For the smart grid, the weather is going to be murder.

The proliferation of solar panels will effectively transform commercial districts and neighborhoods into small, localized power plants. While that will allow utilities to cut back on coal, the unpredictable, varying nature of solar power will force grid operators to dispatch or throttle power rapidly. Solar-balancing smart grid systems now confined to pilot projects will need to become common features pretty soon.

"That future is coming, and it's coming a lot faster than I think many people are aware of," said Chris Baker, CIO of utility San Diego Gas & Electric. (Execs from SDG&E and other utilities will speak on this and other topics at Greentech Media's The Networked Grid conference in San Francisco on Nov. 4.)

SDG&E has about 6,600 customers with solar rooftops, he said. While that's growing by about 60 customers a month, it still only represents about 50 megawatts of generation, or about 1 percent of the utility's 5,000-megawatt total load. The utility can't monitor or control it, but there isn't enough of it to matter that much, he said — yet (see Having Solar Energy System Trouble? Don't Call Your Utilities). 

But what happens when 20 percent or more of the homes in a neighborhood go solar and a cloud passes overhead? That changes a neighborhood of solar power producers to utility power customers in a matter of minutes – and grids built to deliver power one way at constant voltages and frequencies have trouble accommodating that two-way, intermittent flow.

Too much solar power, and local grid voltage could rise, causing potential problems for motors, lights and other equipment. Too little, and voltage can sag. That may only flicker light bulbs at home, but it can lead to million-dollar work stoppages for customers like semiconductor manufacturers and server farms that need clean power at a near-to-constant voltage and frequency.

Energy Secretary Steven Chu has pointed to the challenges of integrating intermittent wind and solar power once it gets to about 20 percent of the capacity of the Bonneville Power Administration, a big, sprawling, power delivery entity rich with reliable hydropower.

What's the maximum amount that the neighborhood distribution grids of today can take – or, that is, what is the maximum amount that a utility is willing to support?

"We don't know," said Matt Wakefield, a senior project manager with the Electric Power Research Institute's smart grid demonstration program. "Utilities are asking, where do we draw the line... They're not sure what is the magic point on penetration and how do you manage that issue with [customers] who really want to get the benefits of solar."

EPRI is tackling the problem in a project with Albuquerque, N.M.-based PNM Resources, aimed at an overarching grid automation system to balance out a neighborhood's solar power with energy storage and demand response systems, Wakefield said. The Department of Energy gave out $11.8 million in July to five "solar energy grid integration system" projects with similar goals.

"Reality is going to come into play here," Wakefield said. "We haven't seen enough of this to see the full effects."

Solving that problem is the focus of smart grid projects for utilities across the country (see stories here, here and here). Some of them could get a boost from the $3.4 billion in DOE smart grid grants announced last week, though more are awaiting word on a $615 million pool for more experimental projects (see Green Light posts here, here and here).


Take California, which plans to get one-third of its power from renewable sources by 2020 – the most aggressive renewable portfolio standard (RPS) of any state in the country.

The Renewable Energy Transmission Initiative, a group of California regulators and utilities, has said that looking to distributed solar power to meet a large portion of those needs could reduce the number of new transmission lines that will need to be built to carry power from far-off desert solar power plants and mountain wind farms to cities (see California 'Green' Transmission Lines Could Cost $15.7B).

But that distributed scenario – which mainly looked at utility-owned rooftop solar – also adds to the costs of that power, the group warned in August. One big uncertainty over its feasibility, the report said, is its "impacts on grid reliability."

Some California utilities have already pushed back against efforts in the state legislature to expand the 2.5 percent cap on how much customer-owned solar power they're obliged to credit to customers, though they pointed to extra costs imposed on non-solar customers, rather than grid stability, as their main objection (see Cal Net Metering Bill Stalls).

But Pacific Gas & Electric has asked the California Public Utilities Commission for permission to expand its cap to 3.5 percent, since it's expected to hit the lower cap by the end of the year.

Right now, PG&E is seeing "some localized issues" with grid instability in neighborhoods where rooftop solar penetration has grown to around 5 percent, said Hal LaFlash, the utility's director of emerging clean technologies. It's responded with existing distribution grid systems that are suitable to the task at hand, he said.

But both the Federal Energy Regulatory Commission and the Institute of Electrical and Electronics Engineers, a standards-setting body, have put limits on distributed power sources like solar panels making up more than 15 percent of a distribution substation's load, he noted. 

"We know there's a lot more coming," LaFlash said. PG&E is working with various initiatives, including DOE's Solar Vision Study, to keep abreast of the issue. But a project the utility proposed to pilot technologies to integrate solar panels with smart grid systems in San Jose didn't receive DOE funding last week, leaving its future up in the air (see Green Light post).

PG&E and other California utilities might have to face a much larger percentage of their power coming from such sources in the near future, according to Edward Cazalet, co-founder of grid energy storage startup MegaWatt Storage Farms.

"The reality is, to achieve 33 percent, most of it is going to be done close to the load," he said. "We're just not going to be able to build the transmission fast enough" to bring far-off solar and wind farm power to cities, he said.

Cazalet's main solution, as befits his company's name, is lots of energy storage – 4 gigawatts in California to meet its 2020 renewable power goals, to be specific. He'd like state regulators to consider making it mandatory for utilities (see Green Light post).

A lot of that would be big storage, he said. That could include pumped hydro and compressed air energy storage, which EPRI's Robert Schainker and others say will remain the most economical form of energy storage, as well as large-scale batteries like the one Southern California Edison wants A123 Systems to build to manage wind turbines in the state's eastern Tehachapi mountains (see SoCal Edison Wants A123's Biggest Grid Battery Ever).

Other batteries could fit into neighborhoods, where they could balance out rooftop solar sags and surges.

[pagebreak:Will Solar Crash the Smart Grid?]


At San Diego Gas & Electric, "Storage is absolutely part of the solution," Baker said, all the way down to batteries at customer premises.

To the north, utility Southern California Edison is asking DOE for $35 million to fund a $70 million smart grid project in Irvine, Calif. that will include household batteries linked to solar panels.

General Electric, a project partner, has said it will bring "net-zero energy homes" to market by 2015, with a concept diagram that includes basement batteries to store rooftop solar power (see GE Unveils Net Zero Energy Home Strategy).

In fact, much of today's rooftop PV comes with small batteries to cushion household wiring from the panels' ups and downs in power delivery. That's according to Michael Ozog, COO of smart grid products at Integral Analytics, which is working on integrating such household systems for utilities including Duke Energy (see Integral Analytics: Orchestrating Duke's 'Virtual Power Plant').

Utilities such as the Sacramento Municipal Utility District are looking at ways to control existing PV systems, he added. The problem is that most systems out there today weren't made to be controlled by, say, a smart meter calling for a battery to discharge power to meet a utility's peak demand – something Integral Analytics has found out in field trials, he said.

More smart grid-ready systems involving new batteries and inverter technologies would be useful, he said (see Rooftop Solar, Batteries Included). But they could add too much to the already daunting costs of a home solar system for most customers, he said.

"I think you'll have to have some storage someplace else to manage that volatility," he said – and if utility customers won't buy it, the utility may have to.

Some utilities are looking at cargo container-sized flow batteries or sodium-sulfur batteries at substations (see Grid Energy Storage: Big Market, Tough to Tackle). Other utilities are considering a similar, small-scale distributed approach known as community energy storage, Cazalet noted (see Utility to Try Backyard Storage).

"That's where you put a battery on a box next to the transformer on the pole, and the benefit for them is, they can put that into the distribution rate base," he explained. "It could help some of the distribution problems caused by PV, and changing loads from charging plug-in vehicles." IDC analyst Nadav Enbar and now Nissan have floated the idea of using aged lithium-ion battery packs from cars for utility storage.

"But if you put that same battery in the home, you're going to have to finance that yourself," Cazelet said. "And if you're providing benefits to the grid, are you going to be compensated for that?"


Maybe utilities could balance growing amounts of rooftop solar power without a lot of expensive batteries.

Inverters, which convert solar panel DC output into grid-ready AC, could help by delivering short bursts of energy, EPRI's Wakefield said.

That could give them a role in niche utility power needs such as frequency regulation, or the task of keeping the grid humming at 60 hertz, which now takes up about 1.5 percent of the nation's generating capacity.

Inverter companies such as Satcon and Petra Solar have discussed using their products for these kind of grid-balancing tasks. Inverters could also help with voltage regulation and VAR, or reactive power, problems, Wakefield said. Those can sap distribution grids of some 5 percent of their efficiency (see Notes From a National Smart Grid Experiment).

"The inverters have the technology to do it," Wakefield said. Unfortunately, "There are some functions we just cannot do, because the standards development isn't in place."

Specifically, the IEEE 1547 standards for linking distributed power to the grid don't support inverters for voltage regulation and VAR support, though proposals are in works to include them, he said. Also, many inverters use unique communications that would be easier to integrate into smart grid systems if they were standardized, he said.

There's another form of balancing that can be done by subtracting, rather than adding, power to the grid – demand response, or the ability for utilities to turn down customer's power use when they're facing peak loads on the system.

Demand response tends to be a centralized affair at present, with utility dispatchers paying agreeable customers and giving them ample advance warning to turn down their big power loads – sometimes via radio signals, text messages, emails or  phone calls – when needed.

But automated systems running over smart meter communications networks could open up new ways to tap the demand response potential in the home – imagine turning down the air conditioner when the solar panels detect a cloud passing overhead.

Tendril Networks, which makes gear and software to manage household energy use, and Fat Spaniel Technologies, which monitors solar panel output, announced a partnership to try just that. Some of Tendril's 30 or so utility pilot project partners have been asking for such technology, CEO Adrian Tuck said last week (see Tendril Wants to Link to Solar Panels).

Of course, utilities that can predict when solar panels are about to fade could fire up peaker plants or dip into demand response capacity to match it. That's where microclimate forecasting comes in, EPRI's Wakefield said.

Every utility forecasts the weather as a part of their day-to-day business. It's hard to miss a heat wave's effect on air conditioning loads. Pinpointing local weather conditions in real time – and then adjusting the local grid in response – is a more complicated matter.

Given the complications, projects to integrate solar panels, storage and smart grid systems are more easily contemplated in smaller units. A host of startups are looking to provide the intelligence to do that, as well as find ways to make the more controllable power pay off by selling into various markets.

Well-funded smart grid software startup GridPoint is helping Xcel Energy manage rooftop solar for Boulder, Colo. residents in its SmartGridCity project. Duke Energy has enlisted Integral Analytics to manage solar, storage and demand response for about 500 customers on its McAlpine substation in its headquarters city of Charlotte, N.C.

Fort Collins, Colo.'s FortZED (Zero Energy District) project has enlisted hometown company Spirae to help (see Green Light post), and Viridity Energy has partnered with Siemens to use $1.2 million of utility PECO Energy's $200 million DOE smart grid grant to set up a microgrid project at Pittsburgh's Drexel University.

San Diego Gas & Electric is using the town of Borrego Springs as such a test bed, and has asked for DOE grants to connect it to a second such "microgrid" at the University of California at San Diego, Baker said (see Microgrids: $2.1B Market By 2015).

Interact with smart grid industry visionaries from North American utilities, innovative hardware and software vendors and leading industry consortiums at The Networked Grid on November 4 in San Francisco.