In the first of a three-part series on the development of the future power grid, Wolfgang Ketter, Professor of Information Systems for a Sustainability Society, co-Director of the Institute of Energy Economics at the University of Cologne and energy policy adviser to the German Government, looks at the ongoing challenges of maintaining a stable and reliable grid during the energy transition

I assume – as FGS is an industry-related magazine – many of you know the ins-and-outs of electricity grids: their importance, their interplay with markets, and the challenges grid operators face in keeping electricity affordable and reliable. But as an advocate for digitalization of energy and electricity landscape as a means to a faster, more stable transition to a clean energy future, I interact regularly with business people who are struggling to understand how the transition will affect their industry. Some look at the changes from a cost-management or risk-mitigation perspective; others are already innovating business models to use their assets to profit from the transition. Either way, both perspectives are reasonable. The future energy landscape is closer than most of us imagine.

For that reason, I will offer a short ‘primer’ on electricity grid management at the micro- and macro-level for readers who are not electricity engineers. In the second and third installations of this series of articles I will dive into my specialty: how artificial intelligence and machine learning will help the world make a quick change to the dynamic, real-time electricity markets we need to make this transition work for all of us.

But first, to understand the challenges that electricity grid operators face during the energy transition, we need to look at how the grid works and what their primary responsibility is. Then we will look at how modern the grid is – and why modernity matters.

Grids distribute electricity from the source of generation to the points of consumption. A long time ago, power production was decentralised because electricity grids were so unreliable that companies needed to have their own power plant. As grids became more reliable, power production became more centralised and remained so until the arrival of renewable energy sources, when it began decentralising again.

Distribution System Operators (DSOs) manage the grid at the local distribution level. Transmission System Operators (TSOs, as they are known in the EU – in the US they are generally referred to as Independent System Operators) manage grids at the high-voltage level. They also often manage 1) the wholesale forward electricity markets that buy future commitments to deliver electricity on a given day to local DSOs as well as 2) the retail balancing markets that make up the difference between expected supply and demand and actual supply and demand. Managing the gap between expected and actual supply and demand is critical in maintaining grid reliability and even stability.

So the TSOs have a crucial job. Electricity, like healthcare and transportation, is a critical public infrastructure. As soon as any critical public infrastructure is disrupted on a large scale, you start having severe repercussions in other critical public arenas. Without access to reliable power, hospitals cannot run their life-saving equipment. Without access to a healthy workforce, road crews and power stations cannot maintain their infrastructure.

Complexity of critical infrastructure
With COVID-19 we are seeing the interrelated nature of our societies and economies and the complexity of managing critical public infrastructure. The healthcare axiom is ‘first, do no harm’; for electricity, it is ‘keep the lights on’. Individuals, cities, and nations run on reliable access to electricity. To keep the electricity flowing, a grid operator needs to make sure that there is neither too much nor too little electricity in the transmission and distribution lines at any given time. Predicting both the supply going in and the demand taking it out of the lines is critical to providing reliable service.

For decades, grid operators have used sophisticated models to predict demand depending on time of day, typical patterns during the year, weather predictions; all kinds of information go into predicting demand. Those predictions are then aggregated, and an advance auction is held to arrange contracts for future power generation. Big centralised power plants, whether nuclear, coal, or gas-fuelled, make a fairly accurate prediction of how much electricity they can provide on that future date and how much it will cost them. It is a sophisticated dance, but grid operators had the moves down, until renewable energy came along.

In many ways, renewable energy is everything that fossil-fuel generated electricity is not. It is cleaner, cheaper, and getting cheaper every year, and often decentralised. But to the grid operators, most importantly, renewable energy makes supply and demand harder to predict and ultimately, it is making grid stability more tenuous. That has a lot to do with the difficulty they are having balancing supply and demand.

Balancing Supply and Demand: variability of supply and demand is increasing
If everyone in your neighbourhood turns on all their appliances at the same time, you might see the lights dim – the health of the entire electricity distribution infrastructure is sensitive to changes in supply and demand. Conversely, if your neighbours all simultaneously switch off all their appliances, that could present a challenge as well. Electricity that is generated, whether by a large central fossil-fuel-powered plant or a solar panel, must flow somewhere: there is not widespread, centralized on-site storage capacity (yet). On the connection points along a distribution line (demand), there needs to be enough – and not too much – electricity flowing to satisfy demand at any given moment. If there is too much or too little electricity in the lines, the system can be damaged.

One challenge operators face is simply that the amount of demand for electricity is growing faster than the capacity of the lines to deliver. Adding capacity to the grid is costly and time consuming. Our ability to come to public agreement to fund and build grid capacity is outpaced by our ability to demand more electricity. It is unlikely that we will overcome this problem any time soon.

But regardless of increasing demand over time, to ensure reliable electricity delivery, grid operators must achieve a ‘balance’ between supply and demand at any given moment. So apart from growing demand, we need to look at how the ‘balancing challenge’ becomes trickier simply because of renewables.

Modern grids use increasingly modern technology, but the system of grid management was designed a long time ago and some fundamental aspects of it are hard to change. Our grid design is still best suited to electricity produced centrally in large power plants. Those plants are powered mechanically in large part by carbon-emitting fossil fuels that drive rotating machines. Those turbines possess a lot of inertia and that inertia means that the electricity supply does not change rapidly and is easier to predict at any given moment.

The transition to renewable energy is great for the planet and for our economies because renewable energy is cheap and clean. But there are several aspects to the increasing contribution of renewable energy generation to electrical grids that make reliability a challenge for TSO and DSO grid operators.

Keeping the Lights on
Remember, the motto is ‘keep the lights on’. Reliable electricity. If there is too little electricity in the grid, your lights are dim, your air conditioning will not work, and your EV will not charge. There might even be some damage to the infrastructure. If there is too much electricity in the grid, the damage could be quite severe. Grid operators have several ways they can ‘balance’ the available supply with demand.

If there is too much demand or not enough supply, grid operators can add electricity from ‘quick-ramping’ gas turbines into the grid. But that still takes a while, those gas turbines cost a lot to ‘ramp-up’ on short notice and they are particularly polluting. So, if you use them as a back-up to cover the fluctuations of renewable energy production, the cost and pollution of gas turbines can negate the positive value of the renewables you used in the first place. But if an operator must call on gas turbines to prevent damage to the grid, they will.

In worst cases, to prevent damage, operators might even engage in ‘load shedding’ – a controlled shut-off of a whole neighbourhood or part of town during periods of high demand to prevent demand congestion from causing damage to the grid. When this load shedding moves from one hour to the next and from one neighbourhood to the next, that is a planned ‘rolling blackout’. In that case, grid instability leads operators to sacrifice reliability now (delivery of electricity to some connection points) for reliability later (ability to provide electricity for a longer time to far more connection points.)

If there is too much power coming onto the grid, they can ask generators to curtail production. But that also takes a while. They can give the electricity away to ‘neighbours’ for free, or even more financially destabilising, they can pay large industrial customers to consume the surplus. Negative pricing discourages investment into electricity production capacity so that is not a good sign for economies in which demand is increasing over time.

Widespread damage to the grid is not only costly, it is dangerous for the local society and economy so grid operators work hard to prevent it. Ironically, grid operators will break with reliability and affordability/investment attractiveness on the short term to preserve the stability of the grid so that reliability is easier to provide on the longer term.

Managing Capacity
In the world of electricity generation, distribution, and consumption, the word capacity comes up a lot. It can be confusing. In generation, capacity is how much electricity a power plant or wind farm could produce if running at full speed, so to speak. A coal-fired power plant can quite accurately predict its top generation capacity, but it does not have to always run at top capacity. It can take measures to control its output.

A wind farm has a top limit to capacity as well, but how much of that capacity is actually generating electricity depends on the strength of the wind at any given moment, and even different windmills in the same farm may be generating different percentages of full capacity depending on the wind strength at each mill. This makes the actual generation capacity of renewables more difficult to predict.

In transmission and distribution, capacity is how much electricity the grid can handle transmitting or distributing. Just like your water pipes – bigger pipes can carry more water – higher grid capacity can deliver more electricity.

In consumption, instead of saying ‘capacity’ to consume, you measure (potential) electricity consumption by ‘load’. How much ‘load’ a house can put on the grid is equal to the cumulative watts of power of all electricity-consuming devices operating simultaneously. Some devices require a lot of power, which is the amount of electricity delivered over time. A high power-drawing appliance uses more electricity in a shorter period than a low-power drawing appliance. You will often notice that a home’s main breaker will ‘trip’ when it cannot handle the load of all the lights and appliances running simultaneously.

With water, you can install a water-conserving shower head to limit the amount of water per second that can come out of the tap, so you are using less water in the same three-minute shower. Electrical appliances do not work the same way usually. You need to buy energy efficient appliances to control how much power they are using. But overall, some types of home uses have the capacity to draw a lot more electricity from the grid than others. The biggest and newest is the home-charging of electric vehicles.

Matching Demand
Remember, grid operators predict electricity demand in advance. Forward market participants bid to supply that demand. Electricity is a commodity – there is no difference in the electricity that comes out of the socket whether it has been produced by a windmill or a coal-fired plant – so the cheapest bids are awarded a contract first. Suppliers need to predict how much fuel they will need to be able to generate the electricity they have committed to provide in future contracts – and they need to know at what price they can afford to sell it and still make a profit. But renewables, of course, do not need fuel so they do not have that issue.

Traditional power plants experience some financial risk in the price of coal or gas and they use different strategies to mitigate that risk throughout the year so that they have enough supply of their fuel around to satisfy the bids they will win. They need to make sure they earn enough to pay for the cost of the fuel as well as the cost of the plant itself, which is typically spread out over many decades.

It is getting more and more difficult to get those plants financed now that renewable energy is increasing in aggregate generation capacity. But as long as electricity demand grows as well as renewable energy generation capacity, we need either to ensure that we have back-up capacity that can be ramped-up quickly or we need to find a way to ‘flatten the curve’ of demand at times of the day when the gap between the drop in renewable generation and growth in demand outstrips predictions.

This mismatch between peak demand and drops in supply of renewables has a few underlying reasons. One of the classic problems is illustrated by California’s famous duck curve. In the late afternoon, the sun begins to set and (home) solar starts to produce less just at a time when people arrive home from work and start to turn on lights and plug in EVs. If demand peaks during that time, California needs to quickly ramp up generation. If the sun is going down and they are not able to turn to big solar farms, they need to compensate with other forms of quick-ramping energy generation.

Wind often contributes to another variability problem – oversupply. When the wind blows harder than expected and the demand is slightly less than expected, grid operators can be confronted with too much electricity. They will ask generators to curtail production and if necessary, they will even pay big industrial consumers to use electricity. This is called negative pricing and it happened in a big way last May 29, 2019 in the United Kingdom.

Germany has a lot of wind farms in the north of the country. On a regular basis these days they must take exceptional measures to deal with oversupply. The wind farms in the north often produce more than the high voltage lines have the capacity to carry down to the high-demand areas in the southern part of the country. Germany has even paid retail consumers to consume energy when there was an oversupply. They often turn to the Netherlands to deliver oversupply, but now even sometimes the Netherlands cannot handle the free electricity, so they turn to other neighbours. Markets have a really hard time understanding how to predict the performance of electricity and energy investments in that kind of volatility.

Renewable energy generators do not have any financial risk in the cost of the wind or the sun – that is free – but their risk is in how the wind and sun will affect their ability to generate. If they do not provide the power they promised, they will pay a penalty for the power that they do not provide. But strangely, over the course of a year, for a large wind farm, they can still make a profit if they reliably bid to provide a little more electricity than they predict they will be able to provide. It is like driving your car a little over the speed limit. You do not always get a ticket and on the whole, if you are making more money by arriving at your destination earlier, it might make financial sense to speed enough that you can afford the tickets for the amount of times you get caught.

So, there are a lot of ways advance predictions can go wrong – and remember – on the day of delivery itself, there needs to be a balance between aggregate supply and demand. If that balance is tipping one way or the other, the grid operator responds in the retail ‘balancing’ market.

Read the second part of Professor Ketter’s three-part series on the modern smart grid in the next issue of Future Grid Solutions