Spatial impacts of electric vehicle charging on power grid stability: A downtown Atlanta case study

Abstract:

Background: The rapid rise in electric vehicle (EV) charging demand poses a significant risk of overloading the power grid and complicating infrastructure planning. While many studies have developed methods to estimate urban-scale charging demand, the spatial distribution of this demand and its implications for power grid stability remain underexplored due to the lack of geographic association in grid data.

Objective: The objective of this study is to estimate the spatial distribution of EV charging energy consumption and assess its impact on power grid performance using real-world geographic data. Our study contributes to the literature by integrating EV charging demand estimation with spatial analysis of power grid stability using real-world geographic data, providing valuable insights for practical EV and grid infrastructure planning.

Method: We estimated charging energy consumption in downtown Atlanta, Georgia, and mapped it spatially onto the power grid network using the IEEE 37 Node Test Feeder to assess the impact of EV charging on grid performance. We analyzed power system violations caused by EV charging in residential buildings under various conditions, including different EV charging penetration rates and types of charging scenarios, specifically home (single-family housing) and non-home charging (multi-family housing). The methodology involves three steps: (1) simulating trip chains at the Census Block Group level, (2) disaggregating trip chains to individual buildings, and (3) estimating EV energy consumption and assessing grid impacts from residential charging at the Census Block Group level.
1) Trip chains were simulated using an agent-based approach based on origin-destination data at the Census Block Group level from Fulton County (2019) and the 2017 National Household Travel Survey (NHTS). Since the NHTS provides limited trip data for EVs, we assumed EVs follow the same travel behavior as conventional vehicles, without detours for charging or extended stops. EVs were set to charge when their State of Charge dropped below 50%, using the nearest available station. We assumed charging could occur at single-family homes and existing stations for multi-family dwellings, with infrastructure data sourced from the Alternative Fuels Data Center.
2) Simulated trip ends were proportionally assigned to individual buildings based on their capacity to estimate how many vehicles end their trips at each building. The building capacity data were sourced from the Model America dataset, which provides information on building type and capacity for each building in the U.S. We then focused specifically on residential buildings to estimate their charging energy consumption.
3) The grid simulations were conducted in OpenDSS, a widely used power system simulation software. Grid performance was evaluated using two key metrics: minimum grid voltage, which indicates the lowest voltage level across the network, and voltage imbalance, defined as the maximum deviation from the average line voltage divided by the average voltage.

Results: The case study covered 51 Census Block Groups, with 2,785 residential buildings and a 10% EV penetration rate, simulating 67,748 EVs. Preliminary findings indicated that vehicle arrival rates peak in the afternoon, while energy consumption and grid minimum voltage are lowest between 5 p.m. and 8 p.m.