Utilities are at the forefront of a data analytics challenge many industries may soon face: making sense of data from millions of connected devices, such as connected smart meters, thermostats and sensors.
Machines that weren’t traditionally considered computers are increasingly being equipped with controllers and networked, a trend some call the Internet of things. Cars, for example, are instrumented to send diagnostic data back to engineers, consumers can remotely control home alarms and thermostats, and medical devices report patient data to doctors off-site.
Utilities that invested in smart meters and other connected devices are in the throes of integrating these new sources of information and, in some cases, using analytics to get more value from that data.
For the most part, the utility industry justified the investment in smart meters to make their operations more efficient by automating billing or perhaps locating buildings that were knocked offline after a big storm.
Now some utilities are analyzing data from meters, thermostats, and sensors for different business functions, such as customer service, product development, or marketing, says Rick Nicholson, an analyst at IDC Energy Insights.
Data-Rich Utilities Search for Business Value
The biggest challenge in this transition isn’t so much the volume or even the speed of the data being produced, he says. Instead, it’s finding ways to get business value from these new sources of information. “Utilities now have lots of data but they are still trying to figure out what to do with it—how to create value,” Nicholson says.
Some utilities have developed their own systems or work with professional services companies, but a number of niche software companies have developed applications to help utilities better monetize their investments in two-way meters, thermostats, and other smart devices, such as sensors that check the status of transmission lines.
One of those companies is Opower, which specializes in providing customer-engagement tools to utilities. Opower announced on Nov. 19 the latest version its software platform, called Opower 4, which can now handle data from smart meters and wireless thermostats.
Opower is the company behind the paper reports that compare one household’s energy use to neighbors and national averages, which it does by processing data from 50 million households. It’s proved to be an effective way for utilities to motivate customers to improve home energy efficiency and participate in programs, such as rebates for programmable thermostats and efficient appliances.
Now that Opower collects and analyzes data from smart meters, it can provide more personalized and fine-grained reports to utility customers, the company says. The meter can indicate, for example, that the air conditioning load during the past month was particularly high and recommend in a weekly email that the consumer change the temperature settings or perhaps have it serviced.
Connected thermostats generate useful information as well. By using analytics, a utility can see whether a person has actually programmed a programmable thermostat (most people don’t) and tailor recommendations to individuals.
Managing all that data and launching these types of applications requires specialized skills, says Opower chief marketing officer Roderick Morris. The volume and speed of data is significant, but processing data from meters and thermostats, which may have unusual data formats, is a particular technical challenge.
“You have to actually access all the datasets, put it together and run analytics on top of it,” Morris says. “Then you need the ability to understand what datasets are important for messaging for consumers.”
The utility industry has stringent requirements on security and privacy, Morris added, and the meter and thermostat data need to meld with their existing back-office systems, such as billing. In the case of smart meters, limited network bandwidth can constrain how much data can be collected, too.
Building for scalability and multiple data types is vital. Opower’s Hadoop-based platform will process 96 billion meter reads in a year and 100,000 events from thermostats in a month. To generate useful information for consumers from thermostat data, its software needs to do 12 billion calculations a month.
One Building Means Petabytes of Sensor Data
The amount of data from sensors and other connected devices is projected to keep growing. Southern California Edison, for example, will collect 12 terabytes of data from its four million meters, which provide snapshots of power usage every 15 minutes, says Amit Narayan, the CEO of startup AutoGrid and a former Stanford University smart grid researcher. But once a building is filled with sensors to monitor room occupancy and lighting, for example, the volume of data will multiply into the tens of petabytes a year.
“As you move to finer granularity and get data from sensors on the grid and sensors behind the meter to measure things like load and voltage, then the data starts exploding and you start getting into volumes that are really unprecedented even in the Internet world,” he says.
Utilities are starting to lean on big data analytics to keep the lights on, too. Equipment, such as substations, transformers and transmission lines, are in some cases being monitored in real time to spot outages quickly. Texas utility Oncor, for example, recently deployed a power outage detection system built around data from billions of measurement points including smart meters, power line sensors, and transmission equipment.
For grid operators, closely monitoring consumption helps them avoid power outages during peak times, such as a hot summer day, and maintain a balance between power supply and demand.
NV Energy, for example, is making smart thermostats available to consumers who want to participate in demand-response programs (consumers typically get a rebate of some kind). Using software from analytics startup EcoFactor, NV Energy will be able to remotely change thermostat settings to lower peak load during the hottest days of the year. EcoFactor also does analytics year round to make minor thermostat changes to cut consumers’ energy bills and improve the utility’s energy efficiency.
Utilities are still in the early days of using analytics to get more from their investments in meters and other connected points on the grid. As with other industries, finding the right technical talent is difficult for utilities, one reason why many are going to outside tech providers, says IDC Energy Insight’s Nicholson. “A few utilities are doing the work themselves but a lack of technical and subject matter expertise relative to analytics in the industry is a constraining factor,” he says.
Martin LaMonica is a technology journalist in the Boston area. Follow him on Twitter @mlamonica.