Consider that, according to industry researchers, if smart meters were incorporated across the U.S., they would generate five times the data of the current AT&T network. That’s a lot of data to manage, but there’s a huge advantage: Accenture consultants estimate that those same smart meters could, if properly deployed, save each electricity customer $40 to $70 per year.

Big Data, the ability to combine and analyze huge amounts of varied information, presents enormous opportunities to save energy and curb emissions, as well as major management and logistical challenges that are only now being addressed. The topic was explored last March in a conference entitled “Sustainability in the Age of Big Data,” hosted by the Initiative for Global Environmental Leadership (IGEL) at Wharton. 

According to Arthur van Benthem, a professor of business economics and public policy at Wharton, “Big Data is an enormous opportunity for making environmental improvements and harnessing energy-efficiency savings. We seem to observe an ‘energy-efficiency gap’: People do not seem to adopt efficient technologies that appear financially attractive. One of the commonly cited reasons is that information about how to save energy is hard or time-consuming to collect, or that the efficient devices are hard to use. Firms that employ Big Data can help consumers overcome these challenges.”

The promise is clear. “Connected machines could eliminate up to $150 billion in waste across industries,” said Paul Rogers, GE’s chief development officer and a keynote speaker at the IGEL conference. “In aviation, a 1% fuel savings would have a value over 15 years of $30 billion. For natural gas-fired power generation it translates to $66 billion.” A 2010 McKinsey & Company study concluded that a holistic energy-efficiency program could produce energy savings worth more than $1.2 trillion, reducing end-use consumption in 2020 by 9.1 quadrillion BTUs (British thermal units) and eliminating up to 1.1 gigatons of greenhouse gas every year.

But the hurdles are just as clear. The technology that promises these savings is also adding new sources of energy consumption. According to Digital Power Group, the world’s new technology — everything from smartphones to data centers — is now using an estimated 1,500 terawatt-hours of power annually, or about 10% of all the generation in the world. Thanks in part to this growth, McKinsey reports that all of the possible energy-efficiency savings mentioned in its report represent just 23% of the projected demand in 2020. 

“Big Data is an enormous opportunity for making environmental improvements and harnessing energy-efficiency savings.” –Arthur van Benthem

Ruben Lobel, a management professor at Wharton, was the moderator for the energy panel at the Big Data conference. “The market for energy efficiency, in my view, has a couple of key problems,” he said. “There are payoffs that accumulate over the long run, but they come with some set-up cost.” 

Lobel added that those costs can be offset with financial innovation. “Maybe 10 years ago,” he said, “the solar technology cost was the key problem. Then the next bottleneck became the financing, which was partially solved with the leasing business model of companies like SolarCity and SunRun. There is a lesson here that business model innovation is perhaps as important as technology innovation.”

Finding Efficiencies

The rapid growth of Big Data is occurring just as saving energy and increasing fuel efficiency (with attendant climate benefits) has become a top priority for government and industry. Fortunately, harnessing information has already yielded big energy gains, and considerably more are promised. 

“We’re becoming instrumented, interconnected and intelligent,” said Wayne S. Balta, vice president for corporate environmental affairs and product safety at IBM. “We have the ability to measure, sense and see the exact condition of everything,” he noted at the Wharton conference. For energy, that means using smart meters to track consumption, analyzing traffic data to reduce congestion and harnessing the tools of distributed generation to incorporate renewable energy into the grid.           

Balta said that IBM has pushed for energy efficiency ever since the Arab oil embargo of 1973, which means that many of the easy targets have already been hit. But the company’s goal is reducing energy consumption by 3.5% annually, and it has been exceeding that with 6% gains. “Each year, our annual bill for energy would be at least $30 or $40 million more without conservation,” Balta noted. 

IBM’s own data centers represent a ripe target for energy-efficiency gains. According to Balta, IBM researchers have produced color-coded mobile measurement technology that can track “hot spots” in data centers and target them with air conditioning or isolate them with chimneys — making it unnecessary to cool the whole facility. 

Microsoft, too, is using Big Data for energy gains. A team led by facilities director Darrell Smith spent three years organizing 30,000 existing sensors (many from different eras) at the company’s Redmond, Wash., headquarters into a single energy-efficiency system. The network yields billions of data points each week on such cost areas as air-conditioning, heaters, lights and fans, the company said. In one case, analyzing the data turned up a garage exhaust fan that had been left running for a year, costing the company $66,000. Overall, the system avoids what would have been a $60 million capital investment in energy-efficiency technology. 

Rob Bernard, Microsoft’s chief environmental strategist, said the “canonical example” of energy waste is a company building that feels comfortable but is actually hugely inefficient because both the air conditioning and heating system are operating — and canceling each other out. “Humans may not be able to realize it’s happening, but integrated systems can talk to each other and uncover the problem.”

Microsoft has also invited in more than a dozen vendors to use its platform to develop efficiency solutions for the corporate campus in Redmond. It’s a pilot program. “Our goal is to manage our worldwide facilities from a relatively small and centralized infrastructure of data optimization,” Bernard said. 

Bernard added that Microsoft is in a period of global experimentation to learn what works and what is cost effective. “For example, we can use weather information to super-cool a building the night before a hot day. We can use information to move energy around.” 

In addition to reducing energy consumption, Big Data analysis can improve the efficiency of the energy industry itself. According to SAS, a privately held software company with a 12-acre solar farm, the drilling analytics it provided for one national oil and natural gas company increased foot-per-day penetration by 12% and cut nonproductive time by 18%.

A Boon for Utilities

According to a 2013 study from Tata Consultancy Services, utilities and energy/resources industries have the biggest expectations of achieving returns from their Big Data investments. It’s an embryonic field, so the low-hanging fruit has yet to be picked. A Ventana Research study found that only 12% of all manufacturers have reached full maturity in their use of analytics.  

Houston-based CenterPoint Energy is an electric and natural gas utility with more than 2.2 million customers that is maximizing its Big Data potential. Until recently, the company read 88,000 meters every day with physical inspections (which cost $75 each). Now, with a mature network of smart meters, it can gather 221 million readings a day and leave the trucks in the garage. 

For CenterPoint, the advantages go beyond savings to public safety — data analysis can quickly reveal that a power line is down and presenting a hazard. Access to detailed weather data allows utilities to have better performance during storms. 

Utilities are adding renewable sources, and Big Data is helpful there, since managing solar and wind systems, and collecting, aggregating and analyzing huge amounts of information (on solar gain and tracking for instance) yields big efficiency gains. 

The individual turbines in large wind farms are now communicating with each other to improve efficiency. According to Rogers, “If a turbine loses wind speed or wind direction, it simply asks what its neighbor is doing and replicates the action, improving availability and power output.” Every second, GE’s wind farms analyze 150,000 data points to optimize the delivery of 400 megawatts to the grid. 

Utilities almost universally recognize the promise of Big Data, but that’s not the same as being ready to maximize its effectiveness. According to a 2013 Oracle white paper, only 17% of utilities “are completely prepared for the data influx” (up from 9% in 2012). And only 20% give themselves a grade of “A+” for getting information, big or not, to the people who need it (up from 8% in 2012). 

Oracle Utilities’ second annual Big Data study, also 2013, found (after interviews with 151 senior electric utility executives) that less than 50% are using new data to provide alerts or make direct customer-service improvements. Further, only half were analyzing the impact of distributed generation, 39% were reducing the cost of generation operations and 26% were assessing what electric vehicles mean to their business. 

“We’re becoming instrumented, interconnected and intelligent. We have the ability to measure, sense and see the exact condition of everything.” –Wayne S. Balta

“Utilities are using more data today, [but] many are not using that data as efficiently as possible,” the survey said. It posited that “historic industry silos need to be pulled down, allowing a more open, holistic and collaborative environment in which data, in particular, is owned and used by the entire enterprise, rather than by specific utility departments.” 

Success takes going beyond business as usual. Lobel pointed out that utilities “are not typically encouraged to promote energy efficiency unless mandated to do so by regulators.” Revenue, he said, is tied to the amount of electricity they produce, and only so-called “peak shaving” — aimed at cutting consumption during high-demand times — is a universal priority. Utilities, he warned, “can improve the meters, collect data and offer pricing schemes to promote efficiency, but they won’t do so if their incentives are not aligned in that direction.”

Changing Energy Behavior

Ory Zik is founder and CEO of Energy Points, which tracks and analyzes the route that electric power takes from source to consumption. He said at the IGEL conference that, in buildings, much of the low-hanging fruit — such as changing traditional light bulbs to LEDs and using more efficient HVAC systems — has already been picked. Changing consumer behavior is one of the key hurdles moving forward. 

That’s where companies like Opower, which allow utility customers to compare their electric usage with that of their neighbors, comes in. Opower has accumulated energy-use information from 100 million homes, and is adding 80 to 100 million smart data sets annually.

Wayne Lin, a senior director in product management at Opower, said at the conference that if an average consumer with a bill of $100 a month is told he or she can reduce that by $2 or $3 with efficiency improvements, “They’ll say it isn’t worth their time.” But those same people will pay attention, he noted, if the next door neighbors are spending far less than they are. The potential energy savings from what Opower calls “behavioral efficiency” in the U.S. amounts to 18,677 gigawatt-hours annually, or 238 kilowatt-hours per household. 

“Today,” Lin said, “30 states have energy-efficiency standards, and 20 don’t. We’re just scratching the surface of what we can do with technology. But if a utility doesn’t need to add a power plant or upgrade the grid, they also don’t have to go to regulators and ask for a rate increase.” 

In addition to changing the behavior of honest customers, utilities are using Big Data to help deter the misbehavior of some less-than-honest consumers. A Deloitte/CIO Journal study estimated that electricity theft — sometimes accomplished by such low-tech methods as turning meters upside down or gumming them up with glue so they’ll run slower — costs utilities in the U.S. $6 billion annually. The Pepco utility reports that stealing electricity is the third-largest form of theft in the country, following shoplifting and stealing copper. 

Utilities are developing analytics-based fraud detection systems that, according to the Deloitte study, can collect several million dollars over five years, more than paying for the cost of developing them. Still, many utilities lack the necessary information infrastructure to allow them to deal with large amounts of data on generation, consumption and billing — sometimes for financial reasons, but just as often because information is held captive in the aforementioned silos. 

But progress is being made. The good news, according to IBM’s Balta, is that “organizations around the world are turning “too much data” into better decisions. Speaking at the IGEL conference, he said that “the walls between companies and customers are breaking down, and we’re developing a two-way street of information. Computers are thinking and acting more like people, and people are being enriched with new levels of computing power.”

Tracking the Transportation Sector

Transportation is America’s number-two consumer cost (after housing.) Though cars are only driven 4% of the time, the average annual bill for U.S. drivers is $8,000. According to the Rocky Mountain Institute (RMI), a more productive use of our existing transportation services (tapping into creative software applications) would yield the same travel results, but with 46% to 84% less driving — and huge energy savings.

To reduce fuel consumption and increase efficiency in the transportation sector, Balta said, customers can now often check bus locations in real time, and book travel appointments through their smartphones. Other technology making travel more efficient, he said, includes accurate and frequently updated traffic information, and dynamic pricing for both road use (tolls) and parking. Consumers are being encouraged to travel and park in off-peak periods. 

The nonprofit Green Parking Council is working on those solutions, and also certifying parking garages for taking energy-reducing steps like installing motion detectors and LED lighting. It’s also encouraging creative solutions such as smartphone-enabled reserved parking with staggered hours to avoid traffic jams during special events or the morning rush (see The Green Sports Movement, a Special Report by IGEL and Knowledge at Wharton). 

According to Environmental Building News, “The idea of a green parking garage might strike some as an oxymoron, but … since LEED specifically excludes stand-alone parking garages, the niche is open.” 

Kelly Schwager is chief marketing officer at Streetline, whose embedded sensors are now sending data from parking spaces in 45 cities around the world. Armed with an app, Streetline customers don’t have to circle for hours looking for parking — they can go straight to an available space. “We’ve recorded 250 million parking arrivals and departures on our system,” Schwager said. “We’re working with Cisco on a video solution that can capture and interpret what is an actual parking event, and not just a driver pulling over to pick up or drop off a passenger.”

“Today, 30 states have energy-efficiency standards, and 20 don’t. We’re just scratching the surface of what we can do with technology.” –Wayne Lin

Streetline, drawing on work from Donald Shoup, a professor of urban planning at UCLA, estimates that 30% of urban traffic is the result of frustrated drivers hunting for parking. The average hunt is six to 14 minutes. “Parking is not always considered when cities are looking for ways to reduce congestion,” Schwager said. “But it can have a meaningful impact if, based on our data, we can tell drivers whether to turn right or left to find parking.” 

Streetline’s data explodes the myth that most city parking spaces are always occupied. “We’ve yet to go into a single city where that was true, even in downtown Los Angeles,” Schwager said. “What often happens is that you go straight and the available parking was to the right.” 

In one RMI scenario, a smartphone alerts a morning commuter of a traffic delay and suggests an alternative route that saves 15 minutes. On the drive, the car alerts you of a pair of highly rated rideshare patrons along the route, with their payments (made on their mobile equipment) offsetting travel costs. Your phone also helps out at the garage by producing and redeeming a discount coupon, and also noting the location of your parking spot so it can be easily accessed for the ride home. 

According to Greg Rucks, a senior consultant at RMI, we’re “making strides” toward a transportation system that is “instrumented, interconnected and intelligent. It helps if apps take an “open data” approach that breaks through the silos. “Imagine,” he added, “instead of a few municipal staff thinking about how to improve their isolated systems, millions of software developers look to gain lucrative market share in the information space — from established powerhouses such as Google and IBM to individual programmers.”

For example, 4% of the 400,000 monthly trips on San Francisco’s Bay Area Rapid Transit (BART) are planned using Embark, a free-to-customers smartphone app started by a trio of college students. 

Zik said that UPS is working with electronic route maps that reduce both left turns and truck idle times. At Dubai Airport, he said, planes that are still 40 miles away from the airport are coordinating with catering crews about the food and water aboard, the number of passengers and other variables, so that servicing can be sped up, and energy consumption reduced. “It avoids planes just sitting 20 feet from the gate,” he said. “If you multiply that over and over again with many flights, you see big impacts on sustainability.” 

Similarly, GE’s Paul Rogers said at the Wharton conference that eliminating system inefficiencies in freight rail operations through strategic use of data could save $27 billion over 15 years, and capital expenditures as part of oil and gas exploration and development could be reduced $90 billion. SAP’s Parker added that spillage is a “big problem” in oil and gas operations, causing delays, shutdowns, fines and accidents. 

“It’s all about running optimally,” he said. “Two thirds of a typical time in a railcar trip is now spent doing nothing but waiting for clear track, but that downtime can be vastly improved. If there’s a problem with an oil-pumping operation, someone would have had to drive by and see it happening in person. It typically has taken three weeks to discover an issue, fix it and get back online, but now we have real-time analytics. We can, for instance, turn off pump 72, because it’s pumping dirt.” 

A single plane flight yields a half to a full terabyte of data, which begs the question of why the lost Malaysia Airlines Flight 370 couldn’t be found using that information. Unfortunately, GE’s Rogers said at the IGEL conference that standard practice is for a technician to download most of that information once the flight lands, which isn’t much use in crashes.

Rogers said that even the Aircraft Communications Addressing and Reporting System (ACARS) system in airplanes, which sends out automated text messages, doesn’t always include location. Sometimes that has to be roughly calculated using the time it takes for the signal to travel to nearby satellites. “We’re working with the airlines now to expand ACARS,” Rogers said. 

Beyond simply finding lost flights, Rogers said, instant access to GPS information for planes could enable maximizing the flight plan on the fly to use the least amount of fuel. Since for many people airplane flights are the biggest part of their carbon footprint, adding efficiency wherever possible is very important. 

Electric vehicles haven’t yet made much a dent in the gasoline engine hegemony, with approximately 100,000 sold in the U.S. in 2013, but there’s hope for streamlining the process. Lobel is intrigued by Tesla Motors’ recent declaration that it would open its patents, especially on its “Supercharger” technology, to other manufacturers.

“Having each manufacturer develop its own charging network is completely redundant and inefficient,” Lobel said. “If EV companies can pool resources and share charging networks, we can seriously speed up the adoption of EVs. It is not clear how costs should be distributed here in this case, but I will be interested to see how this will develop.” 

At Syracuse University’s School of Information Studies, students are analyzing millions of time-stamped electricity records collected through Pecan Street, the environmentally themed development in Austin, Texas. Each smart meter monitors power usage on 50 Pecan Street households, which incorporate rooftop solar panels and electric vehicle charging. 

The Pecan Street Research Institute (PSRI) has already analyzed data to reach an interesting conclusion from the development’s concentration of electric vehicles (50 in a single half-square mile neighborhood). EV owners aren’t charging as much on high-demand summer afternoons as behavioral models had predicted. PSRI said the findings “could significantly increase utility industry estimates on the number of EVs the electric grid could handle without triggering disruptions or requiring major system upgrades.”

The challenge is not generating Big Data, or finding creative, energy-saving uses for it, but zeroing in on the opportunity while keeping up with and managing ever-expanding information streams.