With two grid-scale batteries newly in place, Arizona Public Service is learning the elements of energy storage operation that you can't read in a book.
The utility is using the pair of 2 megawatt / 2 megawatt-hour lithium-ion battery systems from AES to test how the technology performs in the desert climate of the greater Phoenix area, where summer temperatures routinely crest 100 degrees Fahrenheit.
The Advancion systems have been inserted into a distribution grid that has seen rapid uptake in distributed solar generation as a result of the Solar Partner Program, which places APS-owned PV modules and smart inverters on customer homes. Advanced battery vendors tout the technology's ability to ease renewable integration and correct power quality on the lines. APS wants to find out just how well this works in practice.
“No utility, to our knowledge, has ever done two identical batteries specifically sited in two different locations along a feeder,” said Scott Bordenkircher, director of technology innovation and integration at APS.
That distinction enables a controlled experiment whereby Bordenkircher and his colleagues can test the locational value of storage for voltage control and power quality. Typically, equipment at the substation regulates voltage for the whole feeder; voltage runs high at the head of the feeder to ensure it maintains an appropriate level at the end. Grid-scale batteries open up alternative solutions.
Local, reliable data
APS put the batteries on a distribution feeder in a wide open valley northwest of Phoenix, where the Festival Ranch housing development has created a load pocket.
The West Valley has the highest penetration of rooftop solar in the Phoenix area, helped by the frequency of new builds with west-facing roofs that can capture the late afternoon sun. The Festival Ranch feeder already has 68 homes participating in the Solar Partner Program, in addition to other solar customers.
One battery went in at the substation, and the other sits in a walled enclosure halfway down the feeder, on a swath of open land adjacent to the housing tract. Now the utility can test whether one works better than the other, and how they compare to the traditional approach to voltage control.
“You've got a lever point in the middle of the feeder that has a much better chance of acting on both ends as opposed to having to force it from just one end,” Bordenkircher said.
Similarly, it could be better able to smooth out fluctuations from the rooftop solar and smart inverters by being located in the midst of them. If a cloud passes overhead and knocks out 200 kilowatts of midday rooftop capacity, the utility has to call up additional generation elsewhere. The battery can do it from across the street.
“It's a shock absorber — it tends to level out the entire feeder by being able to quickly absorb or discharge depending on the situation,” said project manager John Pinho.
The trial run will also test the batteries for peak shaving, although the 2 megawatt-hour energy capacity somewhat limits the usefulness for absorbing four or five hour peaks.
APS has access to the land for the mid-feeder battery for two years. If the demonstration proves successful, it may seek an extension, but it will also be possible to pull out the 60,000 pounds of batteries, lift the 100,000 pound container out and move it to some other useful location.
Future battery installations will be informed by the data these systems gather on basic operations in the Arizona heat. Bordenkircher wants to make sure they charge and discharge as expected in a real-world field setting. The testing will also measure parasitic load, which is how much energy the battery consumes in the course of operating. That number impacts the ultimate cost and value of ownership.
That kind of data exists for batteries elsewhere, but different feeders have different effects on battery operations, Bordenkircher said. Data on a battery's performance in southern California don't necessarily translate to how it will behave on an APS feeder in the outskirts of Phoenix, so a local calibration is needed.
The climate control keeps the interior of the battery at a balmy 75 degrees, even in the middle of a desert valley. (Image credit: APS)
The batteries aren't doing anything that can't be done by conventional distribution equipment, like regulators, capacitor banks and integrated volt-var control.
For correcting voltage, energy storage currently costs more by 40 to 60 percent, Bordenkircher said. But that's only if you count that one application, and a storage unit can do many.
“We're not 100 percent sure on what the math is going to come out to be yet,” he said. But, factoring in some peak shaving, load following for distributed solar and the declining costs of battery cells, there's reason to believe storage will be a clearer sell very soon.
“There's no question that energy storage is definitely the next frontier, we've just got to get it right,” he said. “We've got to look at it from how do we do it efficiently and in an optimal fashion rather than just throw a bunch of batteries everywhere and kind of cross your fingers.”
It's hard to blame the utility for a methodical approach. Lithium-ion installations are still expensive and new for utilities, and the costs of getting it wrong are high. An early foray into grid-scale energy storage went sour when a 1.5-megawatt Electrovaya storage system APS owned and operated in Flagstaff caught fire in 2012.
The new batteries, which are the utility's first grid-scale acquisition since Flagstaff, come equipped with fire suppression tanks the size of a person, and multiple layers of sensors to provide advance warning and response if anything starts operating erratically. Communication links beam 14,000 data points per second to a 24/7 monitoring station AES set up in Phoenix.
APS is moving ahead with its next battery procurement, having released an RFP for a 2 megawatt / 8 megawatt-hour battery for the town of Punkin Center.
This growing, rural town about 90 minutes northeast of downtown Phoenix needed 20 miles of new line to increase reliability. APS determined that energy storage penciled out as a more cost-effective and versatile investment than replacing the traditional T&D infrastructure. That new math is already coming into force, even without state mandates to drive it.
Communication is key
Field operations can reveal challenges that are hard to envision at the planning table. For the Festival Ranch project, a key takeaway was that reliable communication links are harder to maintain than they might appear.
“We have to be able to know what it's doing and when it's doing it, and if we want to make it change to do something else, we have to be able to communicate,” Pinho said. “If anything fails in the linkage that's required for all this communication, then the system can't necessarily do what you want it to do.”
That sometimes puts the installation's effectiveness at the mercy of the cell signal providers. A blip at a cell repeater could initiate radio silence for crucial seconds when the batteries might be needed to discharge or charge.
The Festival storage system has already experienced brief blips like that, Bordenkircher said. This experience has changed the way his team thinks about the communications side of storage planning. In remote locations, a storage system may even require building out additional infrastructure to create more robust lines of connection.
“If we had it next to the substation, all our substations have fiber. Not an issue,” Pinho said. “But when you're not at a substation or you're in a remote area like Punkin Center — and we're going to have more and more of those places, not less and less — that becomes bigger than probably we anticipated it would be.”
Simpler grid technology from earlier times could sit and operate passively. The evolution of batteries' intelligence simultaneously opens up new potential and heightens the sensitivity of the network to brief lapses in connectivity.