Energy utilities have seen a sea change in the amount of data available to them in recent years. Take Pacific Gas & Electric, for example. In 2013, PG&E became the largest U.S. utility to install smart meters in its customers’ homes and businesses. The company went from one reading per month per customer to the ability to collect as many as 12 readings per minute. Multiply that by the utility’s 5.4 million electricity customers and it is easy to see why the industry is swimming in data.
As a result, utilities are beginning to change how they leverage smart meters. It’s a trend toward data-driven decision-making that aims to balance the demand for electricity with its supply. The trick is to find the right information in all that data.
That’s where a team of Stanford scholars comes in. With support from the TomKat Center for Sustainable Energy, Stanford scientists June Flora, Chin-Woo Tan, Sam Borgeson, and Professor Ram Rajagopal are deploying an open-source platform to analyze and display smart meter data.
The platform is known by its research project acronym, VISDOM, for Visualization and Insight System for Demand Operations and Management. Taking smart meter data as its inputs, VISDOM provides calculations ranging from fairly simple, such as displaying the minimum and maximum power demand for each customer, to complex, such as weather normalization, baselining models, and algorithms to cluster load shapes into similar groups. This approach is gaining unprecedented insights into consumer usage patterns and behavior.
“There are so many potentially beneficial uses of smart meter data,” says Sam Borgeson, the technical lead and head developer for VISDOM. “By interpreting the data we can now present a full spectrum of opportunities that utilities can pursue as they design their efficiency programs.”
Air conditioners, for instance, are major energy hogs that operate when the electrical grid is already maxed out. To reduce their burden on the grid, a utility might decide to offer a rebate for energy-efficient air conditioners or for a tune-up of existing systems. The trick is getting the right message to the right customers. How do they know who has an air conditioner, especially an inefficient one?
The answers are in the data.
VISDOM has an algorithm that filters for customers whose energy consumption reflects excessive air conditioning use at times when the grid is stressed. Utilities can then target these individuals with personalized messages indicating why they are being approached, instantly creating cost savings for the program by lowering marketing costs and improving uptake compared to when the rebates are generically marketed to everyone.
In PG&E’s service territory, a targeted approach focusing on top air-conditioning customers can make related efficiency programs up to four times more effective than those that are untargeted. And while targeted programs are not new, VISDOM can segment out groups more granularly, making them twice as effective as programs that already do target A/C customers by climate.
Broadly, Borgeson explains that the platform allows utilities to sift through the smart meter data in three important ways:
Analyzing smart meter data is a nascent field, and one fraught with some controversy. A small but vocal segment of PG&E’s customers rallied against smart meters due to their concerns about radiation from the wireless technology. Others were unhappy with higher electricity bills due to the utility’s ability to track energy usage at peak times. Confidentiality of consumer data is also a widely shared concern.
The California Public Utilities Commission, which oversees the regulation of this data for the state, has some of the strictest rules in the country for releasing consumer information. In fact, VISDOM began through a first-of-its-kind research collaboration between a utility (PG&E) and researchers (Stanford) in 2009. The data-sharing agreement was years in the making, and enough to lure Borgeson from Berkeley across the San Francisco Bay to Stanford.
The collaboration for residential customers was first made possible as part of the Stanford Energy Behavior Initiative, funded mainly by the U.S. Dept. of Energy's Advanced Research Projects Agency-Energy. A subsequent project, funded by Stanford's Precourt Energy Efficiency Center, applied VISDOM to commercial customers. An Innovation Transfer grant from the TomKat Center has allowed the effort to find the best path from research to actual applications. Today the researchers run a company called Convergence Data Analytics, LLC, which in spring 2016 signed a contract with PG&E to model customer data and consult on energy-saving projects. It also works closely with the Vermont Energy Investment Corporation (VEIC), a Northeast nonprofit that is a thought leader in the field of energy efficiency.
The team has been careful to keep the software available as an open platform. In doing so, they hope to accelerate research into beneficial applications of smart meter data and to lower barriers for interested utilities. As the platform’s reach grows, so does the potential to begin making sense of the energy landscape in the United States and abroad.
“We see this very much as applied research. Tools like VISDOM will have a quantifiable impact on efficiency and demand response programs,” Borgeson says. “The grant from TomKat has helped us begin to deploy our tools and see them adopted in the real world.”
VISDOM, known commercially as Demand Insight, is available for download here.