How Energy Disaggregation Can Inspire Efficiency

Bidgely and PG&E see 7.7 percent efficiency gains from an 800-customer pilot.

Pulling apart whole-home energy data and turning it into appliance-by-appliance breakdowns can help even the most inspired household energy-savers get more efficient.

That’s according to a report released Wednesday by energy disaggregation startup Bidgely and utility Pacific Gas & Electric, tracking the results of an 850-customer pilot project conducted last year. The results showed that participants were able to reduce their energy use by an average 7.7 percent, a pretty impressive result considering that it was driven entirely by behavior change, not with any smart thermostats or other automated energy-saving devices.

Each customer received a gateway device that read second-by-second energy data from their smart meters’ home area network (HAN), and turned that into individual energy usage estimates for air conditioning, pool pumps, and big appliances like refrigerators and washing machines. From August to December, these customers had access to Bidgely’s HomeBeat platform, which delivers energy updates and alerts via web and mobile devices.

Notably, most of PG&E’s test customers were already signed up for the utility’s time-of-use energy rate program, which charges more for energy used during peak afternoon hours, Bidgely CEO Abhay Gupta said in an interview.

“By definition, these people have already optimized some of their energy consumption in the home,” he said. “We were able to demonstrate 7.7 percent energy savings on this time-of-use customer base. Imagine what we could do with a non-optimized home.”

Bidgely is one of a number of energy disaggregation technology providers looking at different ways to bring this core capability to broader market via different channels. Some, such as PlotWatt, have chosen to work directly with commercial businesses such as fast food restaurants. Others, such as Smappee, are selling directly to homeowners. But Bidgely has chosen the utility as its key customer and channel to market, which means that it’s interested in comparing its approach to behavior-driven efficiency against other utility options, Gupta said.

On that front, Bidgely’s test results indicate deeper energy savings than the average 2 percent efficiency gains delivered by Opower, the U.S. leader in utility-to-customer engagement. Of course, whether Bidgely’s results with PG&E can be achieved across the millions of customers that Opower now engages remains an open question. It’s also important to note that Opower has gotten some utility customers to reduce energy consumption by up to 5 percent during peak power events with customer Baltimore Gas & Electric.

Gupta offered another interesting metric to compare Bidgely’s capabilities with the utility status quo. Most California energy efficiency programs run by utilities cost about 10 cents to 20 cents per kilowatt-hour of energy reduced, he said, through energy-efficient light bulb giveaways, appliance and home retrofit rebates, and other traditional methods, he said.

Bidgely’s solution, by contrast, costs about 3 cents to 5 cents per kilowatt-hour reduction, he said. That’s a measure of the cost of the HAN gateway device, plus the cost of paying Bidgely for its software and analytics, and assuming that each homeowner will reduce their typical 6,000 kilowatt-hours per year consumption by an average of 6 percent, he said.

This type of back-of-the-envelope calculation doesn’t necessarily lead straight to utilities directing efficiency program dollars to energy disaggregation, of course. Utilities have to follow specific rules in auditing and verifying how much energy is saved, and at what cost -- and behavioral-based programs are much harder to subject to these rules than are programs that can calculate the difference between an old light bulb and a new one.

It also relies on the real-time nature of the smart meter HAN-to-home gateway setup that PG&E’s test customers were using. Giving people real-time data has been shown to be more effective than giving people day-old information about their energy usage patterns -- but it comes at the cost of installing the equipment to make it possible.

Bidgely has launched its smart meter HAN-connected gateway platform with utilities including TXU Energy, London Hydro and Hawaii’s Energy Excelerator, and it could see customers of California’s big three utilities buy and set up their own gateways under the state’s push to turn on smart meter HAN connectivity to third-party devices. But it and other utility-partnered energy disaggregation providers are also using the smart meter reads being sent back to utility back-office systems as sources of data.

The challenge there is that disaggregation becomes more accurate with more frequently collected data, which means that 15-minute or hourly smart meter reads might not deliver the same accuracy on individual appliance energy usage, or may not be able to differentiate the number of in-home energy loads that people find compelling enough to pay attention to. (Just how accurate different technologies in this field can be is a tricky and debatable issue, with questions of transparency, size of test samples, and other factors coming into play.)

But Bidgely is also working with partners like Silver Spring Networks, PG&E's smart meter networking provider, to get faster, more granular data without the use of an in-home device, Gupta noted. Silver Spring’s Silverlink platform is promising to “pull in high-resolution data, pull that data back to the utility head-end, and make it available to apps like Bidgely,” he noted. “To get HAN-level granular data in the future, we may not have to go through the HAN.”