Balancing the risks of power blackouts with the costs of ensuring adequate supply

Recent years have seen numerous commentaries about the risks of power blackouts in the UK as electricity supply margins get tighter and tighter. However, what often gets left behind in these considerations is a basic question: what cost are we willing to pay to prevent blackouts?

This question becomes of particular interest in light of Scottish Power’s announcement to withdraw 2.4 GW of capacity from the market, in the form of Longannet coal-fired plant, in March 2016, whilst also abandoning plans for a gas-fired power plant at Cockenzie.


Looking back at the most recent 2014-15 winter period did indeed see a period of particularly tight supply in the market. Two days in mid-October 2014 saw the margin between supply and demand narrow appreciably as a result of relatively low levels of wind speeds coupled with the delayed return of coal-fired power plants; that had taken advantage of the summer months to carry out prolonged periods of maintenance.
This left only limited amounts of spare capacity in the market, resulting in power price peaks of à‚£262/MWh and à‚£266/MWh on the 14th and 16th October 2014 respectively, which compares to a market norm of approximately à‚£40-45/MWh for that time of year.

Shortly after this period, power supply units returned to service following their summer outages and by mid-winter 2014-15, the levels of electricity generating capacity at GB power stations had increased by a third from what they had been available in October.

This meant that across the whole winter period there was no repeat of the tight margin seen on those two days in October, with levels of available supply often being high against demand and with daily average power prices commonly averaging below à‚£40/MWh in January.

For all the criticisms of wind farms, from October 2014 to March 2015 the wind fleet saw load factors of approximately 34% of installed capacity as a whole; meanwhile CCGT plants saw estimated load factors of approximately 28% over the same period.

That meant that there was more unused capacity for generation at CCGT plants than at wind farms over the 2014 winter period. By comparison, coal plants saw load factors of 55%, with nuclear plants seeing load factors of 67% over the same period.

Those CCGT plants not generating had been displaced by wind farms and power imports from the continent. Overall demand for electricity generation has also remained very modest for the past two years and may continue to do so as efforts to use electricity more efficiently continue to have an impact, further preventing those CCGT plants from generating as much power as they otherwise might.

Aside from the activity in October 2014, when the wind energy has faltered, demand for electricity has also been sufficiently low to leave the system comfortably supplied. As a result, prices never reached the same heights as seen in October 2014 across the rest of the winter period.


Given this analysis, the decision to close plants such as Killingholme (CCGT) and Littlebrook (Oil) is understandable; whilst the Longannet closure is similarly understandable given the oversupply of generating capacity that currently exists within Scotland.

Ultimately, the market does not need these plants 99% of the time and within the current structure does not want to pay for them. As such any new builds entering into the market are just likely to force the displacement of older existing conventional plants out of the market, resulting in no net increase in supply.

The potential problem with no overall increase in capacity is that once or twice in a decade we will see very high levels of peak demand that could be coupled with low levels of wind generation. Given the levels of capacity awards in the Capacity Mechanism, and the potential for lost capacity as a result, this could indeed lead to occasional blackouts.

Following the Capacity Mechanism of winter 2014/15, the net impact is expected to be a 13.6GW loss of capacity at plants that opted-out or that failed to win Capacity Mechanism contracts against 2.6GW of new builds, resulting in a 11.0GW loss of overall capacity in the market. A solution to this disparity is very simple: to contract for additional volumes in the Capacity Mechanism.

Based on the price of the last Capacity Mechanism auction, every extra GW of power can be assumed to cost the system an estimated à‚£20-25m per annum. While if a further 10GW was sourced, this additional cost might increase to around à‚£40m per annum per extra GW procured.

Awarding contracts for an extra 10GW of generation above and beyond the last auction would almost certainly prevent any future blackouts ensuring ample supply margins, albeit at a cost totalling à‚£2-4bn over the next decade.

The potential cost burden to the consumer through the extra investment required to prevent blackouts is expensive. If procuring the extra 10GW of capacity could prevent blackouts but with only 6 hours of actual blackouts expected to be avoided over the next decade, the cost to achieve this would have been ~à‚£300-600m per hour of blackout. At these costs it is clear that other lower cost alternatives are worth exploring.


These alternatives are the main focus of Demand Side Management i.e. turning demand down as opposed to generation up.

To date, the available demand being managed has been too low to make this a viable solution for the ‘low probability blackout event’ and a general concern has been that when required, business and personal interests could prevent power being curtailed leading to blackouts unless rigorously controlled and managed.

Management of such assets has improved with the entry of firms specialising in demand response with a reputation to uphold, but at the same time demand side response has been allowed to participate in the Capacity Mechanism and displace generating assets from the auction process.

Should the capacity auctioned in the Capacity Mechanism be insufficient, these demand response assets will have already been utilised, at which point it may make sense for large electricity users to have mandatory demand side management as a last resort means of balancing the system.


Alongside the need to ensure the system has sufficient supply in the market to meet demand, the risk of blackouts can also come from a second issue that can often fall under the radar.

The greatest risk of blackouts could actually come not from a tight capacity margin between supply and demand, but as a result of the generators in the market being unable to response within the required window to balance the instantaneous supply-demand position.

Where margin does exist within the market, much of that margin can be sat idle and cold, requiring anywhere from between 1 and 6 hours to warm up and start. With an increasing amount of renewables on the system and with an increasing reliance on interconnectors, a blackout in the UK is more likely to come about not due to a lack of capacity, but due to a lack of capacity able to deliver additional power within the required timeframe.

Over the last year or so, EnAppSys has been monitoring grid frequency and analysing large deviations; which if not managed can lead to instability and the lights going out. What EnAppSys has seen is that there are numerous events where unexpected deviations in wind or unexpected losses of interconnector volumes have required some rapid intervention using pumped storage assets. We anticipate this need will become more acute as the amount of gas and coal-fired power generation online at any one time is reduced as a result of displacement by renewables and new interconnector capacity.

The solution to this problem lies in fast response energy storage, fast response demand side management, spinning reserve and potentially new grid frequency regulation equipment/ plants.


Ultimately the consumer pays the price to keep the lights on and there has been no debate yet as to what the cost is before this becomes too expensive. More capacity would result in greater security of supply, but at the cost of higher bills and as a whole consumers need to decide whether the cost or the security of their electricity supply is the higher consideration.

At the same time, new initiatives and investment in new technologies better suited to the new low carbon world will be require to support the assets we already have and aid our transition to a low carbon grid.

In a low carbon grid it does make sense to install wind farms in Scotland and solar farms in southern England. This may result in the closure of Scottish coal stations and English CCGT stations that are seeing reduced running hours. In response, the market needs to ensure that the technologies are in place to overcome the issues created by this transition without consumers paying an undue cost.

No posts to display