Optimizing Battery and Line Undergrounding Investments for Transmission Systems under Wildfire Risk Scenarios: A Benders Decomposition Approach

TL;DR

This paper develops a scenario-based optimization framework using Benders decomposition to strategically plan battery storage and underground line investments in transmission systems, aiming to mitigate wildfire risks and reduce load shedding.

Contribution

It introduces a novel two-stage Benders decomposition approach for large-scale, scenario-based planning of wildfire risk mitigation investments in power transmission networks.

Findings

Effective on large-scale networks with real-world data

Reduces load shedding during wildfire events

Optimizes investment decisions under uncertainty

Abstract

With electric power infrastructure posing an increasing risk of igniting wildfires under continuing climate change, utilities are frequently de-energizing power lines to mitigate wildfire ignition risk, which can cause load shedding. Recent research advocates for installing battery energy storage systems as well as undergrounding risky overhead lines to reduce the load shedding during such de-energizations. Since wildfire ignition risk can exhibit substantial geographic and temporal variations, it is important to plan battery installation and line undergrounding investments while considering multiple possible scenarios. This paper presents a scenario-based framework for optimizing battery installation and line undergrounding investments while considering many scenarios, each consisting of a day-long time series of uncertain parameters for the load demand, renewable generation, and wildfire ignition risks. This problem is difficult to solve due to a large number of scenarios and binary variables associated with the battery placements as well as the lines to be undergrounded. To address the computational challenges, we decompose the problem in a two-stage scheme via a Benders decomposition approach. The first stage is a master problem formulated as a mixed integer linear programming (MILP) model that makes decisions on the locations and sizes of batteries as well as the lines to be undergrounded. The second stage consists of a linear programming model that assesses these battery and line undergrounding decisions as modeled by a DC OPF formulation. We demonstrate the effectiveness of the proposed scheme on a large-scale transmission network with real world data on wildfire ignition risks, load, and renewable generation.