AMI-Driven Insights for Personalizing Utility Programs

Abhay Gupta's picture
Founder & CEO Bidgely

Abhay Gupta is the CEO of Bidgely, a software company that enables utilities to leverage the power of AI to optimize engagement, reduce operational costs, and serve 100% of homes (smart meter and...

  • Member since 2018
  • 10 items added with 20,380 views
  • Nov 11, 2019

This item is part of the Special Issue - 2019-11 - Artificial Intelligence, click here for more

Advanced Metering Infrastructure (AMI) is becoming ubiquitous throughout the utility industry, providing many sources of value for both utilities and ratepayers as utilities unlock previously unavailable insights to help customers understand their consumption and use that information to take action.

Navigant Research summed up the current industry transformation in its most recent Home Energy Management Overview Q3 2019, saying, “hardware devices such as smart meters and advanced thermostats can be useful for the data they provide and for sensing conditions in a home. The problem is not data volume per se. The issue for utilities is how to skillfully organize and analyze data in ways that unlock enduring value for customers. The software tools and approaches from a decade ago must give way to ones that provide deeper insights and can support more personalized customer engagements.”

Just as consumer engagement experts like Netflix leverage customer interaction data to inform their strategies, utilities have the opportunity to create a comprehensive energy profile for every single customer that drives targeted engagement and efficiencies for utility programs.

Consider electric vehicles (EVs), which present both a challenge and an opportunity for utilities. If they are integrated properly, they offer increased revenue potential. But at the same time, they threaten grid stability. AMI-based EV detection and analytics unlock the potential for smart vehicle-to-grid (V2G) integration, targeting EV owners precisely and ensuring the EV adoption curve ramps up smoothly to the benefit of both utilities and customers.

A Deeper Dive: AMI-based Detection and Analytics for EVs

The International Energy Agency (IEA) estimates that there will be between 125-220 million EVs on the road by 2030. Most of those vehicles will be charged at home. Given the massive opportunity for EV ownership to increase utility revenue, reduce carbon emissions and reinforce the utility / customer relationship and the significant potential pitfalls caused by increased localized load on the grid, utilities are moving quickly to introduce AMI-data based platforms. 

In the 2019 update of its Utility 2.0 Long Range Plan, for example, PSEG Long Island identifies EV Analytics as one of two high priority programs for 2019 implementation. The utility tells regulators that its “AMI-based analytics solution provides insights on EV customer behavior, their charging patterns as well as the potential impact of these EVs on the grid” and goes on to say that “PSEG Long Island will be able to better serve its customers by designing new rates options for them along with EV programs tailored to meet their needs.” Further, the report says, “EV Analytics will also mitigate any potential transformer overload conditions associated with projected EV growth and adoption on Long Island” and allow the utility to “proactively repair or replace transformers susceptible to overloads due to multiple EV customers, resulting in a more reliable network for PSEG Long Islands customers.” 

There are three essential AMI data points that enable utilities like PSEG Long Island to more accurately anticipate the effect that EV charging will have on the grid, streamline rate and grid planning, improve customer relationships and realize greater EV revenue potential. 

1)                  Number and location of EVs

Because EV ownership often grows in pockets, the effects on the grid are typically highly localized. Identifying feeder lines and individual transformers that are likely to come under strain due to multiple EVs should inform forecasting, grid upgrades and grid planning, as well as to define new EV tariffs and target roll outs of EV rates or controlled chargers to encourage load shifting. 


2)                  EV Charger size

The size of the chargers in use is directly proportional to the impact on the grid. For example a single 10 kW charger has a much greater impact on the grid than two 5 kW chargers. Data revealing a region’s charger size mix allows utilities to prevent grid strain and successfully implement charger optimization programs. 

3)    Time of charging

Understanding the time basis of EV charging is key in calculating peak load impacts and designing successful EV rate programs and load shifting initiatives. Using AI-powered heat maps, utilities can identify periodic EV charging, which appears on the map as a high intensity horizontal “streak.”

 [Caption: This customer is already exhibiting behaviors consistent with that of a customer on a TOU rate plan or load control, since their EV charger seems to consistently be turned on at approximately 9 pm and runs as long as necessary overnight. This customer could be targeted with positive outreach encouraging them to keep it up and reminding them how much they are saving by charging overnight vs. a typical home, assuming they are already on a TOU rate.] 

 [Caption: This home charges their EV with a ~10kW charger periodically throughout the day. As a result, they are putting a lot of strain on the grid, as they are plugging in during summer peaks when every other home (including this one) would be using a lot of air conditioning.] 

Using these three data points, utilities can then gauge the impact of EV on operations under different scenarios, and identify specific steps to implement more targeted, less costly and more successful EV programs that maximize opportunities while mitigating challenges. 

# of Transformers at Risk per Year


 AMI-informed action plans reveal clear paths to incentivize EV ownership while also maintaining grid balance through EV rebates on high speed charges and EV-specific TOU rates and EV-charger load shifting incentives that increase utility revenue and enable load balancing on the grid. 

Artificial intelligence, AMI analytics and load disaggregation are opening up new potential in connection with myriad use cases across utility organizations, such as EV programs outlined above. With the benefit of previously unavailable information and intelligence, it is now possible to better understand key drivers of various utility programs, streamline decision making, prove return on AMI investments to regulators and create a personalized energy experience for every customer. 

To download a copy of Bidgely’s AMI-Driven Insights Report that details other utility use cases for AMI analytics beyond EVs like non-wires solutions, demand side management and rooftop PV, go to:

Abhay Gupta's picture
Thank Abhay for the Post!
Energy Central contributors share their experience and insights for the benefit of other Members (like you). Please show them your appreciation by leaving a comment, 'liking' this post, or following this Member.
More posts from this member


Spell checking: Press the CTRL or COMMAND key then click on the underlined misspelled word.
Susan Brissette's picture
Susan Brissette on Dec 5, 2019

Interesting considerations for electric vehicle analytics and electricity consumption considerations for consumers.

Linda Stevens's picture
Linda Stevens on Dec 6, 2019

Extremely interesting article! Are you using and geospatial technology to look at predicting charging station locations and/or how the demand will affect the sizing and location of substations? Analyzing usage patterns by time and location would be extremely interesting and would help focus utility programs for future operations as well as current customer communications.  

Sophia Wen's picture
Sophia Wen on Jan 14, 2020

Hi Linda, our technology allows us to see EV charging by time, location and magnitude down to every home. So certainly with an understand of home charging behavior we can start predicting the impact on local substations. We couldn't agree more that this would be very useful in targeting customer for specific programs with personalization communications. Learn more and reach out to here: 

Matt Chester's picture
Matt Chester on Dec 6, 2019

Artificial intelligence, AMI analytics and load disaggregation are opening up new potential in connection with myriad use cases across utility organizations, such as EV programs outlined above

The use cases seem plentiful and even self-evident in many ways; leads me to wonder-- do you think the utility space is an area where AI integrations are on the cutting edge, or are power providers finally playing catchup to other industries? Utilities have traditionally been quite slow to move and overcome with inertia to change, but recent digital developments have shown that this may finally be changing. I'm curious where utilities rank in terms of 'leading the way' for AI specifically

Sophia Wen's picture
Sophia Wen on Jan 14, 2020

It depends where in the utility you ask. For grid planning and management, investment in AI has been growing in order to optimize truck rolls and predict against service outages. Relative to other industries they're not doing so bad. However on the customer experience and program management side - Utilities still have a long way to go and openly learning from other industries for best practices. 

Jason Price's picture
Jason Price on Dec 12, 2019

Not an EV owner (yet) but such information presented in this excellent report begs the question: what kind of influence can the data and the EV share so that the customer is empowered and optimized to make the right financial charging decision. For example, we are empowered to shop gas stations and can see the cost per gallon to the tenth of a cent. Lundeberg gas report and apps to tell us various gallon rates at nearby gas stations by zip code or geography.  How can the actual EV plus this data inform the consumer that the TOU rate may be better at another charging staton or at a preferred time? Perhaps it is less expensive to go to a charging station, or maybe between 1 and 2 pm at home, who knows. Just any possibility of provding consumer energy price info that could be meaningful. Do you see a future where this could be possible?



Sophia Wen's picture
Sophia Wen on Jan 14, 2020

Yes we do. It's hard to expect customers to understand what they should do with just the raw data so we've been working on a AI-powered interaction engine that essentially pairs energy insights with an actionable recommendation for the customer: for example "you should charge after 9pm because it would save you $30 dollars over X weeks" this reduces the friction of having an EV car, hopefully making it more appealing and easier to have one!  

Get Published - Build a Following

The Energy Central Power Industry Network is based on one core idea - power industry professionals helping each other and advancing the industry by sharing and learning from each other.

If you have an experience or insight to share or have learned something from a conference or seminar, your peers and colleagues on Energy Central want to hear about it. It's also easy to share a link to an article you've liked or an industry resource that you think would be helpful.

                 Learn more about posting on Energy Central »