Electric vehicles (EVs) and other distributed energy resources (DERs) are popping up all over the grid faster than many utilities can keep on top of.
EV detection is how you can start to close the gap this creates: identifying where these assets are and how large they are, using data you already collect.
This blog covers:
- Why isn’t registration data enough to detect EVs?
- Where is the EV visibility gap?
- What does accurate detection require?
- Who benefits from EV detection data, and how?
- Where does Electron fit in?
- FAQs
Why isn’t registration data enough to detect EVs?
EV registration data is a reasonable starting point when looking for answers about how to better understand that EV load, but it doesn’t fully answer everything. For example:
- Location: A registered address isn’t necessarily a charging address. People split time between homes, or register in one place and charge in another.
- Size: Registration data can’t tell you if someone has a level 1 charger or a level 2 charger that can double their home’s peak load.
- Clustering: Adoption bunches geographically. A territory can look fine in aggregate while individual transformers are already stretched.
Read more: Why utilities need more than EV registration data
Where is the EV visibility gap?
Most utilities see clearly at two points: the substation, and the household, but there’s a gap that sits between them at the service transformer and behind-the-meter level.
It’s at this point that EV problems surface first, creating additional challenges:
- Planning uncertainty: Investment is decided without knowing where the load is
- Operational uncertainty: It can cause reliability issues with no visible cause
- Lost value: Flexibility and the value it brings is only partly or not at all tracked
Read more: Why DER detection will make or break the next decade of grid planning and operations
How does EV detection work?
EV detection can support utilities in overcoming the lack of visibility and relevant data. There are different ways to approach it, with differing outcomes:
| Short-run analytics | Long-run analytics |
| Predicts every reading (e.g. every 60 min) | Analyses a wider window (e.g. a week) |
| Solves presence and size in one step | Separates presence from size |
| Noisy. Easily fooled by loads like heat pumps | Filters noise. More defensible results |
Long-run analytics uses standard AMI data that utilities can easily access, typically at 60-minute intervals. It skips past the need for real-time MDMS integration and delivers value in months, or sometimes weeks.
Who benefits from EV detection data?
A utility is made up of different teams that have the same high-level goals – deliver a reliable and affordable grid – and team-based goals, which analytics can support in different ways.
| Team | Their question | What detection gives them |
| EV program teams | Where do we even start? | A targeted list, not a blanket campaign – lower recruitment cost |
| Grid monitoring | How are EVs impacting my system? | Data-driven insights on the impacts of electrification on grid assets, both at the asset level and as general trends |
| Planning | How should EVs impact my planning? | Meter-level penetration that can be rolled up to transformers, feeders, and substation to inform next-generation planning |
| Operations | Which transformers need attention now? | Hyper-local visibility that system-wide signals can’t provide |
Read more: What EV detection can unlock for different utility teams
Where does Electron fit in?
ElectronCompass applies the long-run, decoupled approach above to a utility’s existing AMI data.
It fuses EV and DER detection with transformer loading, giving planning, performance, and operations teams one shared, meter-level view, avoiding the need for a lengthy integration project.
FAQs
EV detection identifies where electric vehicles are charging, and how large their chargers are, using a utility’s existing smart meter data and advanced statistics/machine learning.
Standard revenue-grade AMI (smart meter) data, typically at 60-minute intervals.
No. The long-run approach above works with standard 60-minute AMI data, which most utilities already collect.
No. Level 1 charging draws relatively little extra power making it both difficult to distinguish from other household loads and an insignificant signal for grid planning.
