Joint venture has IBM tackling renewable energy’s grid effects

Wind turbines on the Lowell Mountain ridgeline. File photo by Andrew Stein/VTDigger
Vermont Electric Power Co. and IBM have announced a new venture that will have the computing giant analyzing local weather and electric meter data to anticipate conditions on the grid as increasing reliance on wind and solar energy makes that difficult.

The effort is predicted to save Vermonters money and perhaps someday involve remote control of customers’ electric use to maximize savings.

The partnership comes in the form of a business called Utopus Insights Inc., which is based in Valhalla, New York, and Bengaluru, India.

The company is a startup, with IBM’s help, and with VELCO serving as both an investor and a strategic partner.

The move is an unusual one for both companies, said Utopus Insights President and CEO Chandu Visweswariah, who is an IBM Fellow.

“It’s something very novel that (VELCO is) trying … and I don’t think that IBM has ever done anything like this,” Visweswariah said.

This is IBM’s first entrepreneurial startup, he said, although IBM purchases and divests itself of companies “all the time.”

What VELCO’s done is unusual as well, he said.

The electric transmission company identified future challenges with a changing electrical grid, developed solutions over the past three years in collaboration with IBM, and has now signed on as an investor in the new company, Visweswariah said.

Chris Recchia
Kerrick Johnson, right, is a spokesman for VELCO. File photo by Mike Faher/VTDigger
VELCO is also a strategic partner with Utopus Insights, said VELCO spokesman Kerrick Johnson. The transmission company’s president and CEO, Tom Dunn, will serve as chairman on Utopus’ board of directors.

When VELCO first contacted IBM several years ago, Johnson said, the transmission company faced new difficulties resulting from the state’s increasing dependence on weather-based forms of renewable energy.

Much of the company’s interest at the time was in forecasting weather as a way to anticipate electrical supply.

“We were experiencing much more frequent, much more severe, much more costly extreme weather events,” Johnson said, “and we could see, if this pace of renewable energy growth continues, it’s going to begin to impose significant challenges to balance supply and demand.”

“We felt the riskiest strategy, the riskiest place we could be, would be to do nothing,” Johnson said.

VELCO and IBM first contracted with each other 3½ years ago, Johnson said, and the effort was successful for both entities. Software IBM developed to meet VELCO’s needs at that time has become the industry standard, Johnson said.

This next venture builds on that, but the idea remains the same: improved forecasting for the electrical grid using machine learning.

Because of its dependence on the weather, renewable energy often introduces uncertainty into the electrical grid, Visweswariah said, and his company aims to reduce that by forecasting both the weather and electrical demand.

IBM already owns The Weather Company, after a 2015 purchase that includes the and Weather Underground websites.

Utopus Insights will marry IBM’s weather data with precise, anonymous information on electrical demand that’s gathered at the level of individual electric meters, Johnson said. This will enable electricity supply and demand models for neighborhoods, communities and the entire state, he said.

Vermont possesses several attributes that made it an ideal place to put a concept like this into effect, Johnson said.

For instance, 94 percent of Vermonters have what’s called a smart meter on their home, which is a device that sends and receives information to and from the grid, he said. Vermont also has a 1,500-mile, 72-strand fiber-optic cable running from border to border, connecting many of the state’s electrical substations, he said.

In the end, Vermonters will save money, Johnson and Visweswariah said.

A major part of ratepayers’ bills each month consists of charges related to peak demand, Johnson said. Peak demand occurs when Vermonters consume the greatest amount of electricity, and the New England electrical grid operators assess a fee based on how great that peak demand is.

Reducing peak demand reduces these fees, which would significantly reduce ratepayers’ costs, Johnson said.

This reduction in peak demand “might well be” accomplished someday through remote, centralized control — for example, by preheating homes when the sun’s shining, or by charging vehicles when the wind’s blowing, Visweswariah said.

“If indeed there is control over what happens behind the meter … it will be, and should be, with your knowledge, and cooperation, and to save you money and meet your goals,” Visweswariah said.

Along with predicting and managing peak loads, the Utopus Insights software will anticipate icing events. These can knock down power lines, block solar panels and cause wind turbines to emit unpleasant sounds.

The new company may not employ large numbers of Vermonters, but it will attract people from other states and countries who want to see the concept of electricity supply and demand modeling put into practice, Johnson said.

If you read us, please support us.

Comment Policy requires that all commenters identify themselves by their authentic first and last names. Initials, pseudonyms or screen names are not permissible.

No personal harrassment, abuse, or hate speech is permitted. Comments should be 1000 characters or fewer.

We moderate every comment. Please go to our FAQ for the full policy.

Mike Polhamus

Recent Stories

Thanks for reporting an error with the story, "Joint venture has IBM tackling renewable energy’s grid effects"
  • Matt Fisken

    Vermont has had microwave transmitting “smart” meters for about five years now. By the time any of this pans out, the solid state (computer chip-based) meters will probably need to be replaced as they begin to fail. The smart grid’s dirty little secret is that it actually uses more energy than it could ever save through automated power management. The agenda is to sell more meters and computers while enabling the utilities to gouge with variable pricing schemes.

    The 6% of Vermonters who decided to keep their analog meters were really SMART.

    • JohnGreenberg

      Matt Fisken:

      1) “By the time any of this pans out, the solid state (computer chip-based)
      meters will probably need to be replaced as they begin to fail.” Do you have some evidence about the life span of these meters?

      2) “The smart grid’s dirty little secret is that it actually uses more
      energy than it could ever save through automated power management” Please document this claim.

      • Matt Fisken

        Smart meters have an estimated lifespan of 5-15 years, compared to analog meters which last 30-50 years. Note: Most of the analog/electromechanical meters GMP replaced in 2012 were less than 10 years old when they were scrapped.

        A “secret” is information that is not documented. What we do know is that GMP’s 250,000 smart meters and 250-500 accompanying gatekeepers draw power from the grid to power circuitry and RF transmissions without accounting for that usage. Unlike phantom power draws within a home, there is no way to quantify how much is being used by smart grid equipment.

        The question is, how many antennas, meters, in home devices, smart phones and apps does it take to turn off a light bulb? 😉

        • JohnGreenberg


          In other words, you’re just guessing.

          A Google seach found this:

          “When a smart meter contains an RF transmitter:

          The frequency of operation is
          typically in the 902 MHz and 2.4 GHz bands.

          Power output is typically 1
          watt in the 902 MHz band and much less in the 2.4 GHz band.

          The smart meter only
          communicates when it is commanded to do so, typically several times a day.”

          And another found this: “The actual percent of time the Smart
          Meter is transmitting, especially in the initial years of operation, is very
          small, usually less … than 15 minutes accumulated total per day.”,
          p. 31

          That would mean that the smart meter would require
          approximately 1 watt hour every 4 days = about .09 KWh per year or 1 KWh every 11
          years or so.

          • Matt Fisken

            Power consumption is not the same as antenna power.

            I have found smart meters in Vermont transmitting dozens of times per minute. This was corroborated by the Richard Tell study.

            The big picture is that the unnecessary replacement of functional, passive equipment with more expensive gear that consumes power and has a shorter lifecycle is the opposite of “green.”


            This document underlines the fact that this whole scheme depends on the meters lasting for 20-25 years, despite the strong likelihood that they won’t.

            When all those smart meters need replacing in a few years, they won’t be paid for by matching federal stimulus money. They will become another line item that the utilities simply pass along to ratepayers. I guess we all have different views of what is “sustainable.”

          • JohnGreenberg


            “Power consumption is not the same as antenna power.” How are they different? How is the difference relevant to your assertion that the smart grid “actually uses more energy than it could ever save through automated power management?”

            You say: “This document underlines the fact that this whole scheme depends on the meters lasting for 20-25 years….” It does indeed state that “utilities plan to recover costs and realize additional savings during the remainder of the planned 20 to 25 year period,” (p. 22) but that does NOT imply that it will actually take 20-25 years to fully recover costs. I didn’t see anything in the report indicating when costs would be recovered.

            Your “big picture” assertion ignores all benefits from meter switching to reach your pre-ordained conclusion that the smart grid isn’t green. If, as your link also suggests, these meters save carbon, reduce outage costs, decrease energy costs, and encourage energy conservation, the big
            picture looks pretty different.

          • Matt Fisken

            “how are they different?” does a computer with a 1-2 watt wifi card only draw 1-2 watts of electricity when powered on?

            “how is the difference relevant…?” Because you can’t simply pull random numbers off the internet and say that a smart meter draws .09 KWhs/year, implying that the power consumption of all smart grid infrastructure is insignificant.

            “I didn’t see anything…” While the report could have been more informative, and each utility is in a unique situation (for example, VEC invested in AMI well before ARRA using less expensive technology), it’s best to examine GMP which has spent $105 million and purportedly saved $19.5 million in 3.5 years, or $5.5 million/year. They appear to be on track to recover costs in 19 years. (Burlington Electric is on track to recover in 180 years.)

            So what happens if GMP needs to replace 250,000 meters at $250/meter after 15 years? after 10 years?

            Analogously, if I spend $10 on a new lightbulb because I hope it will save me $20 over 20 years, but that bulb burns out after 10 years, what happens to those expected savings? What happens when I realize there were unintended health consequences associated with that lightbulb?

            But there’s always the Utility Narrative if you prefer it.

          • JohnGreenberg


            Computers have many more functions than just wifi. Point taken.

            So here’s another source: “the meter itself will typically consume less than 1W per hour … for most of its operating life.” 1W per hour would put the consumption at 8.76 KWh per year.

            These numbers provide ballpark estimates which show us the range we’re in. If you have better ones, provide them. At 9KWh per year, it doesn’t
            take much energy savings to realize gains, especially because you entirely ignore non-residential customers. Industrial and large commercial users, in some cases, consume VAST amounts of power, where a savings of 9KWh would occur in seconds.

            This is your assertion we’re discussing, not mine. Feel free to provide some evidence – any evidence – to back it up. Also, please address the benefits I noted which are enumerated from the link you provided above. So far, you’ve ignored them entirely.

          • JohnGreenberg

            “If I spend $10 on a new lightbulb because I hope it will save me $20 over 20 years, but that bulb burns out after 10 years, what happens to those expected savings?”

            Bulb life is rated in median hours; it is obtained empirically. How many YEARS that number represents is simply a multiple of how many hours per day the bulb is used (and actual vs.assumed switching cycles).

            Bulb savings come from reduced energy use, as well as from lower replacement costs: fluorescent and LED bulbs draw far less power to produce a given amount of light and last many times longer than incandescents (fewer bulbs).

            A bulb may reach its life expectancy in fewer years than expected because it is actually used more each day. In that case, the savings are achieved in less time, producing a higher annualized return on investment.

            Alternatively, the bulb burns out in less than the
            average number of hours. Either this isoffset by other bulbs which last LONGER than the average, or the median is empirically wrong.

  • Moshe Braner

    At peak demand times the wholesale cost of power is far higher than the retail price. Thus the rest of us heavily subsidize the high users. It is long past time to use the “smart meters” for their originally stated purpose: variable rate billing. And it should be mandatory. This would shave down the peak demand, and can be done without any further investments in either hardware or studies. No more air-conditioned stores propping their doors open.

    • Matt Fisken

      Many promises were made by Vermont officials and industry representatives regarding “smart” grid and wireless internet schemes, which gleaned more than $200 million in federal funding for the state during the ARRA years. One such claim was that the fiber optic lines spanning the state by VELCO would support broadband services and backhaul meter data. That quickly changed when VTel needed to justify its WoW network by saying it was critical for connecting unserved Vermonters AND backhauling meter data.

      That constantly moving carrot in front of your is called planned obsolescence and it is VERY profitable for the industries involved.

  • Edward Letourneau

    Statements like “…remote, centralized control — for example, by preheating homes when the
    sun’s shining, or by charging vehicles when the wind’s blowing,
    Visweswariah said,” tell us there is a limited understanding of how things actually work. Very few have electric heat, and fewer have electric cars. The projected savings are nil. — The state energy policy of renewable’s cannot technically work. In fact it is driving us towards destabilizing the power grid — which has already happened in California, which is now requiring utilities to install battery storage plants teamed up with gas turbines, pumped storage plants to assure grid stability. And worse for Vermont, if we had 90% of our power from solar and wind, we would need three 600 MW pumped storage plants to balance the renewables. There is no place in vermont to build them, and battery storage plants are fairly limited to about 20 to 40 MW for couple of hours a day — which is why they are being teamed up with gas-turbine generators. The batteries provide power for 20 minutes while the gas turbines warm up and reach and maintain full load. How many of those do the proponents want around? It will take 20 to 40 and no one is going to want them in their communities.

    • Moshe Braner

      We already have all the “pumped storage” we need, in effect. It’s called Hydro Quebec, who can easily ramp up and down as needed for the fluctutations in Vermont’s net demand (except at peak demand times, when what we need is demand-side reductions). We may need more transmission lines though. The new one proposed to go under Lake Champlain is to serve Southern New England, not us.

      • Ken Egnaczak

        ….and the sun “pumped” it, just like for all regular hydro…….see, another solar powered renewable !

      • Edward Letourneau

        So in your view its ok for other people to live with power plants that Vermont doesn’t want to look at, so that Vermont can claim it lives on renewable power. Kind of selfish, don’t you think?

    • bill_christian

      Did you know we already HAVE a 600 MW pumped hydro storage facility that could handle all of Vermont’s existing intermittent renewables and much more, right on our border with Massachusetts? So we just need two more and we are totally good.

      • Edward Letourneau

        Bear Swamp and Northfield do not belong to Vermont. Nor does their power flow from or into Vermont. Now there are interconnections to the grid, but that was not their prime purpose.