Issue Number: 114   November 2014

Will the lights stay on in the UK this Winter?


Figure 1 - Didcot Power Station B Fire

Generation capacity in the UK has fallen in recent years, with some of the issues surrounding the fall covered in this earlier MZINE article[1]. The recent fire at Didcot Power Station B [2], which is the latest in a series of disruptions to the UK's electricity capacity, has highlighted the fall in capacity and reignited the debate on the UK's ability to meet electricity demand over the winter this year. The widespread media coverage devoted to the UK's ability to avoid blackouts led Ed Davey, Energy Secretary, to issue an unequivocal assurance that "there will not be blackouts. Period." In reality, much of the discussion following Didcot Power Station's second dramatic energy-related event this year (the first being the demolition of its coal station's cooling towers) acted as a precursor to the release of National Grid's (NG) Winter Outlook Report. The report notes that despite a corresponding drop in demand the capacity margin for the winter is at its lowest since 2005/2006. Nonetheless, given NG's contingency planning, it seems fairly unlikely the UK will suffer blackouts or any other serious disruptions. However, the demand side management that NG will employ this winter and beyond does not appear to showcase a market that seems to be perennially in transition functioning at an optimum level.


Figure 2 - Didcot Power Station A's cooling towers mid-demolition

Capacity Margin Winter 2014/2015

In its report, National Grid states that de-rated generation capacity, which takes into account availability and historic performance, is 58.2GW. The margin between the Average Cold Spell (ACS) demand [3] (55GW) plus reserve (0.9GW) and de-rated generation availability is 2.3GW (or 4.1%). This figure is the lowest margin since 2005/2006.


Figure 3 - Capacity vs Demand

Unsurprisingly, the calculation of the figures has been criticised. Energy analysts have suggested that National Grid's figures are dependent on the successful and timely return of Heysham and Hartlepool to service, noting that the record of bringing nuclear plants back online on time are questionable at best. [4] This point is actually addressed in the report with NG admitting to some uncertainty and that uncertainty is certainly a factor in the balancing services employed, discussed in more detail later. A report questions the assumption that 90% of nuclear capacity will be working at the point of maximum demand is surprisingly high given the previous year's assumption of 84% and Ofgem's assumption of 81%. In addition, it argues that if mid-December is taken as peak demand, then around 1.6GW of power will not be online due to an outage at Heysham 1, the 20% restriction on Hartlepool and Heysham as units come online and precautionary power reductions of 20% at Hinkley Point and Hunterston.[5] The first two reductions will be true in mid-December, however, at the time of writing the precautionary power reductions are currently only in place at Hunterston's second unit, and in fact, by mid-December may not be in place at all. This leaves a shortfall of around 1.1GW which does certainly exceed the 10% NG suggests, but is not much larger. This leaves little room for manoeuver and there is certainly a point that the availability of nuclear power may have been over estimated but not necessarily by significant margins. [6] Additionally, it should be noted that the 90% figure is likely used for the whole of the winter period used as opposed to certain points. In fact, the graph below does show a tight capacity on the week commencing the 8th December.


Figure 4 - Assumed Generation Availability for Winter 2014/15

Interestingly, whilst many headlines have pointed to the margin of 2.3GW (4.1%), the lowest since 2005/2006, many have neglected to mention the other method of analysis used in the report. In Ofgem's Electricity Capacity Assessment Report 2014 it notes that the industry is beginning to move away from the de-rated margins and towards Loss of Load Expectation (LOLE). This is because margin analysis is becoming outdated due to the increasing use of renewable energy, in particular wind. LOLE is a method that presents the expected risks associated with the actual characteristics of the GB market. It provides a figure relating to the number of hours/periods per annum in which, over the long-term, it is statistically expected that supply will not meet demand in the absence of intervention by the System Operator. In most cases, loss of load would be managed with minimal impact on consumers. LOLE is intended to reflect the economically efficient level of capacity. The target is not zero, thus appreciating the expense of maintaining enough capacity to meet all potential demands in all situations. If we take the LOLE figure at face value, at 1.6 hours, it is well within the reliability standard set of 3 hours. This helps to place the figure within some context and shows a fairly minimal risk of disruptions over the winter. This is before the balancing services, detailed below, are even taken into account.

'1 in 20' Demand

Whilst the main demand figures to come out of the report are based on the Average Cold Spell (ACS), there is also an Arduous Forecast, which has a '1 in 20 demand' - ie a 5% chance that demand will reach this level on average. When compared against assumed generation, NG would be unable to meet the Basic Reserve requirement on two weeks (with 750MW imports from Europe). However, with the European interconnectors at full import there would be a 2.8% margin for those two weeks. It should be noted that full European interconnector imports cannot really be relied upon on a long-term basis, however as a mitigating action sizeable imports from the interconnectors are certainly possible on a shorter basis.


Contingencies

Whilst the graph above might paint a slightly worrying picture, it should be noted that even the ACS demand (the pink line) only has a 50% chance of occurring. In any event, Ed Davey has stated that "we have extra contingencies on top of the caution, and extra contingencies on top of the contingencies" that should ensure supply can meet demand even if plant failures or other events restrict generation capacity. NG's New Balancing Services add another 1.1GW to capacity (increasing the margin to 6.1% or reducing the LOLE to 0.6 hours). Its balancing services include Demand Side Balancing Reserve (DSBR) and Supplemental Balancing Reserve (SBR). In the case of the former if there was a sudden shortage of power, big industrial consumers would be paid to switch on emergency backup generators and produce their own power. 319MW of total DSBR volume has been secured. SBR brings plants that would otherwise have been mothballed back into use. Approximately 700MW is coming from three power stations; Littlebrook, Rye House CCGT and Peterhead CCGT. Although this capacity increases the margin and lowers the LOLE it is not available in the market, but instead held in reserve to be dispatched as a last resort in the event that demand cannot be met. Additionally, NG has 'Maximum Generation' contracts in place, providing potential access to up to 740MW of extra generation. However, NG does not take this into account in any of its analysis as it is a non-firm emergency service and with generation operating under these conditions there is normally a reduced reactive power capability. In a real emergency NG might use voltage reduction (or brownouts) to reduce demand.

Conclusion

The concerns about the UK's ability to keep the lights on this winter have been fairly vocal. However, ironically, these concerns may well be louder next year as capacity margins continue to remain tight until the end of the decade when new builds come online. All but one of Ofgem's forecasts expect LOLE to rise above the reliability of standard of 3 next year. The new balancing services were only meant to be a pilot this year. Next winter, DSBR and SBR volume contracted will likely rise to around 1,800MW. [7]

Whilst the gap between supply and demand appears fairly tight, particularly during mid-December, barring a set of extremely unlikely circumstances, NG appears to have prepared for most eventualities with its various contingencies. Therefore, it seems fairly unlikely that any serious disruption will occur, even brownouts, however, what does seem likely is that a similar discussion over whether there is enough supply to meet demand will probably happen most years for the rest of the decade.


[1] This article focuses on the issues surrounding the fall in EU gas demand in recent years, the particular points of note relevant to this article are those concerning the effect low coal prices have had on capacity, including a lack of investment in power generation infrastructure, the mothballing of uneconomic gas plants and a tighter future margin to support renewable energy sources as old coal plants close.

[2] Didcot B is expected to run at around three-quarters of its total 1.4GW capacity.

[3] It should be noted that the ACS forecast is demand in conditions with a 50% chance of being exceeded. It is higher than the central (or normal) forecast which takes into account historical weather patterns. The ACS forecast is used for many of the margin figures.

[4] http://www.reuters.com/article/2014/10/28/britain-electricity-nationalgrid-idUSL5N0SN24120141028

[5] http://www.theecologist.org/News/news_analysis/2616920/uk_faces_serious_winter_blackout_ risk_national_grids_rosy_nuclear_forecast_fails_reality_test.html

[6] The article also suggests that the baseline Nuclear capacity used in the report is higher (by around 0.1GW) than the figure used in Ofgem's 2014 report.

[7] http://www2.nationalgrid.com/mediacentral/uk-press-releases/2014/national-grid-to-contract-for-new-balancing-services/


November 2014 MZINE