolkit that uses stochastic programming to perform power system production cost model simulations. Named PRESCIENT, the software produces probabilistic forecasts automatically from deterministic historical forecasts for load, solar, and/or wind power production and their respective actuals, using a technology known as epi-splines.

An example of a probablistic solar forecast produced with PRESCIENT.

The software also generates probabilistic scenarios that are fed into the stochastic unit commitment engine. This allows stakeholders who have no access to probabilistic forecasts—or the expertise to generate them—to create an explicit representation of the uncertainty in the system load, solar generation, and/or wind generation. Because of this explicit uncertainty representation, reserve margins and production cost can be substantially reduced, even under high penetration of renewable energy. PRESCIENT’s stochastic unit commitment uses an efficient technique to solve hundreds of scenarios in tens of minutes using commercially available solvers. PRESCIENT also has visualization capabilities; it produces daily generation dispatch stack graphs and a cost breakdown of commitment and fuel costs.

The software was developed using Pyomo, a python optimization modeling language developed at Sandia, and was supported by the US Department of Energy’s SunShot Initiative and ARPA-e GENI program. PRESCIENT was developed in partnership with the University of California–Davis.