Since it was built, our energy system has largely revolved around altering our energy supply to keep up with demand – but attempts to influence demand itself aren’t new. In the UK, for example, Economy 7 was launched in October 1978. Economy 7 uses base load generation to provide cheaper off-peak electricity at night (so that customers can power their electric heating). For the last sixty years, efforts to ‘play with’ energy demand have been defined by (and limited to) the concept of ‘Demand side response’ (DSR). Beyond the UK, DSR was included in the 2005 Energy Policy Act in the USA. In New Zealand and South Africa, peaks in electricity demand have been managed using ‘ripple control’ since the 1950s, and commercial ripple control was first established in France in 1927.
DSR is a neat solution to an old problem. It was designed to keep coal fired and nuclear power stations running smoothly by shifting relatively predictable electricity consumption to times of relatively predictable low demand. However the world has changed since then. The energy system is evolving significantly as we add more intermittent renewables to the grid and electrify heat and transport. As a concept, DSR is at an evolutionary dead-end. It isn’t just DSR’s age that’s a problem – the ideas behind DSR are dated: they won’t get us to net zero. Instead, we should be thinking about ‘Intelligent Demand’.
What is wrong with Demand Side Response, isn’t it what we want?
DSR is fundamentally a static tool. It is predicated on moving demand out of, or occasionally enabling generation supply into, predefined time-periods based on historical peaks in demand. Generation still follows energy demand, so when demand increases, the system operator turns up the coal-fired and gas-fired power stations. Apart from Economy 7 customers, in the UK, consumers generally don’t shift demand. While this has worked up until now, the way that networks and system operators plan for defined consistent peaks needs to dramatically evolve if it’s going to be fit for purpose in a new, net zero world.
The generation mix is changing rapidly (see Figure 1). Renewable generation capacity, primarily wind and solar, has quadrupled over the past 10 years from 12.5 GW in 2012 to 49 GW in 2021.
Figure 1: UK Electricity Capacity 1996 to 2020
The rate of change in the electricity system is forecast to continue (see Figure 2). In 2021, wind and solar made up 43% of domestic electricity generated, but by 2030, these renewable technologies are forecast to dominate, accounting for 66% even in the most pessimistic of the National Grid’s “Future Energy Scenarios”. Intermittent renewable generation cannot be dispatched in the same way as the old coal-fired plants or the gas peaking power stations. If there is more generation than demand, it needs to be stored in batteries, in pumped hydro or curtailed (wasted). Indeed, based on its current forecasts, National Grid estimates that at least 15 TWh will be curtailed in 2030. At 28p/kWh (what consumers pay for an hour of running their house on electricity when their load is roughly 1kW), that is £4bn worth of wasted energy per year, roughly equal to wasting £150 of electricity per home. Conversely, when renewable generation is lower than demand, the system operator turns on CO2 emitting coal and gas peaking power stations to meet demand.
Figure 2: Installed renewable generation capacity by scenarios (GW)
Electricity demand is also changing rapidly. The Committee on Climate Change (CCC) forecasts that total yearly energy demand will roughly double, driven by electrical heat pumps replacing gas boilers and electric vehicle chargers replacing petrol stations. One other source of increased demand is hydrogen – we will need to make green hydrogen to power industrial processes that cannot be electrified. Large parts of these increases in demand are much less predictable and more weather correlated: the colder it gets, the more energy is used in heating, the warmer it gets, the more electricity a heat pump (running backwards ) consumes.
To give you some numbers of the scale of the change in demand, it is worth looking at what industry uses as a basis for planning. In the past, houses had a winter peak usage of around 1kW for customers without electric heating (Figure 3), and around 2kW (at an off-peak time) for customers on Economy 7 (Figure 4). In both cases, demand was also relatively predictable, and the team who ran the electricity system (National Grid Electricity System Operator) could more or less plan on the basis of the load profiles below, and corresponding industrial and commercial load profiles.
Figure 3: Profile Class 1 Customer + Figure 4: Profile class 2 – generally economy 7 customers
Going forward, overall demand is set to increase significantly, and if left unmanaged (an ‘unintelligent demand’ scenario), the peak could increase astronomically. EV chargers are around 7kW for the average single phase charger, and three phase chargers are around 11 – 22kW. Heat pumps for an average house add an additional 2-6kW peak load, but they only draw on electricity half the time (the heat pump turns off and on), so the average load is around 1-3kW. As a result, unintelligent household demand could increase dramatically at certain times of the day. From its traditional 1kW winter peak, an average house with an average car charger and heat pump could see an additional 9-12kW load. If you design power networks and run generation plants in this ‘old world’ context, that sort of increase is pretty scary. But it is also a crazy expensive way to run an energy system.
Up until now, generation has been controlled so as to follow demand. In this context, adding a whole lot of non-dispatchable wind and solar to the grid is also pretty scary, because ‘old’ DSR doesn’t shift around like the weather. We currently have around 10GW of wind generation across Great Britain, which is forecast to increase to around 40GW by 2030, and around 14GW solar is forecast to grow to 2-3 times this level. In the future, on a windy, sunny day across GB, we might have 60-70 GW of renewable generation powering our electricity. This compares to a normal summer peak of electricity demand around 35 GW. If we want to keep costs down – we better have smart and sensible ways to use this excess electricity.
Intelligent Demand can help keep costs down as the grid evolves
This is where Intelligent Demand comes in. In the graphs above, you can see that energy demand ‘peaks’ between 4-7pm on a winter night. Economy 7 was a way to try to move demand to a fixed overnight period, away from this peak. But if there is lots of wind at 6pm on a winter’s night, we want consumers to be charging their cars, and running their heat pumps on boost to use up the surplus energy (so long as the customers’ electricity network can manage this load). In other words, in this scenario, we would actually want to turn-up demand at a peak time.
In contrast to a deficit, a surplus happens when we have lots of wind and sun on the system, or little demand. We want to shift demand into times when there is a surplus of energy, even if this surplus happens at a peak times – because otherwise we will waste this electricity (see graph below of Intelligent Octopus managing EV charging). We have to move demand to the times of the day when electricity is cheapest and plentiful, and out of the most polluting times, when electricity is expensive. In other words, we need Intelligent Demand’.
Figure 5: Graph of wholesale electricity price (grey bars) versus Intelligent Octopus Charging (red line) within a specific distribution zone (Notice the increase in charging in response to negative prices at 23:00 on 01/01/22, prices have gone down, and so Intelligent Octopus sends a command to the EV to start charging)
If you think this sounds futuristic, you’d be mistaken. Information about surplus or scarce generation is already provided via wholesale prices. For example, low wholesale electricity prices correspond to periods of time when available energy outweighs expected demand. What’s more, it’s now easy to get an EV charger to start charging and use up that excess electricity (especially since the clever team at Octopus have launched the Intelligent Octopus tariff to make smart usage easier for ordinary people).
Enabled by our cutting edge Kraken platform, Intelligent Octopus optimises your car charging by looking at the electricity prices over the next 24 hours, the time you need your car ready by, and taking into account the state of other essential components of the power network, for instance how much congestion there is (more on this later). Intelligent Octopus can also work out when your heat pump needs to go into ‘super boost’, and when you should charge or discharge a home battery.
Intelligent Octopus can do ‘turn-down’ too – for instance if Intelligent Octopus is communicating with a customer’s heat pump and their EV, we can turn down both devices (or just one). We’ll always take care to ensure that the car is still ready when the customer needs it, and that we only turn down the heat pump in a way that doesn’t affect the temperature of the house (smart heating). All of this means that we can use electricity when it is cheapest and most plentiful, and pass on the benefit to customers. For instance, if you live in the centre of Manchester, and were with Intelligent Octopus in September 2022, you would pay 7.50p/ kWh between 11:30pm and 5:30am (while we schedule the charging for your car over this time) versus the standard rate during the day of 39.25p/ kWh – that is around one fifth of the day rate in September 2022. At a national level, savings from Intelligent Demand and other forms of flexibility are significant. By 2050, BEIS has forecast related savings of around £10bn per annum, while the Carbon Trust has forecast a higher figure of £16.7bn for their high electric heating scenario.
Can ‘Intelligent Demand’ help ease problems with network capacity?
Of course, the eagle-eyed energy geeks amongst you will have noticed that this only handles the problem where demand turn-up and excess generation are in the same part of the network, or where there are no network constraints, and electricity can flow freely anywhere. But sometimes the electricity network has limited capacity – for instance between Scotland and England, at the B6 boundary (see map) the network is often at capacity.
So what about network capacity? Can Intelligent Demand work its magic here? In short, yes – Intelligent Demand helps here too, but we need to think about two parts to the problem. First, we must consider network constraints at the very high voltage level – the large transmission lines that take bulk high-voltage electricity to grid supply points around the country. The control centre team that manages the electricity system at National Grid ESO (ESO) has a market – the balancing mechanism – to try and manage high voltage level demand imbalances.
The ESO uses the balancing mechanism to buy the right amount of electricity, in the right place, in order to balance the system. Octopus is already participating in the balancing mechanism with consumers’ devices to turn up or down demand for the ESO. Of course, we don’t offer ESO the energy from an individual EV, we bundle all the smart consumer devices together and offer this aggregated turn-down (or turn-up) demand.
What’s next for Intelligent Demand?
Ultimately, all of this is only possible if we stop thinking about the electricity system as it was. Demand side response is an outdated concept. So long as the industry and regulator remains focused on demand side response, it will be more difficult to move to a less expensive, truly flexible, intelligent electricity system. Intelligent Demand is multi-dimensional tool, and can be used to solve a range of problems across the electricity system: it means less wasted renewable energy and therefore reduces the amount of generation that GB needs to build; it enables more efficient use of network assets and lower network costs; it can help fix constraint problems at a local network level, and can help National Grid manage the overall electricity system; and lastly, if properly mobilised, Intelligent Demand can also be used to help provide system stability.
So here’s to increasing peaks in times of energy surplus (and keeping power network engineers sane by using the precision that Intelligent Demand unlocks to help them manage their networks more easily).