Modeling the multi-scale, multi-variate uncertainties that impact electricity system dynamics under global change

Significant changes in electric power system infrastructure and operational capabilities are needed to decarbonize the grid and expand its reach into other sectors. The grid must also prepare to withstand increasingly frequent and severe weather due to climate change. Achieving these goals will require large investment in physical and human capital and directly involve a range of grid participants, including policymakers, regulators, system operators, investor-owned utilities, and ultimately, consumers and investors. These entities share strong interdependencies but can have highly idiosyncratic (and unequal) exposures to risk. Aggregate measures of societal importance (cost to consumers, engineering reliability, and environmental externalities) remain critical in performing decision-relevant power systems analysis. At the same time, accounting for institutional complexity (including the mechanisms that control how risk from failures is transferred among grid participants) is becoming more important as funders continue to emphasize the need for translational research; as expansion and workforce demographics open opportunities for highly skilled scientists and engineers to work in the power sector; and as equity and social justice grow in importance. Even without considering institutional complexity and potential technological and environmental changes ahead, robustly characterizing a power system's performance can be difficult due to the multi-scale, multi-sector nature of the uncertainties involved. This talk will briefly touch on the topology of power systems models used in multi-sector dynamics research, before discussing (what I see as) some interesting research gaps in the area of power systems analysis under global change. I will describe ongoing efforts to develop open source modeling platforms for capturing simultaneous, multi-variate extremes across large geographical extents; computational experiments designed to produce probabilistic measures of system performance; the need for integrating short-, mid- and long-term uncertainties; and open questions regarding the representation of institutions.