Resilient Transmission Planning Using Data-Driven Models Under Climate Change Impacts

Abstract

Transmission expansion planning (TEP) traditionally optimizes investments in new lines, reconductoring, or capacity upgrades to minimize long-term costs while satisfying reliability, economic, and increasingly decarbonization objectives under deterministic or limited uncertainty assumptions. Climate change, however, introduces profound non-stationarities through rising ambient temperatures, more frequent and intense heatwaves, altered precipitation patterns, and compound extremes that derate generation capacities, reduce transmission line ampacity via thermal sag and increased resistance, elevate cooling-driven demand surges, and heighten outage risks. Data-driven models leverage high-resolution climate projections (e.g., Thermodynamic Global Warming or TGW datasets), historical weather reanalysis, SCADA/PMU telemetry, outage records, and geospatial vulnerability layers to embed these impacts directly into planning frameworks. This research paper develops a comprehensive resilient TEP approach using data-driven robust optimization, distributionally robust optimization (DRO), stochastic programming with climate-informed scenario ensembles, and machine learning surrogates for uncertainty quantification and surrogate power flow/heat balance computations.