Last month, the Department of Energy’s ARPA-E program put out a request for information (RFI) for new technologies to help solve a devilishly complex problem: managing today’s advanced battery energy storage systems.
Today’s automotive or grid-storage batteries contain hundreds or thousands of electrochemical cells, all interacting in ways that are difficult, if not impossible to predict. To work around all that uncertainty, today’s batteries tend to be “built and operated well below their theoretical energy and power capacities,” the RFI states. Even when they are, “Many energy storage systems suffer from uncertain or inadequate lifetimes,” which increases costs and risks of deploying them.
What’s needed is technology to monitor and manage the cell-to-cell charging, discharging, degradation and quality issues within the battery system as a whole, whether it’s new sensor technology, data modeling or battery system controls. While ARPA-E notes how complicated the challenge is, it states that finding the right combination “would be a game changer -- significantly accelerating the adoption and value of energy storage.”
This isn’t news in the battery industry world. On the automotive side, we’ve seen General Motors and Nissan pour millions of dollars into battery management technology, both to squeeze maximum electrical efficiency out of the systems and to control temperatures to avoid thermal runaway. Tesla Motors is sharing its battery management system (BMS) with Toyota and Daimler.
On the grid energy storage scale, we have hundreds of megawatts of lithium-ion batteries that are running some sort of hardware-software combination to charge and tap their warehouses full of cells, packs and racks of batteries. Companies like A123, Xtreme Power, Saft, AltairNano and Japanese giants like Mitsubishi, Panasonic and Hitachi are bringing their own battery management technologies to support the grid-scale storage units they've deployed. A123 has supplied more than 100 megawatts of grid storage using its own BMS, and Coda Automotive and BYD Motors have launched grid storage lines that take what they've learned in automotive battery management to apply to grid scale.
But the ARPA-E call for technology makes it clear that there’s still a lot of work to be done to get the industry into new realms of efficiency and reliability. So do reports of “bricked” Tesla Roadster batteries, Chevy Volt battery fires, exploding grid-scale wind power storage units in Japan, and other such bad news.
I recently spoke to two startups, Sendyne Corp. and GreenSmith Energy Management Systems, that said they’re working on solving ARPA-E’s battery management problem. Each has its own techniques, and each is tackling the market, such as it is, in a different way. Each also has its work cut out for it, if it’s to meet ARPA-E’s definition of a “game changer” in the battery management space.
Sendyne Seeks Active Balancing for Better Performance
Sendyne, a New York-based semiconductor developer that also works on portable fuel cell power management, unveiled its battery management plans in February when it announced a partnership with the University of Toronto. The company is providing “patented technology as well as access to intellectual property, development tools and scientific expertise” for the partnership, according to its announcement.
Ellen Gooch, marketing director for Sendyne, wouldn’t disclose details on the company’s funding to date, or which battery cell makers it may be partnering with. Nor did she get into too many details on Sendyne’s technology during a February interview, though she did say it’s based on semiconductors that can accurately measure voltage, temperature and other states within battery cells.
“We are addressing what we believe to be fundamental issues with respect to battery pack management. Among these are accurate measurement and accurate modeling,” she said. Today’s battery systems suffer from a lack of visibility into cell status, as well as the complexity of lining hundreds or thousands of strings of cells together, she added -- “No matter how well you match them, no two cells are exactly alike.”
That means that, over time, certain cells become stronger or weaker than others. Because a battery pack’s overall performance is limited by its weakest cells, battery systems today use a technique called passive balancing, which actually bleeds off power, to try to manage that variation, Gooch said.
Sendyne, on the other hand, is working on active balancing, where power isn’t bled off, she said. Instead, the startup manages relative state of charge at the cell level, sharing charge between cells to balance out weak-strong cell differences. That process can be as much as 15 percent more energy-efficient than the passive process, she said.
Active balancing isn’t a new concept, of course, and there’s active research underway. Tesla buys battery cells from Panasonic and stacks them in the thousands for its Roadster and Model S batteries. Since 2010 it has used power management chips from Infineon in its BMS, which could support active balancing.
Beyond better-balanced cell charging, Sendyne’s system can warn when certain weak cells are about to drop below a critical state of discharge, typically about 15 percent or so, that essentially ruins the cell, Gooch noted. That’s important in preserving the overall life of the battery pack, which can degrade beyond usefulness if too many cells end up failing.
GreenSmith: From Cell Management to Smart Grid Integration
Bethesda, Md.-based GreenSmith is also working on active balancing technology, targeting the control of cells within individual battery packs, CEO John Jung told me in a February interview. The startup was founded in 2008 and has raised $6 million in angel and early-stage financing.
So far, GreenSmith has done demonstrations that show its active balancing system can improve round-trip efficiency of battery systems from the high-70 to low-80 percentage range all the way up to 88 percent, Jung said. Its firmware can also check battery nameplate performance against real-world performance and spot weakening cells and capacity degradation issues, he noted.
In GreenSmith’s case, however, its cell-to-cell management is tied up in a larger Distributed Energy Storage System offering, which includes the industrial servers and underlying software that allows batteries to interact with smart grid, solar panels that deliver power, or plug-in vehicle charging stations that want to tap the power, he said.
GreenSmith is managing a bunch of smaller-scale (100-kilowatt or so) energy storage units that add up to about 4 megawatt-hours of storage in projects with San Diego Gas & Electric, Hawaii Electric Co., Duke Energy and solar manufacturer Amonix, among others, Jung said. The main goal is to make grid-scale storage systems “as easy as distribution transformers” to deploy onto the grid, he said.
The Challenge Ahead: A Comprehensive Set of Solutions
Importantly, neither Sendyne nor GreenSmith are planning on building batteries themselves. Sendyne is aiming at providing components and software for battery management system makers to use for consumer, automotive and grid storage applications, Gooch said. GreenSmith is talking with OEMs interested in partnering with the startup to license its technology or to bring it to market with battery system channel partners, Jung said.
They're not the only startups targeting battery management. RockPort Capital Partners and Terawatt Ventures have backed Newark, Calif.-based Qnovo, which has promised a next-generation electronic management system for lithium-ion batteries, without providing any details on its technology.
Anyone tackling the challenge will have their work cut out for them, on a number of fronts. Let’s take accurate measurements. ARPA-E’s request for information acknowledges that technology exists today to solve the problem, but that it’s very expensive: “The idea of embedding a third reference electrode into a commercial battery has captivated the imagination of battery developers,” but that solution “remains an elusive goal.”
Likewise, getting down to cell-level monitoring in complex battery systems containing up to thousands of cells “is not practical. Series and parallel cell configurations couple the states of groups of cells, and the cost of highly parallelized wiring or sensing is prohibitive,” ARPA-E wrote.
Even when sensors at the individual cell are available, today’s simple voltage, current, and temperature measurements only go so far in giving a true picture of what’s happening inside it. To wit: “Voltage provides a composite measure of changes in potentials, but does not provide critical information on the individual state of either electrode. Temperature measurements in today’s systems only probe the surface temperature of a cell, not sensing localized temperatures that internally drive degradation and failure.”
All in all, ARPA-E sees a need for a comprehensive set of technology solutions, “combining data from novel sensors with advanced models, system designs, and control paradigms.” Any combination of solutions, of course, will need to “provide system level benefits (e.g., increased battery utilization, lifetime, safety, and/or applicability), which far exceed the cost of implementation.”
It’s important to note that there’s no funding attached to ARPA-E’s request, though law firm Wilson Sonsini Goodrich & Rosati noted that it could tap a portion of $100 million in anticipated future funding opportunities at the agency. No doubt the agency will be poring through the submissions to its RFI to see if anyone has come up with persuasive proposals for solving all these problems -- or, more likely, just one of them. In the alchemical world of battery science, any insight into the inner workings of operating cells may be worth investing in.