SAN ANTONIO — Energy disaggregation is a term of art for technologies that can parse out energy usage of individual loads like air conditioning, heating, lighting and appliances, simply by analyzing single sources of data like smart meter feeds or circuit-level sensors.
But this general term masks a variety of different approaches to this underlying data-analytics challenge, largely based on what granularity of data they can access, how quickly they can analyze it, and how deep a historical record they have to draw from.
These differences are important to keep in mind when looking at two major energy disaggregation announcements out this week from two of the world’s biggest smart meter vendors, Itron and Landis+Gyr.
Both are aimed at delivering much more detail on what’s happening with behind-the-meter loads. But each is going about it in a very different way — at least, at first.
Bidgely and Itron mine historical data for personalized insights
Itron announced last week that it’s partnering with Silicon Valley startup Bidgely to provide pre-integrated access to historical data from its entire fleet of nearly 40 million installed electric and gas meters.
In simple terms, it’s a pre-packaged opportunity for Itron’s existing utility partners to mine years of interval data for appliance-level consumption patterns, personal energy usage profiles, and other valuable insights.
Bidgely, which launched in 2010 and got early funding from the Energy Department’s ARPA-E program, has built a roster of more than 25 utility clients using its software to analyze 15-minute or hourly interval meter data and present the findings to their customers. That resolution of data can delivery fairly accurate predictions of how much energy major loads like HVAC or appliances, but it can’t deliver the more fine-grained insights available to technologies that can sample directly from building circuits.
But Bidgely has focused on this interval data because that’s what is available from typical smart meter networks, CEO Abhay Gupta noted in a Tuesday interview at this week’s DistribuTech conference.
“The future business cases for smart meters are being built on the question: What are we going to do with this data?” Gupta said.
Not all utilities have made the most of their smart meters. Utilities that have invested hundreds of millions of dollars in advanced metering infrastructure (AMI) are eager to squeeze more value out of those investments, Gupta said. And utilities planning new AMI deployments are eager to provide state regulators and ratepayers proof that they’ll be cost-effective.
Bidgely has taken on this kind of data-mining with its utility partners, but the new Itron partnership takes the cost and complexity of the effort out of the equation, he said. That could enable more utilities to deploy Bidgely’s customer-facing web portals and alerts services, which is integrated with home energy management partners like EnergyHub.
The technology has many utility-facing values as well, he said. One of the biggest early-stage opportunities are the “load profile surveys” that utilities send out every few years, asking a sampling of perhaps 10,000 customers to identify what they’re plugging in at home and how they’re using it.
To say the least, that process doesn’t yield very clean or granular data for forecasting changes in future load growth patterns. But for decades it was the best most utilities had available, Gupta said.
With Bidgely’s new pre-integration with Itron AMI and meter data management (MDM) systems, “we are giving you, across a million homes, a ground-up load profile of every home,” and “instead of taking months, it takes five minutes.”
Other utility vendors can also tap the same data streams, whether to help customer service reps resolve billing disputes, or to better target efficiency and demand response programs. In recent years, Bidgely has been positioning itself as a natural central data repository for disparate utility software systems.
While Bidgely has developed technology to analyze more high-fidelity data, it hasn’t yet put that to use with Itron, Gupta said. But Itron’s latest generation of technology, OpenWay Riva, comes with on-board computing power and memory to enable more fine-grained sensing of real-time data, he noted.
Landis+Gyr and Sense: real-time, high-resolution disaggregation at the edge
Real-time data is the focus of the second big energy disaggregation announcement of the week.
Sense’s energy disaggregation algorithms are based on decades of speech recognition research by founder and CEO Mike Phillips, and sample building circuits tens of thousands of times per second to capture slight changes in voltage and frequency.
That requires high-fidelity sensors — Sense’s system costs about $299 for retail home installation — but it also reveals loads of data that can’t be captured by algorithms searching through sparse interval data.
“This is why we built our own hardware, because we couldn’t get what we needed from interval meter data,” Philips said in a Wednesday interview.
But L+G’s new meters will come with the sensors Sense needs, as well as the on-board computing power necessary to analyze that raw data to yield valuable insights.
“We can see the HVAC breaking down, because we have this super-detailed view of power,” he said. “We’re starting to use it for load shifting. We’re using it for EV charging management.”
Indeed, a handful of utilities, including Alliant Energy, have bought Sense’s retail circuit monitors to install in customers’ homes, in advance of its integration into smart meters, he said.
Tim Weidenbach, product manager for Landis+Gyr, noted that Sense’s capabilities can also discern problems happening on the utility side of the meter. Much of the new Revelo platform’s promised value for would-be utility clients will come from its “high resolution sensing, based on streaming waveform data” — a capability tied directly to Sense’s technology, he said.
Landis+Gyr has been enhancing the computing power and distributed networking capacity of its smart grid networks, much as rival Itron has done. But the Revelo technology will include high-fidelity sensor technology that’s previously been used primarily for industrial and commercial meters for customers with high power quality needs, allowing them to support Sense’s data requirements in ways that could be challenging for other residential smart meters, Weidenbach said.
That flood of near-instant data will then be turned over to Sense’s pattern-recognition algorithms, to predict grid failures, pinpoint causes of outages, and other such tasks.
“If we start to see systemic issues on multiple feeders, we can tell you, you’ve got a transformer that’s about to fail,” Weidenbach said. “The same concept in the home, we’re doing at the grid level.”
High-fidelity energy disaggregation efforts stretch back more than a decade, including Intel Labs’ experiments with in-home sensoring technology, and Belkin’s integration of Zensi into its WeMo line of smart home products. Other startups, such as San Francisco-based Verdigris, Belgium-based Smappee and Vancouver, Canada-based Neurio, have promised technology that can yield accurate disaggregation of household loads as well.
But Sense’s integration with voice-activated devices like Amazon Alexa and Google, and backing from heavyweights like L+G and Schneider Electric, could mark a new chapter in the spread of energy disaggregation into mass-market electric gear.
At the Consumer Electronics Show in Las Vegas earlier this month, Schneider Electric unveiled its new Square D Connected Home suite of smart electrical panels and controls, which incorporate Sense’s technology as well.