The question of whether, where, when, and what types of transmission facilities to build has been a longstanding challenge in the electric power sector. Transmission planning involves determining how to expand the grid to meet growing energy demands reliably and affordably. Decisions hinge on trade-offs: should we build long transmission lines to transport electricity from distant renewable sources, despite power losses and high costs, or construct new generation resources closer to demand centers?
These challenges are further compounded by increasing renewable energy capacity shares, increasing frequency of extreme weather events, and decarbonization goals, such as California’s target of 90% zero-carbon electricity by 2035.
In response to the challenge of maintaining a resilient power system under new uncertainties, Order 1920 by the Federal Energy Regulatory Commission (FERC) mandates that transmission providers adopt scenario-based planning that accounts for both long-term uncertainties (e.g., policy shifts, electrification trends) and short-term uncertainties (e.g., renewable energy variability, extreme weather). While advances in modeling and optimization have improved planning for long-term uncertainties, existing methods remain insufficient to address the rich diversity of short-term operational scenarios that could result in system failures, such as widespread power outages. Notably, current methods are unable to identify and prioritize all relevant low-probability but high-impact extreme events for transmission infrastructure planning.
This project aims to address this gap through two key innovations:
Efficient Extreme Event Identification:
Leveraging recent advances in optimization under rare events, led by PI Subramanyam, we will create a tool to identify and prioritize extreme event scenarios. The tool will generate and analyze time series profiles of electric load and renewable generation leading to extreme system states, by first identifying the most likely extreme event scenarios and then employing importance sampling to generate and weight other plausible outcomes efficiently.
Enhanced Operational Modeling:
To address the oversimplified system representations common in existing tools, we will develop operational models that incorporate system dynamics such as long-duration energy storage, operating reserves, and unit commitment-based market-clearing mechanisms. Extending work by co-PI Webster in collaboration with the transmission planning group at PJM, this more realistic representation of grid operations will facilitate the identification of resilience-improving investments.
This proposal directly addresses a priority in the power industry to provide system planners with tools to identify the most critical scenarios that could lead to extreme events. Because existing planning processes can only consider a small number of scenarios, the proposed research will provide a systematic way to identify the scenarios that matter most to ensure grid resilience.
Researchers
Anirudh Subramanyam
