New ideas can be very confusing. When Einstein’s theory of general relativity came out it was said that only three people understood it, but apart from the mathematician involved in supporting Einstein and Einstein himself, no one quite knew who the third person was.
However, the verification of relativity made Einstein an instant celebrity; the first celebrity physicist ever.
Somehow, the general public immediately grasped the novelty of the new theory; they had a sense of its potential, even if the experts were still catching up on understanding what it actually meant.
It’s possible that the conflicts and upsets around energy that we’ve described up to now, stem from the fact that many people don’t really understand the new grid and the theory and technology behind it.
For this reason, we’re taking an in-depth look at how the new, distributed grid works.
What’s become apparent is that far from being a plug-in option, and allowing a grid to continue business as usual with a few modifications here and there, renewables seem to be forcing a complete rethink of the entire architecture of the grid system.
In fact, not only do renewables require many elements of the grid to be adjusted, it’s actually simpler to see renewables as inverting every single assumption that once belonged in the old grid.
For where we used to have centralized, we now have distributed; for fossil, replace with solar; for base-load, read intermittent; for AC read DC; for hierarchical we now have peer; and for hub and spoke, read cellular. The list of inversions goes on but I’ll say more about that later.
But first, let’s start with a brief summary of the old grid.
In a classical old grid system we have a generator powered by coal, gas or nuclear.
It’s a thermal device, meaning it takes some kind of fuel, burns it - creating heat - which in turn creates steam and then motion, like a modern day steam engine. The motion drives an armature of metal wires past magnets and this creates electricity.
In the old system, a big powerful singular base load generator sits in the middle of many households and some factories that consume the electricity the generator is producing.
Sometimes, when clusters of households are situated many miles from the generator, large high-tension cables deliver the power over longer distances. The costs of this, not to mention the losses, can build up.
Here is the old system, simplified somewhat.
There can be a series of generators in a large network, but essentially they are few and the household consumers are many.
Classical baseload generators can be located in many different places, but in schematic architecture terms, they’re central and they serve consumers at the periphery. So the power goes outward, and payment and demand for power comes inward.
Apart from being central, the baseload generators are also constant and very dependable. They hum along being dialled up or down according to a typical daily schedule of usage.
While the usage may vary, it’s the job of the central grid manager stationed at the hub to track and supply the overall demand from the consumers.
Occasionally if there’s a spike, a ‘mid-merit’ or even a ‘peaker power’ plant is brought into play.
In the UK, such an event could be a big football match or royal wedding where everyone makes a cup of tea at almost exactly the same time, causing a sudden call for extra power.
In this model, the centre serves the periphery, and because the consumer at the periphery is largely predictable, the baseload can be adapted to meet their needs and maintain the grid.
It’s tempting to think that one might simply be able to slot in some renewable capacity where previously there was thermal.
But as we’ve already seen in Germany, solar and wind don’t have the dispatchability that thermal does. Even when you pair thermal with renewables in tandem the central system shown below starts creaking and becoming more expensive and unmanageable.
A scheme like the one below would obviously fail on many counts.
Turning the centralized scheme on its head and creating a distributed model isn’t just nice to have, nor a gimmick. The new grid is emerging as perhaps the only way renewable energy can overcome the dispatchability issue and be made to work.
So what happens when you turn everything in the rule book upside down? Firstly, the paradigm is no longer command and control, central and periphery. Rather, it operates more like a living cell.
In a household of the near future, you will find the consumer function, in that there’s a household that consumes electricity, but you will also have a solar array on the roof producing kilowatt-hours (kWh). This producer function gets elided with the consumer function giving us the term prosumer.
You will also have some storage capacity because you need to store some surplus kilowatt-hours to keep the whole system working. Right now, there’s not nearly enough battery storage capacity, so peaking power units are often called in to top up the shortfalls.
Because these ‘peakers’ are so expensive, it’s becoming clear that there’s an arbitrage opportunity for anyone who can organize themselves to bring in some storage to the system.
This means prosumers who invest in a battery megapack will find their investment pays for itself in a matter of a few years.
The other benefit of the new architecture is that because of the cellular paradigm, long distance delivery of power is less likely to be needed. In fact, long distance, high-tension cabling could account for as much as 30% of all electricity costs in some cases, so the cellular paradigm is inherently cheaper from this point of view.
In the new grid, production and consumption exist within the same household or factory or campus unit, but at any one time they’re either producing, consuming or storing (or some combination of the three).
They also will be doing some reciprocal activity with another household cell nearby, as seen in the diagram below. Here, one prosumer is selling power to an adjacent or nearby prosumer, either for consumption or storage.
Because there’s such a large number of individual prosumer cells, a market evolves that reflects their state at any one time.
This is the essence of the cellular paradigm and how the power part works.
Note that we still have information power and money flowing through the system just as we did in the old grid.
The difference is that these flows are now bi-directional. Where in the old system, power flowed from the central generator to the households and factories, with the cellular system, power can go from any one cell to any other.
So too can money and information about price and need. That means all cellular units are more or less equal in the hierarchy.
Also note that not every cell has to do all the functions. Some might be houses with no solar panels, or factories with no battery storage.
Perhaps the most obvious example would be an electric vehicle. We could represent it in this scheme as follows: a battery, and an appliance that consumes energy (the vehicle’s motor), but no solar cells.
At the next level, each of these cells has a control function that allows it to interact successfully with the rest of the grid.
The first of these is a kind of artificial intelligence brain that essentially works out the price at which the household appliance or electric vehicle is willing to offer or buy from the other cells.
In the new grid, prices can rise and fall very quickly, for extremely brief intervals of time, because price is the communication that ensures supply always matches demand.
In the old system, which was more like a Soviet-style command economy, matching supply with demand was, of course, done slowly and centrally. These lead to shortages and instabilities which we explored in Part II.
So the price flexibility arises partly from the market conditions at that time and place and also from the need state for the device. Need state is important in determining the bid and offer price, so let’s see how that works in practice.
Let’s take an example of a transaction and see how it plays out.
Let’s say Orange, an orange car, and Green, a green car, are parked on the same street. They are the same make and model and are both roughly half charged and connected to the new grid.
Orange knows it will be used the next day for the school run, and Green knows its owner hardly ever needs it on a weekday, and certainly not before lunchtime.
So Orange needs to charge up, and could buy from Green, which might want to sell some opportunistic kilowatt-hours.
Parked and connected to the grid, each car creates its own bid and offer price for a transaction.
Because the AI unit or ‘brain’ in the orange car knows that it is due for the school run first thing in the morning, Orange has a need state of 8.
Green’s brain checks the diary and service history and evaluates its own need state at 0.
These need states are, of course, reflected in the bid offer prices. Orange shows a bid of 13¢ per kWh and an offer of 27¢ per kWh, while Green bids 5¢, and offers its own stored electricity at 14¢ per kWh.
In this instance, Green is effectively positioning itself as a retailer to Orange the consumer, with a slight discount on the ambient price of 18¢ per kWh.
The two AI brains from the two cars negotiate with each other, and the need state of Orange drives up its bid price to 14¢ and the two cars transact some energy. By the end of the deal and some charging time later they look like this:
One of the interesting features of this process is that the green car has been instrumental in managing a price spike.
It’s saved an expensive ‘peaker’ generator from having to be called in, in the larger system. It has also reduced the need for a big HT power cable to bring the power in over large distances, with the 30% clip on the ticket and electrical losses that go with.
Note that if there had been a local shortage with prices rocketing to say 29¢ per kWh, the orange car might have even sold some of its electricity at 27¢ to help out the grid, rather than buying from Green.
But the kids would have had to walk to school. At the end of the transaction, the ambient price of 18¢ per kWh might also shift upwards to 19¢ per kWh, reflecting the demand made on the system by Orange with its relatively high 8 need state.
Finally, the profit the green car makes will help fund more battery capacity to manage the grid in the future.
We can say this because Green knows that while the neighbor needs electricity first thing in the morning, in the afternoon the offer rate for kilowatt-hours might even be a negative amount, as it was in this example at -3¢ per kWh. With plenty of renewables on the grid, the modern grid must sometimes pay you to take surplus electricity out.
As well as need state, bid and offer, the electric car of the future will also have a geo-positioning function. This means that a car can spot price opportunities from around the surrounding area and work out a route that recharges where there’s a low price recharge available.
To see the geo-position function in action, think about a car taking advantage of a price gradient over a geographical area.
In the above example a car might elect to drive a small detour to charge up where there is a surplus of electricity. At the recharge way-point, the price is low, maybe even negative.
In this example, the car is in a symbiotic relationship with the grid, helping prevent overproduction of kilowatt-hours, at hotspots which could create reverse flows and damage grid components.
Looked at from the point of the grid, the car has performed a similar function to a high tension cable: it is translated power from one location to another.
Finally, the last and arguably most important control function, is a meter identifier that’s connected to a blockchain-based account. This means all of the cell’s transactions can be frictionless, i.e. made without bank charges, and done with high levels of trust.
Remember, in the distributed model of electricity, there’s not necessarily a central electricity board that sends everyone their electricity bill. It can be done in a peer-to-peer fashion.
So these are the five elements that come together in the control of an Internet of Electricity-type device.
To sum up, the control functions are:
1 Artificial Intelligence - to work out bids, offers, and need state based on understanding the market and the humans it serves.
2 Need state - to determine when to accrue power or, alternatively, profit.
3 Blockchain - or accountancy trust and tracking transactions on a ledger.
4 Geo-position - position of the car and likely position at next charge point.
5 Bid and offer - 0n all power going in and out of the unit.
In this chapter we looked at the technical elements of the new grid and saw how they come together to create a distributed version of the hub and spoke ‘old’ grid, with a cellular architecture. We saw how all the major concepts that were present in the old grid are upside down in the new grid.
We saw how entities like power, money and information can come from anywhere, and go to anywhere. In the next chapter we’ll see more about how the new grid performs economically, and how it can deliver non-perverse incentives to the grid.
We’ll also investigate a property that’s really quite different from the old model; demand response. In the new grid it's an ingenious way to match supply and demand, and we'll find out how that is achieved.
This architecture is still evolving and as we speak, new ideas are firming up about how things are going to progress. Like most revolutions, the revolution in energy will take us to places that are much more exciting than anyone ever anticipated at the start, and require more innovation elements coming together than anyone imagined.
Sign up for our newsletter - full of great insights and market updates.
