Why Your Utility Needs to Know the Whens and Wheres of EVs Right Now (and What to Look for in the Right Tool)
image credit: Courtesy of Pixabay (with filter added).
- Aug 1, 2019 7:57 pm GMT
- 2049 views
If the producers of “Who Killed the Electric Car?” and its sequel “Revenge of the Electric Car” completed the trilogy—as all good films come in at least a three-movie box set these days—they should title it “The Rise of the Electric Car.” (And they should call Arnold fresh off his turn as fake car salesman Howard Kleiner.)
In a bit over a decade since that first documentary, my how things have changed.
Today, the numbers predict massive—you could even say explosive—growth in electric vehicles (EV), with some estimates topping 100 million on the road by the end of the next decade. That speaks volumes about changes in social culture, in sustainability thinking and in pure human adaptability.
It also speaks volumes about new demand and disruption coming to every utility’s grid.
To get in front of that demand and disruption—and to make the disruption into a real opportunity—utilities have to do one thing first: know when and where EVs are plugged into their network. Once those EV numbers are proven points in a utility’s data arsenal, it opens all sorts of possibilities from new customer engagement moments around charging options and rates to inventive business models that fold in vendor partners and new market ops.
Now, you may be already on board with this thinking. You may even have been already researching and googling “EV detection” with the idea of getting ahead of this by jumping in with both feet.
You already know the opportunities. You’re excited about them. You want to put them into play at your utility tomorrow and see your customer comms get smarter and more personalized and your SAIDI/CAIFI scores see a nice bump from improved planning and lowered op costs.
But, buyer beware. Not every EV detection (the wheres) and disaggregation (the whens) tool is created equal. And it’s harder than you think to really dis correctly—to properly source a given spike to an EV (rather than, say, an AC or a fridge or a pool pump kicking on).
So, when shopping for that critical EV detection tool, keep these questions in mind:
- Is it based on years of research and lots of eyes on tons and tons of time-series-based end-use data? Or is it just a best guess estimate, a knock-off built on assumptions?
- Did the team behind the tool build model upon model upon classification model to unpack and unearth the very neural network details of the human brain’s learning process? Or did they just look at available data and make a few quick notes?
- Can it distinguish between spike causes down to the minutia because it’s chocked full of machine learning based on literally billions of rows of data about appliance signatures? Or is its reading of charging patterns a bit spotty and unreliable?
- Once deployed, will it not just detect but understand interactions, patterns, choices and clusters? Or is it a simple one-trick tool without adaptive capabilities to grow as EV adoption grows and your grid sees EV connections double or even triple?
- Is it compatible with your data backhaul schedule in place (15-min or hourly), or does it require massive investment in sub-minute or even sub-second data that is completely unsustainable?
The answers to these questions will make all the difference as you plan ahead to successfully manage widespread EV penetration across your network.