How European electricity markets work (part 1)

On the 9th and 10th of June, 2021, Nordpool organised a (virtual) course on electricity markets. The timing was quite fortuitous as it coincided with my visit to KTH (Royal Institute of Technology) in Stockholm, where I am planning to look into electricity markets as well.

Now, the course was only two days long, but it was jam-packed with interesting sessions. This blog post summarises my key takeaways from the course (and may as such not be totally reflective of what the speakers said or intended).

1. All you wanted to ask about the Nord Pool power market model (Haakon Leknes)

The introductory session was by Haakon, which was full of information and catchy sound bites including:

1. The discovery of oil saved the (Nordic) whales (due, in part, to the historical practice of burning whale oil to power lamps),
2. Oil and gas are renewables too but with very long regeneration cycles,
3. Green is good, as opposed to greed is good, in reference to decisions by the Norwegian sovereign fund and EIB (European Investment Bank) to drop / cease funding fossil fuel projects!
4. Once in a century storm, twice in a decade; on the subject of the recent black-outs in Texas.
5. Nord Pool, a happy marriage between engineers and economists, in reference to the company that provides electricity trading, clearing and settlement services, and operates in 16 European countries.

Haakon also discussed the basics of the electricity markets and their history, besides pointing out this interview of everyone's favourite nonagenarian Sir David Attenborough by the right wing bogey-girl Greta Thunberg.

Key take-aways

1. Covid-19 has led to the largest decline in emissions in recorded history in 2020— of course, this has been followed by a huge rebound in 2021 so far.
2. IEA, IMF and other relevant bodies project greater job creation (gains in electricity and biogas, losses in oil and gas — overall jobs in energy will grow from 40 million to 50 million) and lower mortality due to renewables.
3. 3 key costs in electricity delivery: grid, taxes and power price; these vary wildly in space (i.e. country to country) and time (from year to year).
4. Hydrology is a key influencer in electricity prices in the Nordics — this is very different from the situation in Belgium where the most important features are typically load on the grid and the situation in France.
5. Financial markets(forwards / futures) are based on financial settlement as opposed to physical delivery of electricity in the remaining markets. These other markets include day-ahead market, the intraday and the intra-hour markets.

A simplified view of the electricity markets

2. The role of the Transmission System Operator (TSO) (Hans Aarhus)

This talk focused on the TSOs’ responsibilities, regulation and reserve markets. The role of the TSO is primarily to ensure sustainable long and short term energy balance in the grid. To do this, Hans introduced the audience to how Statnett (the Norwegian TSO) reconciles the laws of physics with markets. It guarantees operational safety using something known as N-1 criteria (which simply means that the system should continue operating reliably even if a single cog in the system such as a large generator fails). It was also good to note that Statnett, the Norwegian TSO, saw the future as electric. But perhaps it is just a catchy slogan?

This quirky video from the BBC provides a good overview of the system operator's responsibilities.

Key take-aways

1. Statnett owns / operates over 92% of the main grid in Norway, besides cables to Denmark and Netherlands, as well as 140 transformer stations (and is also part owner of Nordpool). For reference, the Swedish TSO is Svenska Kraftnat, Finnish one is Fingrid, and Danish is Energinet.dk
2. The peak load in the Nordic system is 69 GW, with an installed capacity of 96 GW. Likewise, in Norway, the peak load is 24 GW with wind power contributing >2 GW.
3. Power production is distributed and diversified across the Nordics, with hydropower concentrated in Norway and Northern Sweden and Finland — on the other hand, Denmark as well as Southern Finland and Sweden have wind, thermal and nuclear plants. This can lead to a bottleneck since the different sources have different characteristics.
4. The entire Nordic region is divided into 12 individual bidding areas for physical power. A single bidding zone is a geographical area where there are no grid constraints on internal power flows, as determined by the TSO. Norway thus has 5 zones, Sweden has 4 and Denmark has 2. Finland must do with only one (just like most of the other European countries).
5. Since power cannot be stored on a large scale (economically), it must be produced and consumed at the same time. However, imbalances between supply and demand can be caused by wrong weather prognosis (e.g. temperature etc.), loss of consumption and production, or loss of an electrical component such as a grid interconnector. It is the TSO's responsibility to ensure this does not lead to grid blackouts.
6. The TSO uses three types of reserves (FCR-N/D aka primary reserve, aFRR aka secondary reserve, and mFRR/RR aka tertiary reserve) to ensure system frequency stays as close to 50 Hz as possible. The TSO contracts these reserves, and activates them as the situation demands it (600–700 MW upwards and downwards flexibility on the primary reserve; 400 MW on the secondary reserve). The balance on the system also determines the imbalance fees payable by balance responsible parties. mFRR services (tertiary reserves) were historically activated through calls, but this is being (has been?) digitised now.
7. Case 1: renewable generation with wind. Wind power is developing faster than anticipated: the generation (in TWh/a) between 2018 and 2040 is now expected to grow 10x for Norway (3 to 31 TWh), 4.5x in Sweden (19 TWh to 75 TWh), 11x for Finland (3 TWh to 34 TWh) and 2x for Denmark (15 TWh to 33 TWh). Overall, wind generation is expected to go from 40 TWh to 173 TWh in 2040. This will add a lot more variability to the system.
8. Case 2: electrification of demand. Full electrification of transport in Norway will increase electricity demand by ±50 TWh, but reduce primary energy use by ±80 TWh, alongside a dramatic fall in CO2 emissions.
9. Both of these cases will make the grid operator's job much harder. Therefore, there have been pilots showcasing demand response in a variety of settings, including using electric vehicles (which Norway is a frontrunner in). Recently, a pilot demonstrated 1 MW activation for the mFRR reserve using 160 EV’s, sustained for 36 minutes, with a 55 second response time to full activation. Ramp-down rates can also be controlled.

3. European day-ahead markets (Hilde Rosenblad)

Hilde's talk focused on the day-ahead markets. But what is a day-ahead market? To answer this question, we need to step even further back and consider, 'what is a power exchange'? Well, it is a market place for electricity, where buyers and sellers gather together. The fundamental properties of electricity (that it has to be consumed and produced at the same time, besides limitations on grid infrastructure) make it a bit different than other commodities.

Key take-aways

1. So, how does Nord Pool fit in this? It is there to provide a liquid, efficient and secure power market to customers (both generators and consumers). In doing so, it becomes the counterparty for all trades, i.e. guarantees settlement and delivery. This way sellers are guaranteed to receive money for the electricity they sell, and buyers are guaranteed electricity for the money they pay. Practically, this is achieved by splitting some countries or TSO areas into two or more bidding zones, as mentioned in the previous section.
2. Connecting different Nordic countries results in a lot more stability: the Norwegian supply mix is dominated by hydro, Denmark is split almost evenly between thermal and wind, Sweden has a mix of hydro and nuclear (with some thermal and wind), while Finland has lots of thermal and nuclear, with some hydro. Therefore, in wet years, there is a lot of excess generation in Norway; while on windier days, Denmark is in a similar situation. It was funny (sad?) to note that none of the Nordics is betting big on solar generation (for good reason).
3. SDAC, single day-ahead coupling, is aimed at creating a single pan-European day-ahead electricity market. This day-ahead market represents 97% of total traded volume on Nord Pool.
4. The prices and cross-border capacity in the day-ahead market are set by a single common price coupling algorithm EUPHEMIA. It is run on a day-ahead level (gate closure time for next day is 12.00 CET), takes 12 minutes to run the optimisation (based on the branch and bound technique) and is designed to maximise overall social welfare, while considering network constraints and buyer/seller behaviour. It also ensures that least-cost generators are activated first. More details on EUPHEMIA here.
5. It is important to note that area prices, set by EUPHEMIA, are different from system prices. The system prices are simply a Nordic reference price calculated on a synthetic benchmark disregarding capacity constraints between bidding zones). This system price pops up in some places for financial settlements etc.
6. How do market participants send their energy buy or sell signals to the market? Apparently, in a few different ways. These include single hourly orders and block orders. Single hourly orders consist of 0.1 MWh resolution, apply to every hour of the day, require a buy or sell signal (both price and volume), and there can only be one order per bidding zone. There may be up to 200 price steps (between -500 and 3000 euros). Block orders solve some of the limitations in single hourly orders, such as when the generator is willing to produce variable amounts at different price points, or if there is a link between multiple time steps.
7. The day-ahead plans of generators and consumers do not always pan out (e.g. in case of an unexpectedly sunny or windy day etc.). In this case, SIDC (single intraday coupling), a successor to the XBID project, provides market participants the ability to continuously trade electricity (i.e. there is no gate closure time unlike with SDAC). Intraday trading can also be used for speculative purposes, unlike on the day-ahead markets.
8. Both SDAC and SIDC cover most European countries at this point (with even more to be connected in the near future). A few notable exceptions from both markets include Britain (Hello, Brexit), Switzerland (?) and the Balkans.

Bidding zones across Europe

Cross-Border Electricity Trading: Flow-Based Market Coupling (Matin Bagherpour)

The European electricity grid connects arguably the most countries in the world, or so I think as I listen to Matin Bagherpour's presentation on flow-based coupling. How deep does the rabbit hole of electricity markets goes? Very deep, it turns out. And we have only begun digging.

Key takeaways.

1. What is flow-based coupling and why is it important? To answer this question, we must step back and take a look instead at what is happening today. The grid's physical operation needs to be reconciled with the market bidding, and the delimitation of bidding zones is an attempt at capturing this constraint. The limits on bilateral exchanges between these zones are passed on to the electricity market which ensures that interconnection constraints are met (this is the available transfer capacity regime, as it is implemented today). On the other hand, flow-based coupling includes a better view of physical constraints and therefore provides better utilisation of grid resources (while remaining cognizant of its limits).
2. Electricity flows cannot be restricted by commercial agreements but follow the laws of physics (e.g. exports of electricity from one country to another will follow physics, not order books — this impacts the situation in other countries that are not participating in the trade in the first place — kind of like how climate change also impacts countries that are not responsible for the emissions). Flow-based coupling addresses these issues by including them in the model as well.
3. Such flow-based coupling does have the potential to lead to some seemingly non-intuitive flows. On a physical level, these take place to relieve congestions on the grid, but in practice they may appear as flows from high price zones to low price zones (therefore making them non-intuitive). In ordinary conditions, electricity should always flow from low price zones to high price zones, which makes intuitive sense.
4. According to Matin, this change can effect units in smaller market zones more than those in larger ones (does this put Belgian traders at risk?). In either case, this change has interesting implications for traders, which seems like an interesting topic for further research.

API — Enables trading, Oscar Egnell

And, with that, we turned our attention back to the intraday markets, where it is possible to trade algorithmically / programmatically. This is enabled via an API (duh!) which has seen increasing usage over the past few years. More specifically, API trades now routinely make up more than two thirds of the trades on intraday markets. Five years ago, this number was close to zero!

And with that, we were done with day 1!

Well, not really.

First, we had to pass a Kahoot quiz administered by Haakon, which I joined as colorfulPanda or some such ridiculous name. The questions were fun, and I managed to score a podium finish (#humblebrag)!

And with that, the session came to a close and it really was time for a well-deserved fika and a nap (the first session had started at 08.30 in the morning)!