The case for changing the GB Electricity Reliability Standard

The Great Britain electricity system is set to undergo significant change in the coming years as the system evolves to achieve decarbonisation targets.

As demand changes in volume and nature, and intermittent renewables start to dominate our electricity generation, the tools needed to ensure security of supply will need to change relative to today.  

A key part of the framework for ensuring there is enough capacity in GB currently is the Reliability Standard. Sufficient capacity should be purchased in the GB capacity market to ensure that only three hours of loss of load is to be expected on average each year – a Loss of Load Expectation (LOLE) of three hours. 

Frontier Economics and LCP Delta were commissioned by the Department for Energy Security and Net Zero (DESNZ) to assess whether the current choice of reliability standard metric remains appropriate as GB transitions to a fully decarbonised electricity system. 

The literature on reliability standards is broad, and there are a range of different metrics used in different countries across the world. However, a LOLE standard is fairly commonly applied in Europe and parts of the US. While we found that there is a clear recognition in the literature that reliability standards may need to evolve, there is limited guidance as to what should precisely be done.  

In future, the nature of system stress events is expected to become more complex. Historically, system stress was largely a function of uncorrelated outages of large generation plants or interconnectors occurring at peak. However, as the system evolves, with a greater reliance on weather dependent renewables in particular, the system may experience larger reductions in supply than historically observed due to correlated renewables output; and longer periods of reduced supply due to risks of extended periods of low wind, i.e. the so-called “dunkelflaute”. 

As a result, targeting a single LOLE metric may no longer be appropriate. As system stress events change, other dimensions of stress which are not captured by LOLE could increase or become more extreme. This could include an increase in the volume of energy unserved or the duration of stress events. For any given LOLE, the risks of very extreme events could also increase. Therefore, for a given value of LOLE the costs due to loss of load faced by end-use consumers (i.e. value of lost load) could be quite different, as illustrated below. 

Figure 1	LOLE and different dimensions of stress events

A single superior metric was not identified in the literature. However, there are clear indications that a combination of metrics may be beneficial. There are some limited international examples, e.g. Belgium, where in addition to a LOLE measure, additional capacity is procured to ensure that very rare but more extreme events are limited.  Further development of the theoretical framework is required to inform how the target levels of multiple metrics could be set, i.e. at levels that strike an optimal balance between the costs that consumers face from the different dimensions of outages and the cost of incremental additions to the capacity mix to ensure security of supply.  

Ultimately, the extent to which these findings are applicable to GB will be a function of how our system is expected to evolve as part of the energy transition. Empirical analysis will therefore be needed to determine whether there is a need to adopt multiple metrics and to decide on the optimal choice of metrics. Analysis by LCP Delta as part of this study demonstrated some support for an additional metric in addition to LOLE. However, further work is needed to determine more fully what an ideal combination of metrics should look like. In its second REMA consultation, DESNZ indicated that it is continuing to evaluate the case for change. 

Click here to read the full report: Exploring Reliability Standard Metrics in a Net Zero Transition