Wildfire Risk-Informed Preventive-Corrective Decision Making under Renewable Uncertainty

Abstract

The increasing frequency and intensity of wildfires poses severe threats to the secure and stable operation of power grids, particularly one that is interspersed with renewable generation. Unlike conventional contingencies, wildfires affect multiple assets, leading to cascading outages and rapid degradation of system operability and stability. At the same time, the usual precursors of large wildfires, namely dry and windy conditions, are known with high confidence at least a day in advance. Thus, a coordinated decision-making scheme employing both day-ahead and real-time information has a significant potential to mitigate dynamic wildfire risks in renewable-rich power systems. Such a scheme is developed in this paper through a novel stochastic preventive-corrective cut-set and stability-constrained unit commitment and optimal power flow formulation that also accounts for the variability of renewable generation. The results obtained using a reduced 240-bus system of the US Western Interconnection demonstrate that the proposed approach increases the resilience of power systems across multiple levels of wildfire risks while maintaining economic viability.

Figures (11)

• Figure 1: Contingency analysis and uncertainty modeling done in this paper
• Figure 2: Implementation flowchart of the proposed risk-aware scheduling and dispatch framework integrating contingency analysis with day-ahead S-CSCUC and real-time S-CSCOPF formulations
• Figure 3: Uncertainty analysis for solar and wind generation. Figs. \ref{['fig:a']} and \ref{['fig:b']} show the irradiance and wind speed variations, and Figs. \ref{['fig:c']} and \ref{['fig:d']} map these distributions to the solar and wind generation output
• Figure 4: Uncertainty analysis for wildfire risk. Fig. \ref{['fig:fire_risk_variations']} shows the variation data for the risk of a particular line. Fig. \ref{['fig:fire_risk_regions']} maps the fit distribution to the electrical fire risk $\lambda$
• Figure 5: Geo-referencing the contingency list location. Wildfire risk in the same region (in April 2024) shows high risk in the region the lines run through.