Relying on interconnectors for security of supply carries risks

Last week Will Gardiner, chief executive of Drax Group, told the Guardian newspaper that Britain’s increasing reliance on electricity imports through the use of interconnectors is a threat to energy security, and will result in the country importing carbon emissions. He also said that interconnectors would fail to deliver the aims of the government’s industrial strategy.

With 3 GW of new-build interconnectors securing agreements in the 2018 T-4 capacity auction and up to another 18 GW forecast to come online by the mid-2020s, Britain will be significantly more dependent on interconnectors during system stress events.

That the CEO of a huge coal-fired power station should cast doubt on the use of interconnectors is hardly surprising – interconnector participation in recent capacity auctions contributed to lower prices which pushed Drax’s planned new gas plants out of the auctions. It is also not surprising that the CEO of Eleclink, which is developing one of the new interconnectors disagrees:

“We believe the argument that interconnectors import carbon emissions to be unfounded as the majority of interconnectors connect, or will connect, the UK to countries with a lower carbon intensity,”
– Steven Moore, CEO, ElecLink

The comments by Gardiner are the latest negative remarks from Drax on the subject of interconnectors, and coincide with a new report from Aurora Energy Research questioning the reliance on interconnectors for security of supply.

“The contribution of interconnectors to security of supply is unknown, and, to a significant extent, unknowable. Government has so far overlooked this inconvenient fact,”
– John Feddersen, CEO, Aurora Energy Research

Aurora’s analysis shows that interconnectors have often delivered less power than National Grid assumed they would when demand was at its highest, even undermining security of supply by exporting during peak demand periods. This is nothing new, in previous posts, I described how IFA, the 2 GW interconnector with France, often exports at times of high GB demand.

The contribution from interconnectors in times of system stress is uncertain

The ability of interconnectors (or any other form of supply) to deliver in the event of system stress is reflected in the de-rating factors assigned to each technology in the capacity auctions, which go on to determine the level of income earned by each asset. However, it is not straightforward to calculate de-rating factors for interconnectors given the very limited data on the historical performance of interconnectors during genuine system stress events.

Aurora’s analysis calls into question the use of long-term historical average flows in determining de-rating factors since being secure on average does not ensure security during a rare 1-in-5 year event. The report identified a number of risks, suggesting a more conservative approach should be taken in setting de-rating factors for interconnectors:

• Interconnector performance varies significantly from year to year in response to policy and market changes – for example, IFA’s contribution to GB security of supply during winter peaks has been anywhere between 20% and 80% since 2010. Interconnector imports during periods of peak demand in GB have consistently failed to match their de-ratings, falling short as much as 50% of the time from France and close to all of the time in the case of the East-West link to Ireland.
• Interconnectors can make a negative contribution to security of supply by exporting at times of high GB demand, something that is not currently captured in the de-rating methodology. The fact that interconnectors can export as well as import means the range of possible de-rating is from -100% to 100%, rather than having a minimum of 0% as for generation assets. The risk that interconnectors undermine system security by exporting at times of stress could increase in the future with the introduction of more generous capacity market remuneration in neighbouring markets, particularly since weather correlation means instances of system stress may well occur in interconnected markets at the same time.
• Interconnector dispatch based on half-hourly price differentials is difficult to forecast with any degree of certainty, particularly since policy and technology change occur faster than data can be collected on extreme stress events, which are rare (there has yet to be a stress event in GB since the introduction of the Capacity Market). There are also questions around the extent to which the limited available data are relevant for future stress events, particularly after the introduction of the new Irish Capacity Market, with its substantial penalties for non-delivery of electricity from GB to Ireland during system stress.
• Policy developments in GB and other European countries have the potential to fundamentally alter the underlying economics on which current de-rating factors are based, for example, the introduction of Capacity Markets in other European countries means that interconnectors could be “over-committed” in two different markets. The 500 MW East-West interconnector is de-rated at 59% in the UK and 46.9% in Ireland – if it is exactly meeting its obligations in Ireland by delivering 46.9% of total capacity, its contribution to GB supply will be negative: an outflow of 46.9% of total capacity, which is a substantial 105.9% (529.5MW) in deficit on its GB obligations. Differences in capacity market penalty regimes have the potential to distort interconnector behaviour during correlated stress events, while trade between Transmission System Operators in interconnected markets adds a further layer of uncertainty.
• Increased reliance on renewables exacerbates the impact of low-wind periods across Europe – plausible future scenarios involving faster-than-anticipated renewables build-out, correlated renewables output, and higher interconnection between countries with correlated demand all compromise security of supply in GB.
• Higher levels of interconnection call for lower de-ratings as the additional marginal unit of interconnection contributes less to security of supply. The existence of more interconnectors increases the likelihood of unexpected exports during periods of system tightness.
• The risks described above are not independent, increasing the uncertainty around the ability of interconnectors to deliver during stress events. In plausible scenarios combining low wind output, high demand, and a harmonised carbon price, interconnector flows could easily fall to zero, or become negative (ie exporting).

“With existing and contracted interconnector capacity, GB will be reliant on 7.4 GW of interconnectors to meet our peak demand of approximately 60 GW by 2021/22. Given the need for a surplus capacity margin during peak periods of approximately 5% — or around 3 GW — interconnector performance is already critical. If interconnectors fail to deliver power during peak demand in 2021/22, substantial load will need to be shed.”

Aurora assesses that current de-rating factors are at the upper end of potential future interconnector contributions during system stress events. In calculating the de-rating factors, National Grid analysed interconnector performance between 2010 and 2016 during the top 50% of peak demand periods (7am-7pm on business days between November and February). The probability that an interconnector was importing during these periods was used to set a lower bound on de-rating, with an upper-bound established by forward looking modelling of interconnector flows during tight capacity margins (<500 MW) and winter evening peak periods.

De-rating factors for thermal technologies are based primarily on technical availability and tend to be fairly constant over time since the probability of outages does not change significantly from one year to the next. Interconnector de-ratings on the other hand can be much more variable given the volatility of the price differentials on which interconnector use is generally based – this can have a significant impact as the de-rating factors are calculated 4 years ahead of delivery.

It is also far from clear that examining the likelihood of imports based on price differentials during weekday peaks in the winter provides any useful information about the likelihood of imports during a period of system stress. The choice of time period over which these probabilities is analysed is also arbitrary, further undermining the usefulness of the results.

Behaviour of TSOs may also threaten the use of interconnectors in times of system stress

The price difference between the interconnected markets is the main driver of interconnector use, with electricity flowing from the lower priced to the higher priced market, however transmission system operators (“TSOs”) also engage in interconnector trading after gate closure, based on bilateral agreements whose terms are not public.

Weather correlation between GB and its neighbours is fairly high, meaning that periods of high demand will often occur at the same time in nearby, interconnected markets. If those markets have a higher level of temperature sensitivity than GB, as is the case with France, demand would rise faster in those markets, leading to pressure for the interconnectors to switch into export mode.

Although TSOs are not generally responsible for security of supply, they are responsible for ensuring their systems are balanced, so when demand rises, it is the responsibility of the TSO to call on available capacity to meet that demand. It is far from clear that any TSO would allow exports to occur when its own supply and demand balance is tight.

“The reasons for this trading are opaque and it is therefore difficult to identify how the TSOs at either end would trade in the case of a system stress event. Absent past data, it is conceivable that during a correlated system stress event, neither TSO would be willing to export power and flows would fall to zero,”
– Aurora Energy Research

These suspicions have not been tested in a stress event, however the behaviours of TSOs within the flow-based market coupling (“FBMC”) region illustrates that TSOs will seek to protect their own markets even if their actions are contrary to agreed market norms (or even EU law). As the EU moves towards intraday market coupling, these factors may have even greater impact on interconnector use, potentially reducing their utilisation even in normal market conditions, as is currently being observed with day-ahead FBMC. Although Britain will not be part of intraday coupling before the next decade, it is unlikely to be immune from the effects of such behaviours.
Despite the doubts over security of supply, new interconnector projects continue to move forwards. In the past two weeks, the first cable that will form part of the North Sea Link interconnector with Norway was winched ashore at Browns Farm, Cambois, and the public consultation on the Greenlink project with Ireland opened.

However, project developers should be aware of a changing tone about the contribution of interconnectors to system security. This change is welcome, and policy-makers should take note – assumptions to date that interconnectors are unambiguously positive for security of supply are not supported by the data, and therefore a conservative approach should be taken in assessing their contribution.