Attendees of last week’s Grid Edge Live conference agreed that widespread solar PV penetration will be the key disruptive influence on utilities in years to come. But just when and where this disruption will happen first, and how much money utilities must spend to mitigate it, is harder to pin down.

That makes potential solutions to that disruption subject to far more debate, particularly when it comes to getting regulators to approve big increases in customer rates to pay for it.

Technology vendors seeking to manage rooftop solar, behind-the-meter energy storage, demand-response-enabled homes and other agents of the grid edge need to have some ROI-positive solutions to problems utilities face right now. If they can leverage existing assets -- say, millions of smart meters -- all the better.

For Dominion Voltage Inc. (DVI), controlling grid voltage by using smart meters as bellwether sensors is today’s use case. The company, a creation of Virginia-based utility company Dominion Resources and Lockheed Martin, has deployed its patented technology to deliver volt/VAR optimization (VVO) and conservation voltage reduction (CVR) in projects with Dominion Virginia, Naperville, Illinois, Oregon’s Central Lincoln People's Utility District, Hawaiian Electric and California’s Modesto Irrigation District and Glendale Water & Power.

But Todd Headlee, DVI’s executive director, said at last week’s conference that VVO and CVR are just two possible uses for his company’s underlying technology. He mentioned two new projects -- one with Hawaiian Electric Co. and another with an as-yet-unnamed California utility -- that are looking at rooftop solar inverter data to help solve the voltage problems that come with lots of PV panels on a grid made for one-way power flows.

“Smart inverters, by themselves, can have a positive impact, not only for the homes they’re connected to, but for the grid,” he said. “When you control inverters as a fleet with our Edge product, the value is much greater than when they’re working autonomously. We’re helping them figure out how to do that.”

It also presents new challenges, including the need to anticipate second-by-second changes in sun and cloud cover, and how to bring customer-owned assets into alignment with utility needs. But with an architecture already optimized for deployment on the smart meter network, Headlee sees DVI scaling up to the solar challenge quickly. Here’s how he described DVI’s unique meter-as-sensor approach to conserving grid voltage, and how it lends itself to more challenging grid edge tasks.

From model information to grid automation

DVI differs from the vast majority of smart meter-assisted CVR technologies in the field today in that it doesn’t use grid models to do its job, Headlee said. That’s an important distinction, because it highlights a break from traditional top-down approaches to the challenge.

The vast majority of CVR and VVO schemes are based on transmission grid power flow modeling, scaled down to the distribution grid’s scale. These systems feed their data into a software model of the grid, which informs grid operators and automated devices what voltages on the system should be, he explained. In other words, the smart meter data is informing the model, rather than enabling action on the grid.

But those model-based systems, while accurate at the transmission-grid level, are problematic for the distribution grid. The original GIS data tends to be spottier, and constant repairs and retrofits are often left unrecorded. In cases like these, AMI voltage detection can tell there’s a violation, but it may not be able to tell why exactly in the context of the model.

Without knowing the precise voltage at the endpoint, utilities have to play it safe and keep voltages higher. Each percentage point of voltage reduced equates to roughly 0.8 percent reduction in energy bills, which adds up quickly across millions of endpoints.

By using a distributed, self-informing, method that can actually activate voltage regulation equipment on the grid, DVI can squeeze several additional percentage points of efficiency more than typical model-based CVR deployments -- in some cases, double the savings, Headlee said. No need for models also means it’s faster to deploy.

DVI isn’t the only CVR technology provider moving past model-based approaches. Utilidata, a twenty-year-old technology company that’s raised $25 million in venture capital in the past few years and landed projects with American Electric Power and National Grid, also distributes its voltage management intelligence in hardened substation routers that connect to grid voltage sensors and control devices.  

But where Utilidata requires those sensors to do its distributed voltage optimization tasks, DVI’s sensors are already deployed, in the form of smart meters from its partners, Headlee noted. That list includes Silver Spring Networks, Landis+Gyr, Elster and most recently, Itron, and the company expects to announce work with Sensus in the near future, which closes the list on big North American meter vendors.

That opens up potential for adding CVR to existing smart-metered parts of the grid, or helping to pay for new deployments, he noted. In some of DVI’s most successful deployments, “the benefits are so strong, they pay for the entire deployment of the AMI,” not just the costs of DVI’s software, he said.

Building the bridge from AMI to smart inverters

Smart meters have their limits. Most of the wireless mesh advanced metering infrastructure (AMI) network out there today has practical limits to how quickly it can transmit data, for instance.

“AMI has a kind of a latency -- it can be five, ten minutes before you get the data from the system,” Headlee said. “AMI is not perfect. Many times there are gaps in the data, and your system has to continue to work appropriately.”

DVI combines faster data from a select set of meters with historic data to predict and adjust its voltage control regime to suit these data constraints, he said. But solar’s second-by-second fluctuations are another matter. “If you’re trying to detect a cloud bank coming over, it isn’t quick enough,” he said.

DVI and its partners are looking at several options to solve that problem, including routes around the smart meter like home broadband connections. “Homeowners can make their Wi-Fi available to their smart inverter, or to their regular inverter for that matter,” said Headlee.

Silver Spring Networks, DVI’s partner on its 5,200-customer Hawaii solar-grid management project, is also integrating its software into its SilverLink Sensor Network platform, which promises to put the computing power of smart meters and other networked devices to use for fast-reacting grid tasks.

Rooftop solar can also lead to overvoltage situations that don’t fit into the response profiles of centralized CVR and VVO systems. Some circuits can be inundated with power from one or two big PV systems, while others are still demanding grid electricity as usual.

Traditional grid equipment breaks down more quickly under these kinds of hour-to-hour stresses. But smart inverters, along with grid power electronics from companies like Varentec, Gridco, and GridBridge, could help address these localized challenges.

As for the spread of smart inverters, Germany already requires certain advanced features from its solar inverter fleet, which has driven broader adoption in European markets. In the United States, California is on its way to crafting smart inverter standards that could serve as a nationwide model.

Determining how best to network, monitor and control these assets is a more complicated matter. Research projects around the country are integrating solar PV with grid controls, energy storage and demand response to see which grid management technologies work best at the job. Grid giants like Siemens, Alstom, General Electric, Schneider Electric, Hitachi and Toshiba are promising distributed energy resource management systems (DERMS) to come from efforts like these -- though the market for commercializing such technologies isn’t there yet. 

Solar-rich California may be the first place to watch that market emerge. The state is mandating new distribution grid planning, smart inverter tests and grid-scale energy storage for utilities manage its growing share of customer-owned solar power. It's also the home of the first commercial forays into battery-backed distributed solar systems from the likes of SolarCity, SunPower and Sunverge.