The power market design column – The vulnerability paradox, the double vulnerability paradox and the new vulnerability paradox

The vulnerability paradox can be defined as follows: if supply of a good in a country is less vulnerable, a supply disruption will hit that country all the harder. In 1994 a Dutch study?Stroomloos?introduced the double vulnerability paradox for electricity supply: not only the lower vulnerability of electricity supply, but also the increased dependence on electricity leads to greater social vulnerability in the event of a disruption of electricity supply.

Monitoring reliability is important

We all know how important reliable electricity supply is. And the trend of increased dependence on electricity as mentioned in the study of 1994 has only continued since then. Gas supply, telecommunications, internet, sewage system, traffic control etcetera, it all depends on reliable power supply. This trend will certainly continue in the next decades as electrification is a key element of a successful transition towards a decarbonized energy supply.

It thus does make sense to monitor the reliability of our electricity supply system especially as the industry is liberalized and market forces are assumed to take care of reliability, at least to deliver an adequate amount of generation and storage capacity to meet demand. But reliability monitoring is not a panacea. On the contrary.

First of all, what indicator to use and especially at which value? The Loss of Load Expectation (LOLE) is useful indicator and is often set at 4 hours/year. This value could be based on Cost of New Entry (CONE) of 50 000 Euro/MW per year and a Value of Lost Load (VoLL) of 12 500 Euro/MWh. In that case a system with a LOLE of 4 is assumed to be at the right level of reliability. The CONE can be assessed rather well but is impossible to have an accurate guess for the VoLL. So nobody knows whether a LOLE of 4 is good.?

Secondly, the monitor will see a declining LOLE looking into the future if demand is supposed to grow, older power plants are assumed to be taken out of operation and investments in the future in new generation capacity are, almost by definition, uncertain. It is therefore completely normal that a reliability monitor will result in LOLE values above 4 from a certain year in the future onwards. So the monitor will note a declining level of reliability. But is that a problem? Or should one conclude that the system is moving towards an optimum.?

in?2019?I introduced a new vulnerability paradox:??the more important monitoring of security of supply becomes, the less useful it is or the greater the chance will be that the monitor will be used incorrectly.?

Reliability standard is not met for first time in 20 years

In the Netherlands, the TSO (TenneT) is charged to yearly monitor the level of security of supply. The first monitor was published in 2003. Almost 20 years ago! The conclusion at the time was: "Based on the results of the monitoring, we currently see no reason for the government to take measures for the future security of supply of the electricity production sector on the basis of the data currently available." Since then, the conclusions were almost very similar. Until now. Last week TenneT published its?monitor 2022?(in Dutch).??This monitor looks forward up to 2030.?

Now for the first time a different conclusion is drawn. TenneT writes: “The decline of resource adequacy for the first time leads to an exceedance of the reliability standard in a central scenario for 2030.” In particular the LOLE reaches 4,5 hour (and thus exceeds the 4 hours value) in the main scenario in the year 2030. TenneT is a bit cautious, but the last line of the advice is rather clear: "It is therefore important that action will be taken to ensure that the market is able to timely realize adequate demand side and/or supply side flexibility to ensure resource adequacy."?

So let us look at this monitor from a few different angles.

Too high or too low reliability?

The Dutch newspapers and sites picked up the news directly. “TenneT warns for electricity shortages.” But as explained before, it is totally unsurprising that the monitor shows a declining level of security of supply. Another conclusion based on the same results could be that the level of security of supply in the Netherlands is too high up to 2029 and too low in 2030, as a LOLE of 4 hours per year is an optimal value.?

A reliability standard of just an indicator?

The two main indicators that TenneT uses are the LOLE and the EENS (Expected Energy not Served). As mentioned, TenneT uses a value of 4 hours per year for the LOLE. It is said to be the norm or the reliability standard for the Netherlands. But is that correct? If I am not mistaken there is no law or regulatory decision in the Netherlands hat states that the level of security of supply should be kept at a level with a LOLE below 4 hours per year. At least TenneT does not refer to any such regulation. As explained before it is a guess of what the optimal level of security of supply could be.?

Also interesting is that TenneT does not refer to a decision taken last year by ACM, the Dutch regulatory authority, in which the VoLL for the Netherlands is set at 68 887 €/MWh. This decision is the follow-up of Article 11 of EU Regulation 2019/943 that requires Member States to determine a single estimate of the VoLL for their territory.?

This value only has a direct meaning if a Member State has a capacity mechanism. The Netherlands does not have such mechanism. However, if the Netherlands would want to apply for a capacity mechanism it has to define a reliability standard that is based on a VoLL that is now set at 68 887 Euro/MWh.?If we assume a CONE of 50 000 Euro/MW per year and use a VoLL of?68 887 Euro/MWh, the value for the LOLE becomes only 44 minutes! Suddenly everything changes. A much more reliable system would be needed and TenneT would have to conclude that security of supply is in immediate danger!

If the Netherlands would indeed want to implement a capacity mechanism, like a capacity market or a strategic reserve, and if it then would use a reliability standard based on a VoLL of 68 887 Euro/MWh it should not have a difficult time to explain to the EU Commission that such mechanism is urgently needed. However, if it then would have to dimension that mechanism in such way that it aims to ensure a LOLE of 44 minutes, it would become an expensive mechanism.?

It is understandable that TenneT did not use a LOLE of 44 minutes. The VoLL value of 68 887 Euro/MWh is a surprisingly high value. TenneT did apply a LOLE of 4 hours during all these 20 years and it is line with practices in neigbouring countries. At the same time, it is incorrect to write that 4 hours is the Dutch reliability standard.

Results depend on assumptions

TenneT uses multiple scenario’s and climate years for its analysis and correctly points out that the results will depend on various assumptions. For example, TenneT underlines that the results strongly change if the assumed amount of installed, controllable (dispatchable) capacity is changed.?

It also means that it is easy to question the outcome of the monitor by criticizing the assumptions. TenneT assumes 15.8 GW installed off-shore wind capacity in 2030 in the current policy scenario and 17.5 GW in the high ambition scenario. However current?policy?is aiming at 21 GW in 2030. So, it seems that TenneT is pessimistic.?

For batteries, TenneT assumes 10.3 GW respectively 13.3 GW for the two scenario’s current policy and high ambition. This is based on the requests for connecting batteries to the grid that TenneT did receive. It is however well possible that the high amount of requested grid connection capacity for batteries, is inflated by the current lack of grid capacity. Battery developers may be incentivized to ask as much as possible in the hope to be allocated at least some connection capacity. If that is the case, TenneT would be optimistic.?

Anyhow, forecasting how the market will develop is notoriously difficult even for those technologies for which national policies are being developed. Those policies may change but also lack of grid capacity might have a big impact on the actual developments. In its monitor, TenneT doesn’t dwell on the challenge to expand the grid in due time. But it is a majopr issue as the Dutch grid is already congested which is causing problems for wind and PV developers to connect to the grid.

Controllable, weather dependent and dispatchable capacity

TenneT distinguishes between controllable and non-controllable generation capacity. Wind and PV is in the latter category and all other technologies including biomass and nuclear are considered controllable. This strict separation is debatable.?

Actually, wind and PV is very well controllable, obviously within the limitations of the actual wind and sun conditions, But at least ramping down is always possible and normally at much higher rates than fossil fueled generation. Even ramping up is no problem from a technical point of view. It requires that the plant is pro-actively curtailed, which normally makes little sense from an economical point of view.?

Instead of non-controllable, one could also label wind and PV as weather-dependent, although even that is not fully correct, as the other technologies are also weather dependent, albeit to a much lesser extent. The maximum output of a gas turbine depends on the ambient temperature, high temperatures may also result in lower capacities because of cooling water restrictions and a long period of drought can impact the possibilities to ship coal over rivers.?

Maybe the best way to categorize wind and PV is to label them as non-dispatchable. Or we simply refrain from any categorization, as every technology has its own characteristics regarding availability, dispatchability and controllability. But OK, all these nuances do not matter much. It is obvious that a reliable power system requires a mix of technologies and relying solely on wind and PV is not an option.?

Interconnectors and a supergrid

In its summary, TenneT also explicitly mentions the importance of interconnectors: “Expanding the capacity of interconnectors with the United Kingdom and Scandinavia can also contribute to greater security of supply”. This is not a surprise as the contribution of interconnection capacity to reliability, or better put, as the cost-efficient contribution to reliability of interconnection capacity is well known. It is however somewhat remarkable that this statement is formulated in a such an open way, as TenneT, in cooperation with its neighbouring TSOs, is itself responsible to develop such interconnectors. In its assumptions TenneT only counts on an increase of the import capacity from Belgium. The values for the UK, Norway and Denmark in 2030 remain at the current values of 1.0, 0.7 and 0.7 GW respectively. The market would want to know how these numbers might change in the future. Uncertainty on the import and export capacity increases the risk for merchant investments in new generation and storage capacity. And an increasing risk means lower reliability.?

So instead of just stating that expanding the interconnection capacity can contribute to security of supply, it would be better if TenneT explains what could happen after 2030 and which possible hurdles for such expansion would need to be tackled. It is well possible that nothing spectacular will happen before 2030, but an off-shore, meshed grid will develop sooner or later. Recently the Belgian and Danish TSOs launched a?tender?for consultancy services for the development of the subsea cable route for a new interconnector project in the North Sea. It is called the Triton Link interconnector project which will connect the Belgium and Denmark through two energy islands in the North Sea.?Will TenneT seek to connect the Dutch high-voltage grid with such islands?

Another interesting option is the development of a European overlay grid or supergrid. Such a grid has the advantage that variations in output of wind and PV plants are less correlated over long distances. At the same time, it is questionable whether the possibility of such supergrid will be duly considered. The current approach, with the development of Ten Year Network Development Plans by ENTSO-E, is still mainly a bottom-up approach. This approach tends to result in projects to strengthen the existing national grids and the interconnection between these national grids. It would be interesting to know how to ensure that the possibility of a European supergrid will be considered. Does this require new EU rules?

Conclusion: What is the conclusion?

As mentioned before, the main conclusion of the monitor reads as follows:

“It is therefore important that action will be taken to ensure that the market is able to timely realize adequate demand side and/or supply side flexibility to ensure resource adequacy."?

I have explained that this conclusion is less obvious as it looks. The LOLE can be used as an indicator, but the correct value is difficult to quantify. Secondly it is not a reliability standard that needs to be maintained, and the calculated LOLE-values strongly depend on many uncertain assumptions.?

The wording of this conclusion is also interesting. TenneT does not write that market participants should take action although that is what one expects market participants to do if the market is getting tight. It is not explained in the monitor, but apparently TenneT expects that the market is not able to timely realize such investments. And TenneT may well be right. Especially investments in CO2-free, flexible generation capacity are extremely risky. I am not talking about normal market and price risks. But the business case for such investments is heavily impacted by governmental climate policy, especially as such policy is normally not stable. A good example are decisions to close nuclear capacity, which are then reversed some years later and the lifetime of the same plants is prolonged. Moreover the political and public acceptance of certain technologies like biomass, CCS and nuclear is very uncertain. And finally, we have experienced in 2022 that market interventions in times of high prices are not unlikely. Scarcity prices and scarcity revenues are however necessary for generators to be able to recover fixed costs.?

Overall, we may conclude that the conclusion of the monitor is correct. Actions need to be taken to ensure that the market is able to invest so that resource adequacy can be ensured. However, strictly speaking this conclusion cannot be derived from the monitoring results.

This is my 36th?column on power market design issues. The earlier columns covered the following topics:?EU power market reform,?EU market interventions,?review of the CACM Regulation,?market myths and price formation,?system support balancing,?Blackouts,?the importance of ACER,?Flexibility and foisonnement,?reliability and load shedding,?regulation of congestion income,?dynamic network tariffs,?energy communities,?scarcity pricing,?the Florence Forum,?active system management,?network planning & sector coupling,?off-shore assets,?intraday capacity hoarding and pricing,?interconnectors,?international comparison of market designs,?cross-border capacity calculation,?flexibility,?cross-border capacity,?electric time and unintended exchanges,?EU Network Codes,?price formation and zero marginal cost generation,?simplicity in the Clean Energy Package,?smart grids,?storage,?auto-generation,?balancing,?VoLL,?demand side response,?interconnectors?and the?Economist on market design.

Disclaimer: The views as expressed in this column do not necessarily reflect the views of Energie-Nederland

Paul Giesbertz

[email protected]