Demand might outstrip supply this winter in Great Britain. National Grid ESO published a very tight Winter Outlook - which includes a chance of blackouts and loss of load.
The new Demand Flexibility Service has just started, in a bid to reduce demand from domestic and industrial consumers at times when generation may struggle to meet it. In all, it could be a very challenging few months.
In this article, we explore:
- Some of the key terms used to describe load shedding. These include: loss of load; loss of load probability; de-rated margin; value of lost load; and reserve scarcity price.
- The past impact of loss of load probabilities on power prices.
- National Grid ESO’s position on taking very expensive balancing actions this winter - and what this might mean for prices in different markets.
Loss of load
Electricity supply to customers is cut off when demand exceeds supply. This is called a loss of load event.
Loss of load probability (LoLP) measures the likelihood of a shortfall in generation capacity relative to demand, in a given settlement period. The closer the loss of load probability is to 1, the greater the risk of customer supply interruption (or load shedding).
Figure 1, below, shows the average loss of load probability for each hour of the day. The ESO measures loss of load probability at five different times of the day. It is measured at midday (for each settlement period), and at various intervals relative to each settlement period, from eight hours before down to one hour before.
Note: We have measured probability as a percentage. 0.5% is actually equivalent to a probability of 0.005.
- 99.6% of the time, loss of load probability is below 0.5% (or a probability of 0.005). It has only exceeded 0.5% for 200 hours since March 2018.
- On average, the probability increases over the peak period of the day. Higher demand between 4pm and 8pm means that margins are often narrower.
- Loss of load probability tends to drop as we get closer to actual delivery. This is because the market takes actions to increase margins in the intervening period.
De-rated margin
De-rated margin indicates the balance of supply and demand. It is the difference between:
- The forecasted available generation capacity from conventional sources (i.e. thermal) and wind; and
- the capacity requirement (i.e. forecasted demand), which includes interconnector exports and additional headroom for managing the system.
Figure 2, below, shows the relationship between de-rated margin and loss of load probability.
- De-rated margin is the sole input for calculating loss of load probability via Elexon’s Static LoLP Function Method, shown above in pink. Lower (or negative) values for de-rated margin correspond to a high loss of load probability.
- A de-rated margin of -1,500 MW (equivalent to a 3% shortfall of 2022/23 winter peak demand) would result in a loss of load probability of 1.
- The ESO uses de-rated margin to trigger system warnings e.g. capacity market notices.
Value of lost load
The value of lost load (VoLL) is supposed to indicate what consumers are willing to pay to avoid interruption to their electricity supply. It considers several customer types, including domestic, commercial & industrial.
There are two definitions of value of lost load:
- The true value of lost load is £16,940-£17,500/MWh. It was set in a 2013 white paper for Ofgem, prepared by consultancy London Economics.
- The administrative value of lost load is £6,000/MWh. This was set in Balancing Settlement Code in November 2018.
This administrative value of lost load matters because it sets the price at which demand-shedding actions enter the calculation of the imbalance price. It is also used to set the reserve scarcity price - the price at which some reserve actions enter the calculation of the imbalance price. The reserve scarcity price is the product of loss of load probability and this administrative value of lost load. These factors act to drive prices higher when margins are low, and ultimately secure more generation on the grid.
Does a high loss of load probability indicate extreme prices?
Both loss of load probability and market prices spike when demand begins to outstrip supply - but how are these things related? Figure 3, below, shows the average price across various power markets for high loss of load probability (> 1%, or a probability of 0.01) and low loss of load probability (< 1%).
- Across all the above markets, prices are typically higher when loss of load probability is above 1%.
- This trend is most evident in the Balancing Mechanism - likely because of the real-time nature of this market.
- A low loss of load probability does not necessarily mean low prices, as other factors (e.g. gas prices) can drive power prices in the absence of tight margins.
What does this mean for Winter 2022/2023?
National Grid ESO’s Winter Outlook discussed several scenarios for the winter ahead. It mentioned the potential for scheduled, controlled load shedding. You can read Modo’s take on National Grid ESO’s Winter Outlook here, along with what it means for batteries.
National Grid ESO also recently published a memo confirming their policy to preserve demand. But why does this matter?
- This memo refers to taking “commercial and market actions” that exceed the £6,000/MWh value of lost load.
- This actually happened in July 2022, when the ESO took an interconnector action priced at £9,725/MWh.
- This set the precedent for taking “actions above VoLL”.
- The ESO has also indicated that the £6,000/MWh value of lost load should be updated - to more accurately reflect the cost of keeping the lights on.
Due to ongoing low French nuclear availability, increased interconnector exports are likely (more here). In the event of tight margins in GB, the ESO may have to pay high prices (potentially in excess of VoLL) to flip interconnector trading - as we saw in summer.
The ESO has also indicated an increased likelihood of high Balancing Mechanism prices in response to tight margins. Despite batteries’ historically low participation in the Balancing Mechanism, this could present significant value for storage.
Key Takeaways
- Loss of load occurs when there isn’t enough supply to meet demand.
- National Grid ESO provides a view of the likelihood of this happening ahead of time, via forecasts for de-rated margin and the loss of load probability.
- While historically, periods of time with a high loss of load probability
- Historically in Great Britain, high loss of load probability events are rare. 99.6% of the time, loss of load probability is < 0.5% (or a probability of 0.005).
- Loss of load probabilities above 1% (or a probability of 0.01) signal to the markets that additional supply is needed. We usually see higher market prices as a result, particularly in those markets which run closer to real-time.
- There are two ways to value loss of load: the true value of loss of load (between £16,940-£17,500/MWh which reflects consumer inconvenience) and the administrative value of lost load set by Elexon at £6,000/MWh.
- Ahead of this winter, the ESO published a memo discussing its position on taking actions above £6,000/MWh this winter. In the face of a pan-European power market crisis, we could see high costs to resolve tight margins via interconnector and the Balancing Mechanism.
- Increased likelihood of high loss of load probability this winter suggests high prices in wholesale markets - in turn signaling lucrative opportunities for storage.