Quantifying the Effect of Plug-In Electric Vehicles on Future Grid Operations and Ancillary Service Procurement Requirements

Abstract

As plug-in electric vehicles (PEVs) grow in popularity, there is increasing research interest in the interaction between PEVs and the electric grid. Much of the previous work in the literature relies on an assumption that PEV charging will be scheduled, and that the duration and magnitude of charging loads can be modulated to suit the needs of the utility and the system operator. While access to the data or owner input necessary for charge scheduling and management might be technically feasible today, it is unclear whether vehicle owners will be amenable to providing these data or accepting utility control of their charging choices. Because of these uncertainties in the future relationship between electric utilities and PEV owners, this study seeks to examine the market effects of vehicles in the absence of the additional data utilities would need to realize these alternate, “optimal” PEV charging scenarios. In particular, this study focuses on quantifying the potential uncertainty in vehicle charging loads on an energy and power basis. Monte Carlo methods were applied to vehicle trip data from the National Household Travel Survey (NHTS) to generate simulated driving profiles for individual vehicles. Using these profiles, six PEV fleet sizes were studied, ranging from 1,000 to 500,000 vehicles, to determine whether fleet size had a linear effect on the stochasticity of vehicle charging loads. Following the Monte Carlo simulations, these loads were examined independent of and compared to net load (load minus wind generation). Results from the Monte Carlo simulations indicate that even for the largest PEV fleet sizes studied, variability in average charging loads is on the order of 10 MW, less than 0.2% of the magnitude of charging load for those fleet sizes. In comparison with electricity demand in the Electric Reliability Council of Texas’ (ERCOT) operating area, these charging loads represent a 1% increase above typical summer peak loads. Unfortunately, while the relative increase in demand is small, the timing of peak charging load is nearly coincident with summer peak demand. The simulation approach was validated by comparing the results against empirical vehicle charging data collected by the Pecan Street Research Consortium from households in Austin, Texas. Simulated and empirical vehicle charging trends showed generally good agreement, with similar charging times but slightly different charging durations. The alignment between the two charging profiles indicates that the simulation methodology applied here with NHTS travel data can be used to predict electric load for vehicle charging when empirical historical charging data are not available.