The traditional design of the electric power infrastructure, from generation to transmission and distribution, was built to meet the peak demand plus an extra reserve margin just in case.
If the airline industry were to do the same, they would design their system to have sufficient capacity to handle the peak demand period – during the summer holidays or on Thanksgiving weekend – which is the heaviest travel period in the US.
While this would mean sufficient capacity to handle the crowds on the busiest times of the year, it would result in uneconomical spare capacity on nearly all other days of the year – an expensive proposition that no airline could afford.
The result is an expensive network with a lot of capacity sitting idle most of the time, a burden collectively paid by the hapless customers.
It has historically been justified since no utility CEO, grid operator, utility regulator or politician would want to experience a power blackout on her/his watch – as happened in the Iberian peninsula in April.
When the industry was experiencing declining costs, customers didn’t mind it as much. And because everyone was paying for the upstream infrastructure in proportion to their consumption, the costs were spread across large volumes of sales. It was an expensive form of burden sharing.
But as technologies to self-generate and store electricity on ever smaller scale are gaining traction, some customers can buy less from the network thus reducing their contribution to the burden sharing.
For this and other reasons, many believe that the time has arrived to rethink not only how to collect sufficient revenues to maintain the infrastructure but also to design it for a future where different customers derive different services from the network and have to be charged accordingly.
Rob West of Thunder Said Energy (TSE) recently said that the grid connection capacity needed by a typical UK home could be reduced by an astonishing 80% with significant cost savings.
He claims that a town of 20–50,000 homes could be powered with a 40% smaller distribution network, which doubles the grid’s effective utilization rate while reducing the costs.
Since different customers’ reliance on the distribution network varies by time and location, and because electricity can be stored in batteries and other devices or media, experts believe that the cost of delivering grid services can be significantly slashed by optimizing the generation, delivery and storage.
Artificial intelligence (AI) and machine learning (ML) should come handy in deciding how best to maintain and operate the complex utility infrastructure in real time.
This may be part of the reason for a projected 7-fold increase in the global deployment of batteries from 70 GW in 2024 to 500 GW by 2050, creating a $400 billion per annum battery market.
Other attractive opportunities exist in improving how individual retail customers are served. TSE has catalogued the power consumption of a “typical” UK home – device-by-device, every 6 seconds.
A typical home has a minimum load of roughly 500 Watts – higher for the US where homes and appliances tend to be bigger. This is the minimum load to keep devices that are always on. The median load is 630 W with 800 W for the upper quartile.
The boiling kettle, a fixture in all British homes, uses 3 kW when on, which usually takes 2–4 minutes when used. It is off 99.7% of the time. But during big sporting events and popular TV programs, virtually every kettle is turned on during commercials which results in a noticeable spike in demand. Grid operators must have sufficient spinning reserves to meet the increased demand even if it only lasts a few minutes.
Other electrical devices are also innocuous individually – especially since different customers use them at different times.
A typical dishwasher, according to TSE, is on for only 3% of the hours each year, using 1.7 kW for 30-minute periods at the start and the end of the wash cycle. Similarly, the washer-dryer is on 3% of the time, running from 40 minutes to 2 hours, varying from 250 W to 2.5 kW depending on the model and vintage.
While individually insignificant, if used together, they can spike the household’s load to 10 kW, which occurred at 7:15 pm as illustrated in the chart below.
This happens because several appliances are on at the same time – the kettle, the dishwasher, washer-dryer, TV, lights and other appliances. Even if this happens infrequently, it does happen frequently enough.
And if many homes experience spikes in the demand at the same time and on the same distribution lines, the network may get overloaded. The networks are oversized exactly to handle such situations, which makes them expensive to operate and maintain.
One way to address the issue is to reduce the grid-carrying capacity of the distribution network by including batteries in strategic locations. According to TSE, “It would be fully feasible and more economical, for this particular house to have a 2 kW grid connection and a 10 kWh battery, rather than a 10 kW grid connection.”
Improvements such as this would increase the utilization rate of the grid infrastructure while reducing the costs of delivering services.
There are many who are convinced that bottoms-up distribution-level planning, if routinely applied, will identify such cost-saving opportunities, allowing improved services at lower cost.
The idea would be to examine customers’ energy service needs by recognizing how much can be generated, consumed and stored locally through cost-effective self-generation and storage.
What customers cannot produce themselves can be delivered by the network while recognizing the diversity of loads and resources across neighborhoods and localities.
Such schemes not only reduce costs but increase resilience and improve the reliability of the network while distributing the costs and benefits in a more equitable and fair way among the customers. The days of delivering undifferentiated services to all customers and charging them uniform volumetric rates appear to be numbered.
