ARPA-E, the government's energy research agency, backs a lot of complex projects. But its mission is very simple: find and support groundbreaking technologies that could have a transformative impact on the energy sector at scale.
Traditionally, the agency has been focused on capital- or materials-intensive solutions like batterystorage fuel cells, biofuels, electric motors and power electronics. But in recent years, ARPA-E has ramped up its support for companies and university research teams that are developing software analytics for the transmission and distribution grids.
"A couple of years ago, we asked what happens if that doesn't come true," said ARPA-E Program Director Timothy Heidel, referring to new types of battery storage for balancing the grid. "What happens if it takes longer?"
Addressing that question meant looking at grid solutions that were easier to scale than batteries or fuel cells. Grid optimization was the natural choice.
In 2011, the agency formed the Green Electricity Network Integration (GENI) program to support grid modernization technologies like power flow controls, new types of circuit breakers, outage response systems, distribution automation and low-cost power electronics. But the team at ARPA-E eventually realized that those projects only represented half of the potential.
"That naturally leaves you with some on the hardware side. But it's not only the hardware," said Heidel, speaking at GTM's Soft Grid conference last week.
Software analytics became a priority for the $39 million GENI program as well. Since 2012, the agency has supported five software projects designed to model and manage the grid in real time. One prominent startup in the analytics space, AutoGrid, was awarded $3.4 million to build a platform for managing demand response and dynamic pricing programs in real time. The software, which was developed in partnership with the Lawrence Berkeley National Laboratory and Columbia University, is now being used by Austin Energy, Palo Alto’s municipal utility, the Sacramento Municipal Utility District and Oklahoma Gas & Electric.
While many of the teams developing new hardware are hitting their milestones, Heidel noted that progress has been faster within teams developing software.
That's because a new software developer can go to a utility and use a synthetic data set to prove its value. Once the firm has proved that the algorithms work, the software can often run apart from other critical functions. "You can rarely do that with hardware," said Heidel.
Consequently, a few of the teams building software analytics tools have moved well beyond their expected milestones and have chosen to focus on immediate issues for customers, not necessarily the big multi-year challenges that ARPA-E grant recipients usually tackle.
"There’s no question that they’ve been able to gain traction a lot faster. Small vendors are really starting to gain a lot more traction now," said Heidel.
Of course, ARPA-E's mission isn't to support technologies that can scale quickly. The agency was created to support the kind of technologies that even the most ambitious investors wouldn't touch. However, Heidel and his team understood that the more difficult, hardware-based power control solutions would only be as strong as the supporting analytical capabilities.
"Having a single hardware device is never going to be sufficient. You need the optimization systems to wrap around that," Heidel said.
To hear more about how ARPA-E is shaping its investments in data analytics, watch the panel below from last week's Soft Grid conference in Menlo Park, Calif. The conversation also includes Silver Spring Networks and startups Grid4C and TempoIQ.