Latest Issue

Demand side, it’s all about flexibility

By Jon Bernstein  

This article was produced in association with Shell

Find Prospect’s full report, “The Future of Energy” here

It’s a common trope beloved by PRs and headline writers alike. When seeking to explain the benefits of a new product or service, businesses seek to compare it with something else that’s equally shiny and new but, crucially, something that’s already caught the public imagination. So the new but unknown service becomes the iPhone, the Uber or the Bitcoin of the industry in which the unknown start-up operates. The result? Credibility, understanding and brand recognition by association. Or at least that’s the hope.

So when Open Energi, pioneers in demand-side smart energy technology, wanted to tell its story to a mainstream audience, it reached for something suitably zeitgeisty: the sharing economy. For the duration of one Huffington Post article, at least, Open Energi became the Airbnb of energy.

The piece included the following passage: “By connecting everyday equipment to a smart platform (just as you might upload your property to Airbnb), it’s now possible for consumers to take advantage of small amounts of “flexible demand” in their existing assets and processes—be it a fridge, a water pump, or an office air conditioning unit—and sell it to organisations tasked with keeping the lights on—like the National Grid.”

The analogy works up to a point but it doesn’t really capture the full nature of the challenge and the nuanced response that demand-side flexibility promises.

The challenge? How to deliver flexible forms of capacity when dealing with renewable sources of energy. “It’s not just capacity,” David Hill, Open Energi’s Commercial Director—and author of that Huffington Post article—tells me. “It’s flexible capacity that can ramp up and down at very short notice.” When dealing with intermittent supply, such as wind and solar power, then flexible capacity management becomes a fundamental need and challenge.

The promise, meanwhile, is two-fold. First, energy use can be deferred without any adverse impact on the intended heating or cooling. Second, it makes it possible for equipment to switch energy source as appropriate—to take advantage of excess wind on a windy day, for example. Using demand-flexibility techniques, wind energy that would otherwise be lost can be captured and used.

Demand-side flexibility works by taking advantage of thermal or pumped energy stored in a wide range of commercial and industrial equipment and processes, from a supermarket refrigerator to water pumped into a reservoir. Provided they work within defined parameters—think temperature or water level—consumption patterns of these functions are rarely time dependent. That means energy use can be flexible, equipment can postpone usage and organisations can make that excess capacity available to the grid.

Open Energi believes the impact of this approach will be significant. The company estimates there’s about six gigawatts of innate demand-side flexibility within everyday business processes. “To put that into context,” Hill says, “that could power 2.5m homes for an hour without any need to ramp up a power station. That’s about 10 per cent of the UK’s peak winter demand for an hour, without any impact. Now that’s quite a change in the way the electricity system works.”

Once energy use is deferred and energy sources interchanged, then it’s possible to move to a model that doesn’t require power stations built for peak demand. A more agile system takes its place.

“We’re not saying we can solve the flexibility problem indefinitely—because you can’t shift demand indefinitely. Yet without the need to build new gas fire power stations and deploying mass batteries, just by manipulating the way we consume energy in a very predictable, reliable and unobtrusive way you can save a huge amount of power.”

Won’t energy-consuming industrial systems and machinery need to be adapted in order to take advantage of this kind of software intervention? And won’t that upfront investment cost be off-putting? “Certainly when we were starting out it was true that utilities and plant owners had to put some investment in place to make these assets smart and capable,” notes Hill. “It’s of no surprise that a water pump built in the 1970s doesn’t have the controls or the capabilities to do what we wanted it to do. So there was an investment in terms of upgrading.” Not only are those devices much cheaper today but the pay back,
says Hill, is one to two three years compared to a return on investment for a power station of 30 years or more. “It’s a different ball game.”

Currently 50 customers, across 382 sites deploying an aggregate of 3,150 devices, are using Open Energi’s Dynamic Demand technology. Customers include water utilities—Seven Trent, United Utilities and Welsh Water among them. “They are a good fit,” says Hill, “because a huge amount of their demand is flexible. Pumping water or dealing with waste is not time-dependent.” Moreover, water utilities not only use large amounts of energy but because these firms are spread across the country, access to wind energy is likely to vary which in turn means they are unlikely to make simultaneous demands on the same energy source.

Other clients include those from the quarry and mining sector that make intensive use of industrial heating applications. Open Energi also counts supermarket Sainsbury’s as a customer. Technology is installed in 200 of Sainsbury’s largest stores. “Working in real-time we’re listening to the health of the electricity system—and the price of the electricity—to make sure we are shifting the power from their heating and ventilation equipment without anybody noticing.”

The software is connected to around 1,500 devices within each store’s building management system and continually polls these devices to understand real-time consumption. If there’s an event on the electricity grid it would lead to a big shortage of power in the system while, at other times, it may be possible to sell unrequired energy back to the National Grid. “On an aggregate level we may have two megawatts of power that we can shift and turn off instantaneously to bring the system back into balance—and Sainsbury’s doesn’t know about it [in terms of outward effect],” explains Hill. “For that two megawatts we get paid by the
National Grid and we share that revenue with Sainsbury’s.”

What happens if by bringing forward heating or cooling events, Sainsbury’s stores become too cold or refrigerators become too warm? That won’t happen, says Hill, because the algorithm has been designed to ensure energy supply never goes outside “control parameters” set by the customer.

Customer control is a theme Hill is keen to stress. “Trust is a big issue and we have very secure networks. And we are very clear that it’s the customer that is defining the parameters,” says Hill. But he points out that a fridge, for example, is already subject to control. “However, that control signal has no bearing on whether it’s a good or bad time to turn on and off, it’s just related to the temperature. What we’re saying is that if you could just listen to the health of the electricity system—price and capacity constraints—then you could do something [more beneficial].”

The consumer is not to be confused with the domestic consumer, at least not yet.

“It hasn’t gone domestic because there isn’t a huge amount of flexible demand there,” says Hill. “If you think most flexibility comes from heating and most heating is gas, then what’s left—fridges—
is quite small, really. That will change with electric vehicles, heat pumps and battery storage used to optimise photovoltaic (PV) panels on the roof. That completely changes the game. You go from households having just a few watts of flexibility to two to three kilowatts of flexibility.”

For now then, it remains an industrial commercial proposition. Key to delivering demand flexibility at scale is artificial intelligence. Machine learning allows Open Energi to manage large amounts of data—it collects between 10,000 and 25,000 messages per second which relate to 30 different data points—and reschedules the power consumption of multiple assets simultaneously. “Traditionally every device had a control algorithm specific to it,” notes Hill. “Now self-learning algorithms make much more sense.” Open Energi has been fine-tuning its algorithms over the last five years and claims it can now shift about 20MW of power in real-time without material impact.

Open Energi is not the only organisation to explore demand-side flexibility or the power of artificial intelligence to manage energy flows. For example, Upside Energy’s Virtual Energy Store manages storage assets and offers real-time energy reserves to the grid.

Meanwhile, Google-owned AI specialist DeepMind revealed in March that it is in exploratory talks with the National Grid about applying machine learning to improve supply and demand management. DeepMind CEO and Founder Demis Hassabis told the Financial Times: “It would be amazing if you could save 10 per cent of the country’s energy usage without any new infrastructure, just from optimisation.”

When Google steps on your patch it’s both frightening and heartening. Frightening because there are few competitors with the resources and the reputation of Google. Heartening because Google’s involvement helps legitimise the market. “It’s exciting that they say things like that because it brings a lot of attention to what you have been saying and doing. It aids your own agenda,” says Hill who argues that the market is big enough to accommodate many more entrants. Hill reveals that some of DeepMind’s machine learning protocols have been a reference point for Open Energi’s own algorithms.

Google is making a relatively late entry into smart energy technology. “One of the reasons I guess Google has been biding its time is because the world in which we operate is not as software-led as we’d like it to be because all the assets and machines we want to interact with the electricity system operate behind 1980/90s legacy control systems.”

What, if anything, does Open Energi want the government to do in order to stimulate interest in smart energy technologies? Is there something Hill’s firm needs? “Not really, everything is taking shape,” he says. “There are market incentives to make it possible to compete against power stations. It can always be better but the UK is leading the way in this area. In other countries it can be prohibitive for technology firms to compete against power stations. In the UK, we’re lucky that the National Grid and the electricity market have been incredibly supportive. There is still a challenge—we’re not always on a level playing field with power stations—but the market incentives are there to develop these kind of platforms.”

Hill believes there is now a consensus around renewables. But is that really the case? In political circles, some point to the loss of a climate change portfolio at Cabinet level—following the abolition of the Department for Energy and Climate Change in July 2016—as evidence that the issue has been downgraded by policymakers. “Look, everybody realises there is a capacity flexibility issue,” Hill insists. “Unlike wind and solar there are no subsidies for what we are trying to do. We are trying to build a commercially-viable product in today’s market conditions.”

Subsidies, says Hill, are a double-edged sword. While they can help to kick start a market they also bring undue attention to a fledgling sector. Notwithstanding, Hill believes the recently-formed Department for Business, Energy and Industrial Strategy, does understand the dual benefit of smart energy technologies: they reduce costs for consumers and are a catalyst to stimulate UK economic growth.

Back to that Huffington Post article. Hill, a business school graduate, described himself in his byline as an “economist and optimist.” Why optimist? “I’m not the machine learning brains,” says Hill, “but I guess I’m the person that sees there’s a large inefficiency in the electricity market… that you can collapse that surplus and transfer the benefit to the consumer. That brings optimism, that there’s a huge amount of wastage to return to the customer.”

“I work with seriously smart people who give me this sneaky porthole into the future. If you have a renewable energy system that is subsidy-free and the marginal cost of producing energy is incredibly small—wind and solar—and you have a demand-side energy platform that is optimising our consumption and storage in real time (and has an incredibly low marginal cost because it’s just data) you begin to get towards a system that is massively cheaper than the one we have today.”

As to that future, Hill believes one medium will have a huge role to play. “Battery storage is going to completely change the game because it’s far more intuitive to understand,” Hill says. “On an industrial scale, it’s [easier to] understand that we can manipulate battery energy use without effecting their productivity.”


To find out more about how you can become involved in Prospect’s thought leadership programmes, please contact

You can also receive the full report as a fully designed PDF document by simply entering your email below.

When you sign up for this free report, you will also join our free Prospect newsletter.

Prospect takes your privacy seriously. We promise never to rent or sell your e-mail address to any third party. You can unsubscribe from the Prospect newsletter at any time

We want to hear what you think about this article. Submit a letter to

More From Prospect