Forecasting Pennsylvania’s EV Future: Why Grid Planners Need to Pay Attention Now

As electric vehicles (EVs) roll into garages, highways, and loading docks across the United States, the power grid is facing a significant transformation. While most drivers are focused on how far their next charge will take them, grid operators are asking an equally important question: when and where is all this new EV charging demand going to hit the electric system?

Anticipating and managing these new loads is crucial. If done well, it could help utilities avoid costly and potentially unnecessary grid expansions. If left unaddressed, utilities risk being caught off guard by sudden spikes in electric demand, leading to overloaded feeders, service disruptions, and expensive last-minute infrastructure upgrades.

That is why the Pennsylvania Public Utility Commission (PA PUC) tasked DSA with developing a 15-year granular forecast of the impact of EV adoption as part of a statewide study on transmission and distribution (T&D) capacity. These forecasts can allow utilities to plan for future changes in demand while also identifying opportunities to defer or avoid infrastructure upgrades.

Figure 1: Forecasted HE19 EV Demand across Pennsylvania in 2030
Figure A: PECO	Figure B: PPL
Figure C: Duquesne	Figure D: FirstEnergy

We developed a granular, location-specific EV load forecast across the four major electric utility territories in Pennsylvania:

Step 1 – Forecast Statewide EV Energy Consumption

We used PJM’s zonal EV energy consumption forecasts to develop a 15-year projection of statewide EV energy consumption specific to the service territories of each of Pennsylvania’s four major electric distribution companies (EDCs).

Step 2 – Forecast EV Energy Consumption by Charging Type

We divided the total energy forecast into three main types of EV charging: light-duty vehicle (LDV) at home charging, LDV public/workplace charging, and medium- and heavy-duty vehicle (MHDV) charging. Each charging type was assigned a different hourly load shape; the LDV shapes were borrowed from NREL and the MHDV shape was borrowed from PJM’s electric vehicle forecast.

Step 3 – Estimate Feeder Propensity

After mapping feeders to locations, each feeder on the grid was assigned a propensity score for EV adoption and the three types of EV charging:

• LDV home charging: based on EV percent of current vehicle registrations by zip code (taken from PennDOT data)
• LDV public charging: based on nearby public charger locations (from AFDC data) and proportional to the quantity of public charging ports at each feeder
• MHDV charging: based on commercial property data for transportation-heavy industries, proportional to commercial square footage at each feeder

Step 4 – Calibrate to the Statewide Forecast

Using a calibration function developed by DSA, the forecasted EV growth for each utility was distributed across that utility’s feeders, ensuring that local adoption is calibrated to every location’s propensity and adds up to the utility-wide and statewide total forecasts.

Step 5 – Estimate Hourly Loads

Annual feeder-level EV MWh were converted to hourly loads using the normalized load shapes for each charging type. This enables planners to see not just how much demand will grow, but at what time during the day loads will peak on each circuit.

Figure 2: Example of Projected Hourly EV Load for a Duquesne Feeder in 2025 and 2030

This type of granular forecasting is key to identify T&D upgrade deferral opportunities or target T&D upgrades. Instead of making system-wide upgrades, utilities can use forecasts like these to invest where upgrades are really needed and avoid investments where they are not.

As Pennsylvania and other states prepare for rapid growth in electric vehicles, location-specific T&D forecasting will be essential to making smart, cost-effective infrastructure investment decisions.