Since 2006 I have claimed that the perfect dispatchable unit for balancing purposes has not yet been invented. P-F Bach 
Guest post by Hugh Sharman, extended bio at the end of this post.
1. Thumbnail summary
The UK Government’s ambitious renewable electricity targets are likely to be met. Unfortunately, the effort and financial subsidies that have done so much to cause huge quantities of wind power to be built has not been matched by the serious effort nor finance needed to deliver commensurate quantities of balancing power to keep the electricity system stable and the “lights on” for when the wind does not blow.
The 30 GW of combined cycle gas turbines (CCGTs) that were listed by DECC as operational as of May 2013, even generating plant that was delivered as recently as 2010, are proving unequal to the task of balancing wind power because this task requires greater flexibility, faster start-up and stopping times and relative robustness to frequent starts and stops. These attributes are physically beyond their capability.
The case of Ireland, where wind penetration reached 18% in 2013 and where CCGTs also deliver nearly all the balancing power, demonstrates that these are performing badly, having a fleet efficiency of roughly 40%, compared with its name-plate rating of or over 55% and which in any case suffers accelerated heat rate deterioration when units are ramped up & down. This low and deteriorating fleet efficiency is accompanied by abnormally high rates of wear and tear. The case of Irish CCGTs is a sort of “canary in the coalmine” warning of things to come in UK.
The complete absence of suitable generating plant that is needed to deliver stable balancing power to the stochastically operating renewables will extend the electricity supply crisis by another decade, at the least and cost many £billions of further investment that are not presently recognised by the UK’s policy makers. During the six years remaining before 2020, the quality of supply will worsen. The fact that no proper financial provision has been made for balancing so much stochastically available electricity will also drive up the price of power to the general public.
2. UK’s renewable electricity targets for 2020 are likely to be met 
Renewable electricity generated in the UK during 2013 amounted to 52 TWh, roughly 14% of all electricity generated (and 16% of energy consumed). Of this 27 TWh, or roughly 7% was generated from wind power.
Whether or not one approves of the UK Government’s energy policy, in any case an inheritance of the UK Climate Change Act of 2008 and of the EU Renewables Directive, one thing is clear. The financial support for “low-carbon” but stochastically available generation has unleashed massive spending for this type of power. But this policy has disincentivized investment into the fossil-fired dispatchable capacity that can deliver secure supplies in cold, windless weather.
Table A Notes: a) Load factors derived from DECC, Digest of United Kingdom Energy Statistics (2013), Table 6.5, use the conservative unchanged configuration data where possible. b) For reasons of concision, Geothermal and Wave data have been removed from the table, though their minor contributions are recorded in the totals.
The foregoing table A was compiled by the Renewable Energy Foundation from UK government data held in the Renewable Energy Planning Database, together with technology load factors reported elsewhere in government data. It shows that, if all the consented capacity is built, then the Government’s “renewables” targets for 2020 will be comfortably over-achieved. According to earlier estimates, if renewable electricity is to form 30% of electricity consumed then 15% of all energy consumed in the UK in that year will be from renewable resources. It helps to view this data graphically (Figures 1 and 2).
Thus by 2020, if all consented capacity in 2014 is commissioned, about which one can still retain some scepticism, there will be 53 GW of renewable capacity of which only 6 GW, mostly biomass-fired boilers, like Drax, could be said to be dispatchable in any way. Wind power will constitute 39 GW. This renewables capacity, if built, will deliver an expected 157 TWh in 2020 (Figure 2).
This will constitute 42% of the 370 TWh of electricity that were generated in 2012. If demand stays roughly stable until 2020, after five straight years of demand falling, 42% of electricity will be generated from renewable resources. This is far more than the target of 30% of electricity consumed required by the EU Directive and the Climate Change legislation and the inter-connected EU agreements and “commitments”.
An initial reaction to this data is that the present Coalition and whoever wins the election in May 2015, can relax the apparently relentless drive to encourage the building of even more new and expensive renewable capacity because what is already listed will comfortably deliver all its targets by 2020. It may therefore be rational to expect a slow-down in the rate of new developments. Similar slow-downs in the previously frantic growth of renewable electricity are already occurring in Spain and Germany.
But there is a stronger reason to re-appraise the whole programme. This is because if so much wind and PV capacity actually gets built, the rest of the electricity and especially generating infrastructure is quite unfit for balancing the stochastic generating capacity that looks like being on-line in just over five years from the date of this document which is mid-2014.
The basic physics of maintaining the stability of an island grid have not changed one iota during these past years. Whatever “smart grid” enthusiasts insist that their technology “will change everything”, in the absence of storage, generating capacity delivered into the grid must always be balanced by the demand drawn from it, from second to second. Electricity storage is still in its infancy, with a few “demonstration” plants being built at the scale of less than 10 MW.
It is relatively simple to balance the UK grid, in 2014, while roughly only 10% of its generation is supplied stochastically. Demand is still pretty much predictable and as long as the system still has ample reserves of dispatchable thermal generation. The task of balancing the system will become exponentially more difficult as stochastic inputs to the system grow beyond this. This is because GB’s incumbent, fossil-fuel-fired generating capacity was never designed for performing this task.
3. How wind balancing is performed in 2014
Figure 3 shows that demand for this randomly chosen but still “low-wind” month, varies roughly between 40 GW during the working day and 25 GW valley demand between midnight and about 5 AM. Nuclear power ran continuously at close to capacity. A large part of the coal capacity is turned down or off at night, responding to lower demand (and lower prices) but is kept hot to restart the next day.
Nearly all the rest of the variable demand and generation is managed by starting and stopping CCGTs or by turning these up and down. When wind output is low (figure 4), the daily pattern of operation of the CCGTs allows both the transmission system operator (the TSO is National Grid, referred to hereafter as Grid) can plan and dispatch capacity with little need for more “hot” or “spinning” reserves than is needed to keep the system stable if the largest generator in the system, Sizewell B, suddenly fails.
Figure 5 The operating profile of the CCGTs involved in daily balancing changed when the wind blew harder earlier in September 2013.
All the CCGTs that supplied the “top” 4 – 6 GW of daytime power must be ready to ramp up and down at very short notice or are required to be frequently started and stopped.
Neither oil nor open cycle gas turbines (OCGTs) were operated during the month indicating that, despite the difficult and expensive operating conditions for the CCGTs, it was still more “profitable” to keep the CCGTs in operation for daily balancing.
The operation of the grid became more stressed in the much windier month of December 2013 (Figure 6).
Figure 6 The peak output of the CCGT fleet during the week, from Sunday the 15th thro’ 22nd December, “daylight” CCGT load was 17 GW but on a windy Wednesday was only 14 GW. The lower peaks during the latter part of the week probably reflect the proximity of the pending Christmas holiday. The daytime variation was in the range 3 – 6 GW.
Figure 7 OCGTs delivered 0.01% of all power during the month, illustrating that no matter how uncomfortable it is for the CCGTs to operate in these conditions, they are still bidding into an energy market that makes these more attractive to operate than OCGTs.
The grid-connected wind fleet in GB at the end of 2013, was roughly 8 GW.
By 2020, if (say) 40% of all power is to be generated from renewable resources and 88% of this is planned to be wind, the grid-connected wind fleet will grow from 8.4 GW (June 2014) to 39 GW, by a factor of 4.9.
4. The system in 2020
By 2020, only 5 years from now, it is most unlikely that any new inter-connectors will have been built and commissioned. Due to the combination of the Large Combustion Plant Directive (LCPD) and the Industrial Emissions Directive (IED) it is likely that dispatchable capacity will have shrunk considerably. Of 21 GW of surviving coal capacity at the end of 2013, only Ratcliffe power station in Nottinghamshire, owned by EON, is believed to be fully IED compliant beyond 2023.
All IED-non-compliant generation from the beginning of 2016 must either retrofit new technology to existing plants to ensure they comply with the new pollution limits or agree to a so-called Limited Life Derogation (LLD). This means that from the start on 2016 the LLD plant can only operate for a maximum of 17,500 hours from 1st January 2016 until the end of 2023, or an average of only 2,500 hours per year.
As was the case with the LCPD, it seems highly likely that those generators with older non-compliant plant that will be expensive to upgrade, and taking into account the rising tax on CO2 emissions, will run their equipment through their 17,500 hours as fast as possible rather than upgrade these, causing a sharp reduction of dispatchable capacity towards 2020.
The IED also applies to CCGTs, affected by NOx emissions as well as coal-fired units.
Accordingly it is entirely reasonable to expect quite large scale closures of some older coal and CCGT capacity by 2020, spurred not least by the gradually increasing tax on CO2 emissions. If Labour wins the 2015 election, it has promised to ensure more stringent environmental restrictions on older plant than might be expected if the Conservatives form the new Government.
As regards nuclear, all the Advanced Gas-cooled Reactors (AGRs) are on “life-support” beyond 2016 when most were scheduled to close. EdF, their owner, is working hard to extend their lives. Hopefully, only Dungeness B will be fully de-commissioned by 2020. However, all these plant life extensions will require extensive periods off-line so that the necessary safety-related improvements can be made.
As figure 8 illustrates, the output of the UK nuclear fleet has been highly erratic for the past few years, so it is reasonable to expect an average nuclear fleet output to be 2 GW lower than the 8 GW average achieved through most of 2014.
For the purpose of taking a view on the balancing of power in 2020, I have selected to extrapolate the conditions of December 2013 to 2020.
In the following calculations, the following assumptions were made:
- Demand patterns and aggregate demand remains the same as in 2013
- Inter-connections, remain as per 2013 but always export 3 GW to France and Netherlands whenever the wind exceeds 10 GW.
- No exchange between Ireland and GB during high wind conditions because wind correlation between GB and Ireland is so strong
- Pumped hydro and hydro will remain as 2013
- Coal (and biomass) output will be 7,500 MW less than 2013
- Wind capacity will be 4 times greater than during 2013
- CCGT “will be retained” in the system at whatever level is required to meet peak annual loads
System operation in 2020
Figure 10 It is instructive to see this situation at a higher resolution for the week 15th thro’ 22nd December.
On the sixth night, because of lower demand, fossil-fired generation is briefly turned off completely and the whole of GB is powered by wind and nuclear, despite strong exports to France and The Netherlands.
CCGTs are supplying virtually all the power needed to balance between generation and demand.
Figure 11 How wind power and CCGT output interact during this period of high wind.
It is abundantly clear that with this quantity of wind power in the system, even with 7,500 MW of coal closed down, and wind power taking priority for dispatch, that by 2020, no CCGTs at all will be operating in any way resembling base-load (Figures 9 and 10). Yet, during periods of low wind, there will have to be at least 30 GW of CCGT (or other dispatchable) capacity beyond nuclear, if only “to keep the lights on”.
It is also clear that any more wind power than the 32 GW simulated in these calculations would threaten the base-load status of the diminished nuclear fleet.
However, this type of operating regime, with constant starts, stops and hard ramping will rapidly destroy GB’s large, incumbent, and elderly CCGT fleet.
The unsuitability of the Frame-type CCGT for multiple starts stops and hard ramping
Figure 13 Source: Author, compiled from DECC Table DUKES 5.1 (2013)
Figure 14 Source: www.Corelia.co.uk
This is because the start-up cycle of the typical, incumbent CCGT is so long relative to the average time that it will be operating at full, optimum load (Figure 14).
Synchronisation does not begin until roughly a half an hour after the start button is pressed. The Frame-type turbine does not reach full output for seventy five minutes.
During the whole, 100 minute plus long, start-up operation, full load and optimum efficiency will, in many cases, last just a few hours before the 40 minute shut-down process begins, during which time, once again, the CCGT will be operating with greatly sub-optimal efficiency, causing disproportionate fuel costs and high specific emissions.
Furthermore, in addition to the high fuel and emission costs, which can average £10,000 – £13,000 in fuel and emissions, each new start is costly in wear, tear and shortened component life, estimated by a notable industry consultant as £10,000 – 12,000 per start.
Every time a power plant is turned on and off, the gas turbine, HRSG or boiler, steam lines, steam turbines, and auxiliary components go through unavoidably large thermal and pressure stresses, which cause damage and incur additional future costs. Frequent starts and stops and high ramping and up and down will shorten the periods between serious and costly outages caused by failures in the major components in the steam and gas generating equipment.
This damage is made worse by fatigue and creep-fatigue interaction damage. These cycling-related costs and damages are strongly correlated to start-up, shutdown, and rapid load following. It is therefore necessary to estimate and compare them to similar combined cycle plants with longer histories, where there is more data for start/stop and load following costs, the so-called cycling costs.
In summary, these costs fall into the following main categories
- Increases in maintenance, operation (excluding fixed costs), and overhaul capital expenditures
- Increased time-averaged replacement energy and capacity cost due to increased equivalent forced outage rates (EFOR)
- Increase in the cost of heat rate changes due to low load and variable load operation
- Increase in the cost of start-up fuel, auxiliary power, chemicals, and extra manpower for start-ups
- Cost of long-term heat rate increases (i.e., efficiency loss)
All the UK’s CCGTs are, to a greater or lesser extent vulnerable to these issues which deteriorate as the plant’s running hours increase.
5. How they manage balancing elsewhere
Figure 15 The case of Denmark is most instructive, as always. Denmark’s wind power generated during 2013 was 34% of all power generated in Denmark and the equivalent of 33% of power consumed in Denmark. Denmark was a net importer of power during 2013.
Figure 15 illustrates how, whenever the wind blows strongly, a high proportion of Danish wind power is not actually consumed in country but instead floods into neighbouring systems (Norway, Sweden and Germany) through inter-connectors, the aggregated capacity of which have the same capacity as Denmark’s peak load. This is 5.8 GW and will grow to more than 7 GW by 2020, if current plans are fulfilled.
A similar phenomenon can be observed in Germany where stochastic renewables in 2014 constitute roughly only 15% of all electricity generated.
Figure 16 Germany, power generation, demand and net power flows June 16 – 22, 2014
Most of the time, peak PV output (especially) coincides with up to 10 GW of exports that are flooding into its eight inter-connected, neighbouring systems. Indeed, Germany has been a net exporter of electricity ever since the rapid build-up of PV started in 2009.
The abrupt slowdown in the rate of growth of its stochastic generating infrastructure that is taking place during 2014, is due almost as much to the objections of its neighbours, whose much smaller transmission systems are under pressure from loop-flows generated from Germany, as due to the unacceptably rising costs being imposed on German consumers by the “renewable energy law”.
The inter-connector balancing capacity that is available to Denmark and Germany, will simply not be available to the UK, by 2020, the year that (possibly, if improbably) up to 40% of the UK’s generation will come from stochastic renewables, its inter-connections with France and the Netherlands will be no more than now, 3GW and altogether, with Ireland, 4 GW.
The two foregoing cases demonstrate the central importance of having high inter-connections with neighbouring electricity systems in order to achieve high levels of integration, in the absence of large levels of pumped hydro and/or other storage, as demonstrated in the cases of Spain and Portugal.
All these cases, though interesting, are not relevant to the focus of this paper which is how the UK will raise wind penetration from 7% in 2013 to almost 40% by 2020.
The Republic of Ireland (ROI) generated 18% of its electricity from wind turbines during 2013 and its policy is to generate 40% of its electricity from wind by 2020. These targets are comparable with the UK’s.
Whether this is realistic is another matter. It remains a firm policy intention with bilateral political support in the Irish Parliament.
The Irish system and its experience with wind power integration is especially relevant because of the striking similarity between the two island systems.
The following table, which compares the dispatchable generating of the two island systems, is instructive despite being almost three years out of date.
Figure 17 Source: Eirgrid & DUKES Table 5.11 (2010) compiled by the author
Both islands have tiny inter-connector capacity with their neighbours. Only CCGTs have been built since the early 1990s and coal capacity is in decline.
Overwhelmingly, CCGTs provide the lion’s share of wind balancing.
The author has calculated the specific fuel emissions of the CCGT fleet during 2012, before the E-W inter-connector was commissioned and found that the fleet efficiency of ROI’s CCGTs and therefore their specific fuel and emissions costs are substantially lower than nameplate rating of more than 55% (lower heating value).
These calculations demonstrate very clearly that even at only 16% annual wind penetration, the average fleet efficiency of the CCGTs is substantially less than nameplate rating, being under 40%, causing high specific fuel costs and CO2 emissions.
Furthermore, the anecdotal evidence of serious plant failures is strong. Documentary evidence is scarce because this data is closely held and regarded as commercially sensitive.
The complete unsuitability of CCGTs for the only remaining task they will have in the UK, as so much more wind power becomes installed, is not yet publicly recognised, although there can be no doubt that the generators understand this well enough. The absence of a willingness to invest in new CCGTs is not just because of the uncertainties of the Electricity Market Reform but is driven by the realisation that CCGTs cannot operate profitably in the market being created by so much wind power having priority on the system.
The high likelihood of a pending, forced write-down of the 30 GW of the UK’s CCGT capacity, with a replacement value in excess of £20 billion that must be spent by 2020 in order to “keep the lights on” when the wind is not blowing, needs the most urgent public recognition.
Technical solutions that are better suited to high wind penetration are being developed but do not yet exist. However, the need for the early replacement of most of the incumbent 30 GW CCGT fleet will produce another financial shock in the market for which neither UK policy makers nor the public are properly prepared.
 DECC, DUKES Table 5.11, May 2013
 The heat rate of the UK fleet in recent years has been rising and the efficiency falling to well under 50% (LCV), DECC DUKES 5.10, 2013
 The vote in favour of the motion was overwhelming, quote “The House having divided: Ayes 463, Noes 3”, unquote
 Main driver of UK policy on renewables is the EU Renewables Directive of 2009, which requires that 15% of Final Energy Consumption in the UK should come from renewable sources in 2020. The UK govt. expects that that about half this quantity will come from electricity, entailing that about 30% of final electricity consumption will be renewable
 Consented capacity will generate about 110 TWh. If all capacity in the planning system is also consented, this will add a further 46.6 TWh
 A bold assumption, given the faster economic growth that has been delivered recently?
 UK has 2.8 GW/25 GWh of pumped storage which is located in Wales and Scotland, with 600 MW additional pumped storage that SSE is developing at Loch Ness. This is trivial in relation to the average 1,000 GWh generated and consumed every day but of course, is useful capacity for helping to stabilize the system.
 Dungeness B was unexpectedly forced to close down due to a failure half way through the month, illustrating a normal hazard to be expected in an ancient electricity system, nearing the end of its design life.
 All the oil-fired power stations are now decommissioned
 “Embedded” wind power which is not monitored by Grid is roughly 2 GW. This type of the wind power is most unlikely to grow.
 In Denmark and Germany, where renewable penetration is much higher than in UK, changes in the pattern of use of power intended to exploit changes in the amount of renewables produced, have been negligible.
 Neither the new Channel Tunnel inter-connector nor the proposed Norway inter-connector are assumed to have been commissioned before 2020.
 It is possible but improbable that the 0.6 GW pumped hydro at Loch Ness will be commissioned by 2020.
 32 GW, a more conservative assumption than the 39 GW foreseen in REF’s paper
 For the purpose of this paper, issues of grid stability, inertia and ROCOF (rate of change of frequency) are conveniently ignored. They must be addressed of course but will not be dealt with this paper
 Private communication, generation industry source. Gas priced at £7.50/GJ
 Private communication, generation industry source and http://www.nrel.gov/docs/fy12osti/55433.pdf
 In Ireland, where wind penetration reached 17% by 2013, there is a wealth of closely held data about the maintenance costs of CCGTs stressed by balancing wind power
 Data compiled from www.energinet.dk by the author
 In other words, the inter-connected capacity of the UK would have to be 60 GW to achieve the same flexibility as the Danish system
 Private communication with a retired but still active director of the Danish power sector with unique experience at the top of both power generation and power system operations.
Electricity supply and demand for beginners
How Much Windpower can the UK grid handle?
The changing face of UK electricity supply
Parasitic wind killing its host
Brave Green World and the Cost of Electricity
The Coire Glas pumped storage scheme – a massive but puny beast
Hugh Martyn Sharman
Hugh Sharman is the owner/director of Incoteco (Denmark) ApS, a Danish energy developer and consulting engineer since 1986.
He graduated in civil engineering from Imperial College, London, in 1962.
During his career formative years, until the early 1970s, he specialized in innovative offshore construction with the French group, GEM-Hersent. This work was mainly connected with on oil and gas projects in the Persian Gulf, USA, France and the UK.
He has been involved in energy engineering and developments since the 1970s.
During the mid-1970s, his UK-based energy company, Conservation Tools and Technology, pioneered the popular use of renewable energy and the parallel need for much greater fuel efficiency in UK. Between 1977 and 1986, he was Area Representative in the Caribbean for advancing the power generation activities of Gothenburg-based Swedish Shipyards and Mitsui Engineering-owned BWSC AS. During this time, he was responsible for power station development and sales in Venezuela, Barbados, Puerto Rico, Bahamas and Bermuda, using the world’s then most efficient and fuel-flexible equipment available at the time, being low-speed, 2-stroke, marine-type engines.
He founded Incoteco (Denmark) in 1986 and has since undertaken power station and energy related work, focusing on technically innovative environmental and energy processes. His clients have included TRW Inc., Rolls Royce, Scottish Hydro, Renewable Energy Foundation, Ormat Inc, VRB Power, Danish Energy Agency, Mission Energy, Qatar Petroleum, Norsk Hydro, ECA International, Elsam, Kinder Morgan CO2, Mott Macdonald, Eos Energy Storage, among many others.
In 2004, while performing a study for the Danish Energy Agency on the use of transport hydrogen, generated from “excess” Danish wind energy, to improve the economy and overall efficiency of Danish wind energy, he became convinced that distributed electricity would change energy paradigms such as is becoming clear in 2014.
Since 2008, Hugh has been closely involved with early stage commercial development of several, advanced electrical storage technologies for Canadian, US and Chinese companies. Electricity storage processes will be necessary to address the more efficient integration of increasing quantities of stochastic renewable energy in Europe and elsewhere. He continues to consult to and for this industry and actively monitors global progress towards robust, low cost electricity storage technologies.
At present, he is pursuing the same objective by leading the development of a novel, high efficiency, high-flexibility, thermal generating cycle, for two globally famous equipment manufacturers. When commercialized, hopefully in 2014, this will address the urgent need for the more fuel efficient and flexible integration of large quantities of wind energy into the UK and Irish electricity systems.
He has written many technical and policy articles for energy magazines and has made presentations on energy policy, for, among others Qatar Petroleum, OPEC and The Economist Magazine. He remains the editor of the energy blog www.DimWatt.eu . Such papers are available on request.