Leo Smith from Gridwatch has been monitoring the electricity transfers between Scotland and England since 30 December 2015. This post presents a first look at these data.
- The transfers are almost exclusively Scotland to England
- The transfer peaks define a plateau of 3.5 GW that defines the inter connection capacity
- The pattern of transfers follows wind supply quite closely. An argument can be made that all of Scotland’s wind power is currently exported.
- Scotland currently meets 100% of indigenous demand from conventional sources, mainly nuclear, coal and hydro.
- A diurnal demand cycle is embedded in the export pattern, but curiously, export peaks are centred on midnight.
- There is evidence that wind-curtailment may be used as part of load following strategy.
Just before Christmas, George Lindsay sent me this link from National Grid that monitors electricity flows between Scotland and England. I sent the link to Leo Smith at Gridwatch and enquired if it could be monitored. Leo obliged and we now have the first couple of months of data. Many thanks Leo!
The interest here is to monitor Scotland – England electricity transfers leading up to and following the closure of the 2.4 GW Longannet coal fired power station at the end of March. Longannet is part of the backbone of Scottish electricity supply.
Figure 1 Scotland-England electricity transfers as published by National Grid. The raw data contains a large number of errors and omissions which have been repaired (See appendix). Click all charts for a large readable version.
Figure 1 shows the pattern of Scottish electricity exports to England. There are five occasions when exports fell to zero and went slightly negative but on no occasion was Scotland a significant importer. As we will see below the pattern of exports is correlated with wind, but wind is not the only controlling variable. There are some other key points to be made first.
First, electricity exports peak on a plateau value of 3.5 GW. Presumably this is the capacity of the inter-connection, substantially higher than the 2.66 GW I estimated in a recent post. Finding concrete details on the size of inter-connection has remained elusive, but I believe we now have the answer – 3.5 GW.
Second, there is a clear signal through most of the data that the Scotland-England interconnection is following diurnal load and is therefore engaged in load balancing. My immediate assumption that Scottish hydro was being deployed to help balance national load proved false. Looking closely at the data 22 to 28 February it can be seen clearly that Scottish exports to England are peaking around midnight and are reduced during late afternoon peak demand periods. This happened during a wind lull and Scotland exported less at time of UK peak demand. This is a form of negative balancing. Rather than produce more power when it is needed, Scotland simply exports less. This presumably places greater strain on the system south of the border,
Correlation with Wind
Figure 2 We do not have data on Scottish wind production and the Scottish output is based on pro-rating UK output based on installed capacity. Scotland = 0.41 of the UK.
Electricity exports follow the pattern of wind production but are normally significantly higher than wind supply (Figure 2). It should be clear that most of Scotland’s wind production is exported. If you are upset by a wind farm on the hillside opposite your house you should know that its only function is to send electrical noise to England. But this will change with the closure of Longannet. Cross plotting wind v exports shows R^2=0.49 (Figure 3). Wind clearly influences the export pattern but is not the sole variable. The other main variable is diurnal load following exchanges.
Further observations to make from Figure 2 are that wind peaks are following a plateau of 2.5 GW (41% of total UK wind) suggesting that curtailment may occur at values above this. And inspection of the graphs in UK Grid Graphed does indeed suggest that peak supply is following a plateau. There is also evidence for a diurnal cycle in the wind that is difficult to understand. In summer this could be explained by sea and land breezes but in cyclone dominated winter I don’t see that the wind should decide to blow harder at midnight. I will speculate that the diurnal cycle seen in the wind data may be down to curtailment used to follow load. Once again this is evidence for load following achieved by withholding supply rather than using balancing reserve.
Figure 3 Wind production versus Scotland-England exchanges shows a fairly high degree of correlation. The maximum exchange of 3.5 GW is clearly defined.
Figure 4 National Grid have not yet published Scottish demand data for January and February, hence the model above is based on pro-rating Scottish demand = 10% of UK demand. Adding exports (red) provides a picture of Scottish supply.
Adding exports to Scottish demand shows the pattern of Scottish supply for these two months. Peak supply defines a plateau of about 8 GW. The generating assets at disposal were as follows:
- Hunterstone B nuclear 1 GW
- Torness nuclear 1.2 GW
- Longannet coal 2.4 GW
- Hydro peak 1 GW
- Peterhead gas 0.4 GW (+0.75 GW supplemental reserve)
- Wind peak 2.5 GW (constrained?)
This provides a total capacity of 8.5 GW compared with the 8 GW peak output. Note that I have forgotten to include the 0.25 GW Moyle inter-connector (N Ireland) in this analysis and so peak demand and capacity are almost matched. Including exports to N Ireland, it is evident that all the generating assets have been deployed in January and February and were running flat out when demand was there.
Figure 5 Deducting wind from exports leaves this residual that represents the exported conventional supply.
We can scratch the data a little deeper. With exports normally following but running ahead of wind generated it is clear that a component of exports comes from conventional dispatchable power. Deducting wind from exports leaves a residual that represents this component (Figure 5). Two main observations here. The first is that this dispatchable component peaks at around 2.4 MW coincident with the output from Longannet. Second, the diurnal cycle is more strongly pronounced in the residual. But once again the export peaks are normally centred on midnight. Why on Earth are Scottish power stations sending electricity to England at times of low demand?
Figure 6 shows these various components. Generation from Longannet currently contributes to Scottish demand and provides a surplus for export. Its closure will make Scotland dependant in part on wind to keep the lights on.
Figure 6 Scottish supply broken down into the three components of indigenous demand that is met from conventional sources, exported conventional surplus and exported wind.
The main risk to the Scottish system lies in one of our nuclear power stations tripping in cold calm winter conditions and the risk that this trips the other nuclear power station leading to a nationwide blackout. Power stations require power in order to operate and re-energising the grid from a black start is a recognised risk.
This Scottish Government report describes the black start protocol as follows:
Scotland is a single Black Start contracting zone and currently the SO has Black Start contracts with Crauchan, Foyers, Errochty and Sloy with each TO having 1 LJRP to manage. If Longannet or Peterhead remain open the SO is able to use them with the hydro and pump storage stations to Black Start and provide skeleton restoration in the SP Transmission and SHE Transmission areas within 12 – 18hrs. If Longannet and Peterhead close then the SO will adopt the alternative strategy, which is based on energising from England and Wales transmission system, in conjunction with hydro and pump storage generation in Scotland providing a skeleton restoration in 24+hrs.
SO = system operator
LJRP = local joint restoration plan
TO = transmission owner
SP = Scottish Power (grid operator in the south)
SHE = Scottish Hydro Electric (grid operator in the north)
Cruachan and Foyers are pumped storage hydro schemes
Errochty and Sloy are conventional hydro schemes
Via email, engineers have raised concerns that Scotland will not manage a black start without Longannet that is a massive, robust and flexible beast at the heart of the generating system. And so it appears that Scotland may be dependent on the English grid but the same engineers expressed concern about the viability of black starting Scotland via the England – Scotland inter connections.
If you live in Scotland, there has never been a better time to invest in a wood burner and candles.
The Scotland-England exchange data are being archived alongside the other Gridwatch data but are not yet available from the web site. Leo sent me a link that accesses ALL the Gridwatch data that is very cumbersome and almost crashed my computer. So I’ve not published the link lest too many folks try to access it and crash the Gridwatch server. Anyone who wants the link only needs to ask.
The data are also full of timing errors, essentially missing lines and I had to spend several hours repairing it. This includes 14 hours of missing data on 20th / 21st January.
The data are recorded with 5 minute resolution. This has been converted to hourly averages to make the files more manageable.