Pulling a cohesive narrative out of the DistribuTech conference’s thousands of participants and menagerie of electrical doohickeys is like synopsizing James Joyce’s Ulysses.
Among the cacophany, though, a few trends emerged. The sophistication of software to manage an increasingly decentralized grid continues to grow, as Jeff St. John observed. Meanwhile, utility representatives expressed caution about the dislocations stemming from that decentralization.
“Our main charge is to make sure our customers have reliable power,” said James Boston, manager of market intelligence at San Antonio utility CPS Energy, which recently deployed a demonstration microgrid. “Utilities are generally cautious to make sure these things are tested, are proven, before we put customers’ reliability in any type of jeopardy.”
A spot of optimism from both utilities and grid edge enthusiasts appeared in the form of microgrids.
In the last few years, the drop in prices for solar photovoltaics and energy storage, the arrival of advanced grid-forming inverters, and the impacts of solar penetration on the flow of electrons through the grid have made microgrids more cost effective and attractive.
“Ten years ago, resiliency was the only reason you would buy a microgrid, because the energy would cost too much to create — it would never be cheaper than a grid,” said Mark Feasel, vice president of smart grid at Schneider Electric. “Now with PV and CHP with natural gas, in many states you can generate energy cheaper than you can buy it.”
Cheap technology, though, is not sufficient to usher in the age of the microgrid.
“The challenge is on the commercial side,” said Ken Horne, director for smart grid at Navigant, in a panel at DistribuTech. “Not all value streams from these investments are readily monetizable.”
Now, more than ever before, utilities and other energy companies are experimenting with those business models. Many models are emerging, but three distinct concepts stood out at the conference: microgrids to defer transmission and distribution costs, microgrids as a service to a customer, and microgrids developed for both customer and grid-wide benefit.
Not surprisingly, utilities tend to like microgrids when they can own them, control them and recover their costs through the rate base. If the commercial microgrid-as-service model takes off, though, it could usher in a broad deployment of this technology without asking ratepayers to put their money on the line.
Utility-owned distribution microgrids
San Diego Gas and Electric delivers power to the town of Borrego Springs via a single radial transmission line running through the desert. Lightning strikes and desert flash floods threaten that line, resulting in historically poor reliability, Chief Engineer Thomas Bialek explained at the DistribuTech panel.
The utility needed to maintain or improve reliability for the nearly 2,800 Borrego Springs customers, but the traditional fix, building out a parallel transmission line, was pricey. A microgrid would be three or four times cheaper, Bialek said. So that’s what they did.
The system, paid for by SDG&E, the Department of Energy and other partners, combines diesel generators, large and small batteries and rooftop solar PV.
SDG&E's microgrid kept the desert town of Borrego Springs powered during a flash flood-induced outage. (SDG&E)
The microgrid has already proven itself in the face of adversity. When a flash flood in September 2013 downed transmission poles and lines leading to the town, the microgrid fired up and restored power to 1,056 customers while the grid repairs unfolded. That covered the core city center, so that those residents who didn't have power yet could move to central facilities for shelter from the heat.
In this use case, the microgrid doesn’t challenge or complicate the utility’s role — it helps achieve the core mission of reliability for all customers, however remote.
“If they were redoing from scratch all the rural electric cooperatives, I doubt they would build all those miles and miles of lines,” Bialek said.
Up in metropolitan Toronto, distribution utility PowerStream had a similar conundrum. It powers a remote town called Penetanguishene well north of Toronto, fed by very long transmission lines under threat from storms and snow.
“When we tried to improve reliability in this town where reliability was historically poor, we thought, ‘We could put out more conventional assets to solve the problem or we could do something more innovative,’” said Shuvo Chowdhury, smart grid project lead at the utility.
They decided to test whether a microgrid really could defer a traditional grid investment. The installation took a year from detailed design to deployment and commissioning, and includes a 500 kilowatt-hour lithiuim-ion battery from Samsung SDI with an advanced microgrid controller and 750 kilovolt-ampere inverter.
The microgrid can island and run the town for 42 minutes, considerably longer than the usual outage of 10 minutes. When an outage is not expected, the battery can earn some revenue by arbitraging from cheap to expensive generation.
“The major play is for resiliency, but most of time you don’t use it for that,” Chowdhury said. “We’ve had a very mild winter, so we haven’t had as many outages. Not that I’m asking for some…”
There is still work to be done, then, to optimize the design of such a system for return on investment. It’s one thing to prove a microgrid is cheaper than wires upgrades, and it’s another to design and operate it in a way that pays for itself most quickly.
A different model is emerging to give customers access to microgrid service while shielding them from risk. This approach adopts the power purchase agreement model from the solar industry to sell microgrids as a service.
Schneider Electric will be supplying one of these microgrids at the public safety headquarters and correctional facility in Montgomery County, Maryland, the company announced Wednesday.
Under this structure, Duke Energy’s renewables arm will own and operate the microgrid. Montgomery County will pay a 25-year PPA for the energy created onsite as well as the secure power for the critical facilities in the event of an outage. Schneider connected the customer with the investor and supplies about half of the equipment that Duke will purchase to construct the setup.
Duke has plenty of experience developing renewable energy and maintaining electrical infrastructure, and a big enough balance sheet to absorb a large up-front investment.
“For the past 10 years, Duke Energy Renewables has been successful selling wind and solar power through power purchase agreements, so we extended that model to the Montgomery County microgrid project, meeting the customer’s need for reliability, security and affordability,” said spokesperson Tammie McGee.
Montgomery County has a triple-A credit rating, but limited expertise in the construction of microgrids, not to mention limited budget for capital investment and maintenance.
Thus, Duke can trust the county to pay incrementally over the next 25 years for resilience, energy cost stability and expanded renewables. The structure avoids the risk of using taxpayer dollars to buy the emerging technology itself.
“Not only are we preparing for outages, but these facilities are in a better condition for day-to-day operations,” said Eric Coffman, Montgomery County chief of energy and sustainability, tuning in virtually to a launch event at the conference.
In this case, the local government is the customer, but ratepayers themselves don’t bear the cost or risk of the installation. Duke owns the assets as a private investor, the county pays for service. It’s easy to see how this model would work for commercial and industrial applications.
It also carries implications for the equitable improvement of the grid.
“If this works out in some commercial arrangement, why should we put that in the rate base and hope the business model works?” said Feasel, Schneider’s smart grid leader. “Because eventually, if it doesn't work, poor grandma in public housing is going to pay more, and that's not fair.”
In the pay-for-service model, he continued, the customer pays the agreed upon amount, the developer makes sure everything works, and if something breaks the suppliers have to replace it. Ratepayers are not on the hook for any of that.
Serve one customer and many
Elsewhere, though, utilities are exploring ways to build out microgrids with partial ratepayer support.
The general model derives from the dual nature of most microgrids as both a local, independent service and a grid-tied resource. In the partially rate-based model, the utility builds a customer-sited microgrid in a place where resilience is in high demand, but uses some of the assets for broader grid services when the customer doesn’t need backup.
The customer pays for its expected use of the facility, and the utility asks its public utilities commission for permission to recover costs from the grid-serving portion of the project.
Arizona Public Service has undertaken two such projects, using diesel generators rather than solar and storage. One is at Marine Corps Air Station Yuma, the other is at Aligned Data Center, going up in Phoenix. Data centers place a premium on uninterrupted power, so APS used the microgrid as a way to attract the center to its territory.
The host customers contribute based on their expected usage of the facility, said Scott Bordenkircher, APS director of transmission and distribution technology innovation and integration. When those facilities don't need backup power, though,the microgrids can serve the broader grid with peaking energy, frequency response, voltage regulation and spinning reserve. The utility is asking for cost recovery based on the stacked values the microgrids provide for ratepayers at large.
“I've got to do these certain things anyways, now I can do them with this, what's that look like from a cost comparison standpoint?” Bordenkircher said of the decision making process. “In fact, it's far cheaper using the microgrid generation than it is through other means.”
This is one of those calculations that makes sense in principle but could get tricky when it comes to divvying up the costs in real world situations. The microgrids are up and running, but APS is still waiting to hear back from regulators on the rate case that would reimburse the grid side of the projects.
A rendering depicts the futuristic aerotropolis development Peña Station Next, where Panasonic's new office will host a microgrid (City and County of Denver).
Over in Colorado, utility Xcel Energy has partnered with the city Denver and Panasonic to create a $10.3 million microgrid in Peña Station Next, a “smart city” development under construction near the airport.
Panasonic is anchoring the new development with an office to house the operations center for its nationwide network of utility-scale solar installations. Such a facility needs to run at all times. Meanwhile, Xcel was looking to improve grid reliability in the area and brace for a feeder heading toward 30 percent solar penetration. The company had received permission from the public utility commission to recover costs for testing energy storage on the grid, making the deal even more attractive.
Peter Bronski with Panasonic's marketing and content strategy team described the project as “a microgrid with fuzzy boundaries,” because the components don’t all live within the fence the way one might expect.
The microgrid includes 1.6 megawatts of solar owned and operated by Xcel but installed on a carport at Denver International Airport. Panasonic is footing the bill for 259 kilowatts of solar PV on its own roof and will host the 1 megawatt/ 2 megawatt-hour lithium-ion battery from Younicos. That battery is located in front of the meter, but on Panasonic's side of the islanding switch.
The partnership, then, leverages private land and funding to deploy equipment with the potential to help the broader grid. Panasonic will enjoy an estimated four hours of backup for its critical facilities. The utility gets solar integration, grid peak demand reduction, energy price arbitrage and frequency regulation.
The partners will watch and learn how the batteries work for two years, testing the ability of the battery to stack different jobs in real life. After that study period, Xcel and Panasonic will finalize a more formal agreement for use and payment for the rest of the battery’s life.
It's also possible to do this sort of project without a rate case. CPS used company funds to construct a demonstration microgrid at Fort Sam Houston in Joint Base San Antonio, to operate with analytics funded by a National Renewable Energy Laboratory grant. The company chose a library as the project site, to gain experience with real usage patterns without putting any critical operations on the line just yet.
That approach doesn't constitute a business model per se, but it's a way to develop early microgrids and figure out sustainable business models down the road.
Several microgrids have arrived, then, but microgrids as a class are still in a state of arriving. This class of project has entered the grid and started carrying real loads, but the examples discussed here still serve an exploratory purpose, gathering data for a better understanding of how they can work in the future.
Given the diversity of markets and customers, it's reasonable to expect no one model will win out. The sources interviewed for this story agreed that the microgrid market is still in its infancy, and a thriving market for this asset won't arrive for perhaps five or 10 years.
In some cases, regulatory policies will impose constraints, particularly where utilities are barred from owning distribution and generation, and where storage is classified as the latter. The precedent set by the Arizona Corporation Commission in the APS rate case will also be a sign of the viability of the hybrid model for leveraging private and public capital.
It is noteworthy that utilities are not shying away from this type of grid experiment. Utility representatives repeatedly dismissed the idea that a proliferation of microgrids could catalyze a death spiral like the one once expected from distributed solar generation. They're not worried about grid defection from customers who can run their own miniature grids.
Part of that is logistical. For secure, nonstop power, localized generators would run into air permit issues, and solar-plus-storage still requires a big upfront expenditure, said APS's Bordenkircher.
“I don't see [a microgrid] as becoming my 24/7 power source of choice,” he said. “It's going to provide a secondary service that I need, not a primary service.”
It's cost effective to run a microgrid on grid power most of the time, and island it in an emergency; to achieve full independence drives the costs up fast.
Additionally, microgrids offer utilities a new source of revenue, especially welcome as traditional utility business models face changes in the years ahead.
The skills required to design and operate a microgrid — electrical engineering, grid balancing and the like — are already areas where utilities specialize, noted David Chiesa, who watches over S&C Electric Company's microgrid market segment as senior director of business development.
Meanwhile, distributed solar and energy storage are spreading on their own steam and complicating the business of grid operation, as seen on the feeder where Xcel is building the microgrid. Microgrids offer a way to compartmentalize these distributed resources and mitigate their effects on the grid at large, before things get out of hand.
“Why do you think they’re experimenting with it now?” Chiesa said.
To really enable a utility-driven microgrid market to flourish, microgrid design will need to become more standardized, he added.
“If you look at microgrids today, they're all perfect little snowflakes,” Chiesa said. “That's anathema to fleetwide management… Utilities are going to want their microgrids to look more like a fleet.”
Several pathways forward have become clear. Microgrid developers need to log more hours of run time and shave the costs of designing and building. But they've got an essential crowd pleaser on their hands: A product that combines the safety and reliability of traditional grid architecture with the control and flexibility of distributed energy.