Making Sense of 46M Smart Meters and 1B Data Points Every Day

Soft Grid 2013 brought together experts in the field for an update on how they’re managing that big data challenge.

Greentech Media has been covering the ins and outs of how utilities and smart grid technology providers are making use of the nation’s massive rollout of smart meters for years now. As the Edison Foundation reported last month, that rollout now stands at 46 million smart meters and counting, delivering a collective 1 billion data points per day -- a massive set of information and potential value to those that can make sense of it.

This week’s Soft Grid 2013 conference in San Francisco brought together some experts in the field to update us on how they’re managing that big data challenge. On a Tuesday panel entitled “Last-Mile Analytics: Insights Into AMI, DG, EV and Beyond,” GTM Research analyst Zach Pollock and a host of smart meter experts laid out the current state of play. Here’s a snapshot of what went down.

- Josh Gerber, smart grid manager for San Diego Gas & Electric, started out by reminding the audience that utilities have a core use case for their smart meters that can’t be messed with. “Primarily, the AMI network is there to measure the interval data for customer bills,” he said -- and that means that any suggested use of smart meters that might interfere with the smooth functioning of that system isn’t going to get a good utility reception. 

The value of that AMI data can certainly be extracted in ways it’s not today, he added. At the same time, SDG&E, which has a complex and overarching grid communications and automation plan in place, isn’t looking to its AMI network to serve as the primary conduit for its distribution automation (DA) work, he said: “We try to find the right network solution for the right application.”

Instead, the utility is looking at ways that this customer information can be analyzed and turned into useful insight -- a process that’s going to come not in one big wave, but in wave after wave of innovation. “This is the beginning of a very long evolution that may or may not end in our lifetimes,” he said.

Scott Young, senior director of software platforms for Silver Spring Networks (SSNI), pointed out that while AMI networks are indeed first and foremost meant to manage customer bills, that leaves a “network that is pretty vastly underutilized” for other purposes.

“You do have a tremendous amount of computing power out in the network,” he said, which can be tapped for functions from grid reliability diagnosis to customer-facing insights. But to figure out what information that AMI network can collect from customers and deliver back to them, “You have to reach out to those customers and give them something that interests them,” he said.

Silver Spring is working on customer-facing demand response projects with Oklahoma Gas & Electric and other utility customers, but it’s also looking at going deeper, he said. One intriguing possibility might be to analyze household power usage data to help customers learn where they’re wasting energy, how different appliances are affecting their usage, and the like -- in other words, to turn utilities into “trusted energy advisors” to their customers.

Richard Mora, president and CEO of Landis+Gyr Americas, said that his company, which was bought by Toshiba for $2.3 billion in 2011, is getting ready for significant investments in data analytics. At the same time, it’s been delivering information from existing deployments, such as its 3-million-meter network with Texas utility Oncor, he said.

That project, in conjunction with IBM, uses AMI data in conjunction with meter data management and outage restoration systems to remotely diagnose service call problems that previously would have required utility truck-rolls to fix. “What was key there was the ability to integrate the three systems,” he said.

Beyond that, L+G is working on the customer insight value to be derived from smart meter data with a Texas municipal utility, which is looking at ways to predict individual customers’ likelihood to want to sign up for load curtailment programs (demand response), or perhaps install solar panels on their property.

In other words, “It’s pairing third-party information” of the kind you might get from the retail and customer service bid data world “with the usage details they get from their meters,” he said. (Another Toshiba-Landis+Gyr project in New Mexico is seeking to integrate solar PV, battery, distribution grid and customer behavior information to balance out those resources.)

Franco Castaldini, software solutions product marketing manager for GE Digital Energy, laid out another use case with a “large Midwest utility” that correlated AMI data with weather data to find meters that were operating outside of their normal temperature range to diagnose potential failures.

After a period of building trust with utility field crews, the utility’s data analysis team was able to show good results from this data-driven diagnosis, he said. The next step they’re contemplating is “a big data platform to automate this process.”

GE’s Grid IQ Insight data analytics business has been working on concept-to-execution projects like these with lots of utilities, but that process is going to take quite some time, he noted. The important thing is that utilities and their big data partners take the time to determine which projects are the most valuable and cost-effective, while also fitting smoothly into the way utilities do business today. 

“To bring this to scale, it’s going to take an iterative approach,” he said. “Most of that is going to come from a change management perspective, to get all these people and departments to work together.”

Paul Moon, vice president of corporate development and strategy at SK Telecom’s GridMaven, laid out how his company’s smart grid network management system (NMS) it helping to provide insight into the right mix of communications technologies for smart grid deployments like the one it’s doing with National Grid in Worcester, Mass.

That project includes wireless mesh, cellular, WiMAX and utility-owned point-to-point microwave radios, all performing different functions side by side. GridMaven’s ability to track and analyze all those systems from a single view can help utilities determine the “right technology cocktail” on both operational and economic levels, he said.

GridMaven is also building on that network-of-networks management role to include insight into other grid functions, he said. One example is its project with Duke Energy’s McAlpine smart grid test bed in Charlotte, N.C., where it’s helping to integrate data from grid sensors, smart meters and solar and energy storage devices to “look at what’s going on beyond communications.”

Watch a discussion at Soft Grid 2013 about analytics and consumer engagement:



Video streaming by Ustream