Kim Bauters, Ph.D., is an expert in artificial intelligence. If you want to know the best gelato recipe to use the ingredients you have on hand, he has developed a smartphone app that will help you. If you want to know the best recipe to make soap, he’s put AI to work for you with another app called Soap Genie.
If, however, you want to save electricity and reduce your carbon footprint, Bauters explains, that’s much easier “because there isn’t an artificial intelligence component.” So, he created an app that helps homeowners calculate the best times of day to use their appliances.
Bauters isn’t the only one helping people go green and save money. Karolis Petruskevicius, the founder of Homely Energy, used his parents as guinea pigs to see if the energy-saving technology he was developing would work. And Mick Wall taught himself several programming languages as he developed a web page showing the lowest-cost times to use household appliances.
Across the U.K., lone hackers, a term that has come to mean high-tech tinkerer, are creating ways to harness energy price and usage data that has only recently become available. Their creations tell people how to reduce CO2 emissions and save money by reducing their electricity use. Remarkably, by reducing electricity demand at specific times, they are also helping balance the electrical grid.
Two innovations made this possible. First, an electric and gas company called Octopus Energy created a variable-rate pricing plan called Octopus Agile that uses prices to reward customers for using electricity when supply is abundant and increases prices when demand is high. Second, Octopus made real-time price data available to anyone with enthusiasm and a little programming skill. Octopus encouraged everyone who would listen to create new tools for their customers.
The results are just beginning to appear, but these new innovations are attracting the attention of energy experts and managers in the United Kingdom. Thanks to the democratization of information, individual hackers like Bauters, Petruskevicius, and Wall can create, all by themselves, what used to require a team of experts with a big bank account. Octopus is also adding customers, thanks in part to these innovators who make trying this unique approach to electricity pricing less daunting and more rewarding.
Helping homeowners respond to fluctuations in price—encouraging them to use renewable energy when it is available—will become even more important as the National Grid Electricity System Operator moves toward the goal of operating Great Britain’s grid with zero-carbon by 2025. Meeting that goal while delivering power when people need it will be very difficult without the help of homeowners. That difficulty will become more pronounced as the U.K. adds increased demand for electricity from a growing number of electric vehicles.
In the midst of these large and challenging trends, we are increasingly likely to find individual hackers nibbling away at the problem with tools that can be as simple as sharing information to gadgets driven by artificial intelligence and complex algorithms. Without them, the transition to a carbon-free grid would be more expensive and difficult.
A new approach to electricity pricing
Nobody knows that better than Phil Steele of Octopus Energy, whose official title is “Future Technologies Evangelist.” His job is to proselytize about the need to create the technology and tools customers will need in the changing electricity landscape.
“It is all about trying to get consumers on to renewable energy,” he explains. The best way to do that is to help customers use electricity during the times of day when it is generated by renewables. Currently, that is a challenge.
On a typical day, the amount of electricity that people want increases in the morning as they wake up and head to work. It declines slightly through the middle of the day and then peaks during the late afternoon and evening as people return home, turn on the TV, and cook their dinner. This pattern, while fairly predictable, can double energy demand—or cut it in half—in just a few hours. That large swing requires grid managers to find more supply by turning on and off “dispatchable” sources of electricity, often natural gas or coal. The peaking plants that generate this electricity are not only less environmentally friendly, but they are also the most expensive. Shifting demand away from peak hours would not only cut CO2 emissions, but it would also reduce the need for the most expensive sources of electricity.
Few people, however, realize how expensive it is to turn on their tumble dryer or dishwasher during those peak hours because they have been protected from big fluctuations in the prices electric companies must pay to meet that demand. When customers did not have the capability of following real-time price fluctuations, a stable rate structure made sense. Most electricity rates protected customers, charging based on simple price schemes that often had only two price tiers—one for the daytime, and one for overnight. Most electric companies in the U.K. still use a similar system.
Recognizing the new opportunities created by information technology, Octopus created their Agile rate which varies every half-hour during the day. “We launched Agile Octopus to encourage households to shift their consumption from the daily peak energy demand period of 4 p.m. to 7 p.m. when we’re paying very high wholesale prices,” said Steele. If the electric company can buy less energy when wholesale prices are high, they can pass the savings on to the customer. And the savings can be significant.
For example, at 3:30 p.m., customers could be paying less than three pence for a kilowatt-hour (kWH) of electricity. A half-hour later that price might jump to 17p per kWh. When 7:30 p.m. rolls around, those prices plunge down to eight or nine pence. Price changes during the rest of the day are less dramatic, but they can go from 1p up to 4p in just 30 minutes.
Perhaps most remarkably, prices can even become negative, when Octopus actually pays customers to use more electricity. This typically happens in the middle of the night, when demand is low and there is an excess of wind power. Last December, when Storm Atiyah blew in, the amount of wind energy on the grid, enough to power about 15 million homes in the U.K., doubled in 24 hours.
That electricity had to go somewhere. Rather than turning other generating stations off, which is expensive, Octopus decided to pay people to use the short-term surplus. Between 1:30 a.m. and 6 a.m. on Dec. 8, 2019, Octopus paid customers about 1 pence per kWh. As Octopus staff wrote on their blog, “overnight storms triggered the energy equivalent of an ‘everything must go’ sale.” If you owned an electric car, you hit the jackpot. Imagine what would happen if filling stations paid to fill your tank with petrol.
Grid managers noticed what Octopus was doing. The director of operations for the GB grid tweeted “a big thank you” for helping create demand to balance the excess supply.
Paying people to use electricity, however, doesn’t make a difference if nobody knows about it. To help customers, Octopus publishes its rates online, as well as making them available to programmers using an Application Programming Interface, or API. Put simply, the job of an API is to answer questions asked by another app. Programmers send a query to the API— a more technical version of “what will tomorrow’s electricity prices be in London” —and the API returns the data.
Programmers can ask for the prices customers will pay during the next 24 hours. They can check the energy usage for individual meters. Octopus encouraged programmers to find new ways to use all this data. In 2018, they hosted a “Hack Day,” attracting more than 100 computer programmers and tinkerers. “We wanted to enable other tech innovators in the energy sector access to this incredible data,” Octopus staff explained, “so we …cracked open our Agile Octopus API for public development.”
Individuals have taken up that challenge, and the results are beginning to appear.
Tools to empower energy customers
The most basic way to help consumers adjust to the Agile pricing is to provide timely and useable information. When Mick Wall started playing with the Octopus API, he only wanted the information for himself. His home already had solar panels, so he appreciated the value of using electricity when it was available. As an Octopus customer, he wanted to know more about pricing but the detail he wanted was not available. “You could download small amounts of Agile tariff data for a single U.K. region from the Octopus website,” he said, “but not my region.” That’s when he started his “journey of discovery.”
Teaching himself several programming interfaces, he created a program that downloaded the data into a spreadsheet. Since he already managed the web page for his running club, he created energy-stats.uk and shared what he found. Soon, people began asking him to generate graphs and over his Christmas break, he added the ability for visitors to download the data. The simplicity of the site, providing electricity prices for regions of the U.K., is what makes it useful. When I asked him what innovations were next, Wall said he liked it the way it was. “I fear adding more stuff will take away from what the site is good at, just delivering the Octopus pricing data each day,” he said.
He’s probably right. He notes that the basic information on his site changed how he uses electricity and says many people have been in touch to thank him, saying it has helped them save money too. “The Octopus Agile tariff is a game-changer,” he says. “I defy anyone not to change their habits when they know they can save lots of money.”
That is also what inspired Kim Bauters to put his knowledge of computer science to work. He knew the price data was available, but, “People in my family found it was hard to track, so I decided to create an app.” Initially, he created an app called Octopus Watch for the Apple watch but has now added versions for the iPhone and Android. Unlike his other apps that use AI to create soap and gelato recipes, the programming for Octopus Watch was straightforward – take the data provided by the API and make it digestible. “All of these incentives don’t necessarily translate into a user doing something,” he said. The amount of data could be overwhelming.
“It is like the Google problem of having too much data,” he said, referencing the trend toward “big data,” where individuals and businesses are inundated by the information. Octopus Watch takes all that data and translates it into simple guidelines about when to turn on energy-intensive appliances like clothes dryers and dishwashers. Before the app, “we didn’t use timers on any appliances,” said Bauters, “but now we routinely use the app to find the best slots.” It is paying off.
The combination of the Octopus Agile prices and his app is yielding a “fairly consistent 35 percent savings,” he told me. For owners of electric cars, the savings “tend to be much more toward 60 to 70 percent.”
To take the next step in energy savings, however, people will probably need a little help. You may not need artificial intelligence to find the best time to dry your clothes, but it is handy when trying to keep your home comfortable amidst daily price fluctuations. That is the problem Karolis Petruskevicius and his colleague Ignas Bolsakovas wanted to solve.
A Ph.D. candidate in power networks, Karolis has been focusing on modeling electricity prices in the U.K. with a growing number of residential heat pumps and an increase in power production from intermittent renewable sources such as solar and wind. Working with his advisor, he developed a model and he wanted to test it. His parents had a heat pump and he decided to make them his test subjects.
Ground-source heat pumps are expensive to install but can be extremely efficient. Taking heat from the pipes in the ground, heat pumps return about four units of heat for every one unit of electricity. “You are getting three units of energy for free from renewable sources,” Karolis explained. “It is going to be key to decarbonizing.” The problem is to decide when to use electricity to run the pump.
The algorithm he developed turned on the heat pump before prices jumped during peak demand, relying on the home’s insulation to keep temperatures at a comfortable level until prices fell again. “When price is low you slightly overheat the home. What it would do is overheat the house slightly before 4 p.m. and then let temperatures drop until 7 p.m.” With the tolerance of his parents, he found the system worked.
Monitoring the heat pump is more than most people can do on their own. Partnering with Ignas, a computer scientist, he launched Homely Energy, and created a smart thermostat for heat pumps that uses Octopus Agile price data. Their app also works with heat pumps that are already internet-connected, turning them on and off based on the same price algorithm he used at his parents’ home. “We think that we can save up to 30% compared to fixed tariffs,” he says.
If his projections are right, the market for Homely’s thermostat and app could grow dramatically. Petruskevicius cites estimates from the National Grid that half of the population of the U.K. could have a heat pump by the year 2050. He believes that as people see the economic and environmental benefits of heat pumps, more people will give them a try.
Another person who is convinced is Jan Rosenow, the European Program Director at the Regulatory Assistance Project, an independent energy think tank. “The main reason I chose Agile is that we installed a heat pump,” he told me. “The interest was to see if we could get benefit out of load shifting.”
Where does one of Europe’s leading experts on energy demand turn for information about how to manage his personal energy use? “I rely on third-party apps,” he answered. Prior to switching to Octopus Agile, Rosenow was paying about 14 p per kWh on average. Now, thanks to the price information, he says he pays about 7 or 8 p per kWh.
Rosenow is optimistic about the ability of market prices and consumer information to help adjust to the changing electricity market in the U.K. and Europe will see in the future. He explains that the main benefit of using prices and information to change the way people use electricity is to “shift that additional load outside peak hours and absorb renewables or solar during the day.” By charging an electric car during the day, rather than when people get home, it not only avoids the hours of peak demand, it uses solar energy that is available during the middle of the day. Users get two benefits at once – cheaper energy and smaller environmental impact.
Open data for green energy
The U.K. is just beginning to see these innovations make a difference, and there is still a long way to go. As the number of innovators grows, they realize they are part of a community. Mick Wall is building a list of devices and apps that work with the Octopus API. These range from intelligent plugs like Ecopush, to chargers for electric cars. More innovations are most certainly coming.
The catalyst for those creating these tools that take advantage of Octopus’s variable rates was to make the data available. “Without the APIs we wouldn’t be able to do what we are doing,” says Bolsakovas of Homely Energy. That data is also helping improve the performance of future innovations. “We also need more accurate data to make our algorithms work properly,” and instead of just using data from Petruskevicius parents’ home, they can now use the enormous amount of data provided by the API and their hardware. “The more open data we have the better.”
This seems like small stuff, but the potential is huge. A study by Octopus found that customers with Agile pricing and an electric car, shifted their energy use from the more traditional pattern – peaking around 7 p.m. – to peaking in the early morning hours around 2 a.m. This shift reduced their total electricity use during peak hours by nearly 50 percent.
Open data also means an increasing diversity of solutions to suit the needs of customers. Not everyone has an electric car or heat pump. For many people, the primary electrical draw may be a clothes dryer. Whatever the case may be, from a simple web page like energy-stats.uk to a more sophisticated controller using data-driven algorithms like the Homely heat pump controller, customers can already find something that suits their needs. As more electric companies introduce similar pricing systems, the number of customers who can benefit from these tools will grow, and so will the incentives to create more apps.
Octopus seems to agree. After Storm Atiyah moved through, the company managers wrote on their blog, “For years, the mainstream thinking was that renewable or sustainable products were expensive or niche. Over the weekend, a combination of smart meters, smart energy tariffs and a couple of windy nights helped us make the U.K.’s greenest energy the cheapest too.”
Petruskevicius wants to be part of undermining that mainstream mindset. “I think it is the future,” and services like Homely Energy’s will be more necessary “when we have a lot more loads like batteries, EVs and heat pumps.”
These early energy hackers did not intend to overturn the mindset and change the way the U.K. – and, perhaps, others in Europe and North America – think about and use electricity. Curiosity and self-interest were their primary motives. As the changes in electricity generation and markets accelerate, however, hackers like Wall, Bauters, and Petruskevicius will be pioneers in that important transformation.