Electric Grid UQ

Uncertainty Quantification in Electric Grids

The analysis and reduction of electric grid models under uncertainty is important to developing robust and reliable smart-grid architectures and motivate the following ongoing efforts.

Sandia’s computational researchers are investigating the dynamical behavior of ordinary differential equation network models under uncertainty. They have explored the eigenstructure of these systems, identifying the range of variability in dynamical behavior corresponding to given parametric uncertainties and outlining means of model reduction allowing for these uncertainties.

Also, algorithms have been developed for Bayesian inference of the joint uncertain input parameter space for any given model, in the absence of raw measurements, but given constraints corresponding to information from prior measurements. This data free inference approach is key to establishing self-consistent input uncertainties when original data is not available.

Finally, the context of structural uncertainty, corresponding to uncertain information regarding network integrity in case of faults or other disturbances to normal operation, is being addressed.