Control of the National Grid (Great Britain)

The National Grid is the high-voltage electric power transmission network in Great Britain, connecting power stations and major substations to ensure that electricity generated anywhere in Great Britain can be used to satisfy demand elsewhere. There are also undersea connections to northern France (HVDC Cross-Channel), Northern Ireland (HVDC Moyle), and the Isle of Man (Isle of Man to England Interconnector).

The national grid is required to maintain stability within specified standards of frequency and voltage.

Power generation and transmission statistics

• Total generating capacity is about 70 GW, supplied roughly equally by nuclear, coal fired and gas fired power stations.

• In the UK, the peak winter demand is 57GW (57,000 MW).
  • N.B: This peak would be much higher if it were not suppressed by various mechanisms such as maximum demand tariffs, and the system of Triad warnings and charges.

• Annual power used in the UK is around 3.6 ×1011 kWh
  • N.B: The average load factor is then (3.6 ×1011 /(8760 x 57x 106 ))x 100 = 72%

• There is generally about 1.5 GW of so called spinning reserve – this is typically a large power station paid to produce at less than full output. So, a typical power station, which might have 4 generating sets each of 660 MW, giving a total output of 2.64 GW, might only be operating at 2 GW, with the steam boiler full, but with the steam valve not fully open. At the request from National Grid control centre, or under command from the generator governor this valve can open up and deliver an extra 640 MW in 20 to 30 seconds. This requires the boiler air fans and the coal feeders to increase output accordingly. The greater the total load on the system, and / or the greater the expectation of large demand fluctuations (at the end of popular TV programmes for example) the larger the proportion of spinning reserve set by National Grid Transco (NGT).

• National Grid Transco (NGT) pays to have up to 8.5 GW of additional capacity available to start immediately but not running, referred to as “warming” or "hot standby", that is ready to be used at short notice which could take 1/2 hour - 2 hours to bring on line. Generally, there will be more of such "hot standby" capacity whenever there is a large amount of expected disturbance on the system. The cost of fuel or tonne of CO2 emitted by keeping such plant warm is tiny in comparison with the amount of fuel used to generate power, maybe equivalent to the fuel used to produce a 1/4 of a MW compared to a full load fuel demand for a large set of 1,800 MW. Often quoted talk about the high costs of standby spinning reserve are misleading.

• A similar amount of power stations - 8GW to 10 GW by capacity - are operable from a cold start in about 12 hours for coal burning stations, and 2 hours for gas fired stations.

• At any one time a large amount of power stations are unavailable due to regular maintenance, being off line due to a fault becoming apparent or because of sudden breakdown.

• Other stations are “mothballed” or "deep-mothballed" which means they cannot be readily called upon; even in an emergency it may take several months to de-mothball. Recently (as of Summer 2006), Fawley Power Station near Southampton has been de-mothballed to cope with anticipated power capacity shortages for winter 2006/07.

• Up to 2.25 GW of capacity can be supplied by the Frequency Response participants mentioned below – steel works, cold stores etc.

• 2.25 GW of Standing Reserve diesel generators (as per Wessex Water), small gas turbines (e.g. 22 MW operated by First Energy) which as we shall see augment the Frequency Service arrangements.

• 2 GW Fast response plant such as large open cycle gas turbines (OCGTs). OCGTs are gas turbines that are in the range 25 – 100 MW, and which can start in a few minutes (slower to start than diesel engines and marginally less reliable upon start up). Normally, these are not used for power generation since their low operating efficiency means that the cost per capacity supplied is prohibitively expensive. However, they have low capital cost, as demonstrated by the 100 MW unit at White City, London.

• The pumped storage schemes at Dinorwig and Ffestiniog can offer up to 2 GW of power within 15 seconds. (Incidentally, at the time Dinorwig was built, which was solely to cope with the inflexibility of nuclear power and the inherent unreliability / intermitancy of large power stations, a further similar station was planned on Exmoor but was never built, so presumably it could still be built if the need arose)

• A cross channel EHVC power line can bring in up to 2 GW of power from France, though this tends to be unreliable.

Load and generation response mechanisms

The national grid is organized, and power stations distributed, in such a way as to cope with sudden, unforeseen and dramatic changes in either load or generation. It is designed to cope with the simultaneous or nearly simultaneous failure of 2 x 660 MW sets, which it evidently does with ease as these events happen on average at least once a month.

Frequency Service

NGT (National Grid Transco) who operate the national grid and control the operations of power stations (but does not own them) has a number of partners who are known as NGT Frequency Service participants

These are large power users such as steel works, cold stores, etc. who are happy to enter into a contract to be paid to be automatically disconnected from power supplies whenever grid frequency starts to fall. An example of such a participant would be a large steel melting furnace, which may take a day to heat up using an electric arc or induction heater, and is not adversely affected if the process is delayed by 20 minutes. The same applies to a large cold store where interruption in cooling for 20 minutes can readily be accepted in the same way as the normal domestic freezer can happily be shut off for 36 hours without gaining significant temperature rise. These disconnections can obviously assist enormously if a sudden power demand is made on the grid or if there is the sudden loss of generating capacity.

This instant switch off is achieved using a relay provided by NGT mounted on the power supply to the major plant switch gear. It is set to detect the falling frequency which can occur when a large power station fails suddenly or there is a sudden rise in demand, and opens the circuit breaker to the furnace or cold store. Ancillary circuits in the factory are unaffected, such as lights and power sockets. These Frequency Service participants are contracted to stay off for up to 20 minutes.

The exact setting on the relay can be remotely monitored and controlled by NGT – such as the exact frequency at which the relay disconnects the load, whether the relay is armed or not, whether the customer has temporarily exercised his right to over-ride the relay etc.

These Frequency Service participants receive a fee which is of the order of more than several thousand pounds per MW per year – it is a capacity fee per MW not per MWh.

NGT Standing Reserve

Operating closely with NGT Frequency Response is NGT Standing Reserve. NGT Standing Reserve participants are small diesel engine owners, and Open Cycle gas turbine generator owners, who are paid to start up and connect to the grid within 20 minutes from the time Frequency Response customers are called to disconnect. These participants must be reliable and able to stay on and run for an hour or so, with a repetition rate of 20 hours.

A typical Reserve Service partner would be Wessex Water one of ten water and sewerage companies in England and Wales, covering Somerset, Dorset, Wiltshire and parts of Avon. There are many other Reserve Service partners, hospitals, police headquarters etc. but Wessex Water is used as a typical example.

Wessex Water has about 550 emergency standby diesel engines, totalling 110 MW of capacity whose primary function is to power essential services such as sewage works and water supply works during power failures which happens on average a few hours each year. Of this number, about 33 units totalling 18 MW are also used in a number of non-emergency ways commercially which are called collectively Load Management, and which includes routinely feeding power into the Local Distribution System and ultimately the National Grid. These generators presently have a 4 minute start up and paralleling capability automatically from our control room, and are currently being modified to enable start ups in less than one minute. These units are quite small, 0.24 MW to 1 MW range and it may surprise people to know that these are used by the National Grid on a regular call off basis to supplement its arrangements with power station owners. Essentially this systems drastically cut the amount of expensive Spinning Reserve that would otherwise be needed.

Again substantial fees can be earned simply for making these engines available, again backed up by complex contracts with specified levels of reliability, response times, frequency of use and so on.

Wessex Water is only one of many such participants.

Diesel generators

Wessex Water diesels can all start and be feeding power into the grid within 4 minutes from receiving the start signal from NGT. This delay is mainly due to the time taken to dial up each set. There are currently 4 auto diallers and modems each of which have to make the calls in series. The company is presently switching to broadband where the calls will be instantaneous and the start up time will then be less than 30 seconds to full load, which is far less than any conventional power station.

(With specially modified diesel engines (all standard factory made modifications such as air pressure vessels to start the turbochargers, air start rather than battery start, jacket warming, continuous pre-lubrication, continuous slow rotation - etc.), start up and full loading can be achieved within 1 second)

Use of the Reserve Service and Frequency Service in practice

An example illustrating how the Reserve Service and Frequency Service are used in order to cope with Intermittency / Variability is given below:

• Consider, if a 660 MW turbine generator set, (the standard size of large steam turbine) trips (large power stations usually consist of 2 or 4 sets each of 660 MW).This can happen for all sorts of reasons – a coal crusher might break down, boiler tubes might fail, an alternator might start to overheat, insulation might fail on the alternator. In the event of certain failures the generating set would automatically trip out and because the grid has suddenly lost 660 MW, which on a typical day might be 1.3% of the total national grid output, then due to the immediate imbalance between supply and demand, grid frequency immediately starts to drop from the standard 50 Hz.

• As soon as this happens, the under frequency relays on Frequency Response customers begin to trip of their load as the frequency falls, ultimately to shed total load equal to 660 MW. These relays are set at a range of frequency between 48.5 and 49.5, so the 660 MW of generation that has been lost is not instantly matched by these relays shedding 660 MW of load simultaneously but progressively as the frequency drops, until exactly enough is shed to exactly match the remaining power station capacity. This will then stabilise the frequency at its now lower level – perhaps 49.3 HZ. This all happens in less than a second.

These Frequency Response participants are only contracted to have their shed loads off for up to 20 minutes.

• At the same time the NGT Control Room issues start up signals to sufficient of its Standing Reserve Service participants (that’s people like Wessex ) for 660 MW (or whatever power it is short of) which by contract have to become available within 20 minutes.

• The NGT Control Room is monitoring the situation and if sufficient Standing Reserve capacity does not come on, then it can order more until it has exactly matched what the Frequency Response relays have shed. (These relays are monitored in real time by NGTs telemetry systems)

• Due to differing circumstances on the ground, different Standing Reserve participants will have different start up times and reliabilities which again NGT monitors using telemetry. However, when sufficient Standing Reserve has become available, which would be in less than 20 minutes, the Frequency Response loads (steel furnaces, cold stores etc) are automatically re-connected by the relays again gradually so as not to destabilise the system.

• The original NGT Frequency Response relays are then re-armed by NGT.

• Up to an hour or so later, the output of the Standing Reserve diesels and gas turbines, (which are nearly all in private hands, i.e. not professional power generators as such) will have been augmented, and then replaced with new levels of large gas or coal fired power stations which together will have driven the frequency back to its correct level close to 50Hz. The diesels can then be stood down, ready for the next emergency.

• The replacement generation sources would have come from increased outputs from other power stations on spinning reserve, resulting in increased output.

• At the same time new levels of spinning reserve will have been created, which might have been stations on hot standby / warming now switched to running.

• Increased levels of stations on hot standby will also be called for.

The foregoing is a brief description of how the National Grid organises itself and the dispatch of power stations to cope with sudden, unforeseen and dramatic changes in load or generation. It is designed to cope with the simultaneous or nearly simultaneous failure of 2 x 660 MW sets, which it evidently does with ease as these events happen at least once a month.

The point to note is that complicated as this may sound, this has been going on for many years as the loads imposed on the grid, and supply of power from power stations is by itself extremely intermittent already, simply due to the sudden and unpredictable failure of these large 660 MW generating sets, or sometimes the entire power stations; and the sudden changes in load which can happen at the end of a major TV program, or events such as the last eclipse of the sun. These latter can cause surges of several GW which whilst larger in magnitude than the sudden loss of 2 x 660 MW sets, are not instantaneous and so are not as severe a shock to the national grid system.

It should be noted that however reliable a power station is, grid operators have to assume that it can fail, and that it will fail, so its replacement must always be running and available.

Voltage Control

There are a number of ways that voltage control is undertaken on the National Grid 400/275/132kV system. This can be done by:

• Over/Under excitation of generators
• Switching in/out of Shunt Reactors
• Switching out of Overhead Line and Underground Cable circuits
• Tap staggering of both Supergrid and Grid Transformers
• Under/Over excitation of Synchronous Compensators
• Declutching Open Cycle GTs into the synchronous compensation mode
• Static Variable Compensators
• Manually Switched Capacitor Banks
• Synchronous Compensation of generators

Sources of intermittency on the National Grid

The largest source of intermittency on the National Grid is thus the power stations themselves; in fact, it is Sizewell B nuclear power station. Whenever Sizewell B is operating the entire output is capable of stopping at any time, without warning. Its capacity is 1.3GW or 2.16% of the national grid maximum demand. Thus, this is the largest source of intermittency, since it is the largest power generating unit which can be lost, yet NGT readily copes with it using the methods outlined overleaf including the use of diesel engines. Although it comprises the standard 2 x 660 MW sets, for various reasons, we needn't go into here, the entire output can be lost, which is not the case in coal fired power stations.

(In fact there was a big row when Sizewell came to be commissioned, since the National Grid's licensing regime had changed during the construction phase, and technically the power station could not be connected. It only happened due to Downing Street intervening and putting pressure on National Grid, then in government hands)

So the kind of intermittency that 100% penetration of wind would introduce is nothing like the intermittency already there due to the power stations themselves. Even at the most extreme case, the simultaneously change in output of all wind turbines would take at least an hour to achieve the change in output caused when Sizewell trips which is instantaneous and entirely unpredictable.

Paradoxically, although wind power is inherently intermittent and variable it is in fact much more reliable than conventional plant. Consider a 660 MW set. This could be replaced by perhaps 300 x 3MW wind turbines to give the same annual output of energy. On a day when wind strength is enough to give a total output of 600 MW, then these simply cannot all fail simultaneously. Neither can the wind suddenly stop at all locations simaltaneoulsy, particularly if they are spread around the country.

Furthermore, the most reliable form of wind forecasting is to simply look at the total output of the wind turbine themselves – in all probability, what they are producing at one point in time, is likely to be produced one hour later, or only a small change from that. It this prediction window is decreased – 20 minutes, 10 minutes 5 minutes, the difference in total national wind power output becomes less and less, and even at 5 minutes, that is ample time to raise or lower spinning reserve accordingly. There is thus ample time to cope with these changes by calling up or standing down more or less plant. If the 5 minute estimates are wrong then the Frequency Service and Reserve Service diesels will clearly have the resilience to cope with it.

Though as we shall see later, this already easy to control situation can be made even easier.

Note also that the cost of spinning reserve is not high as is often erroneously stated. The efficiency of a plant might change from say 37% to 36.5% if the output of the set were dropped from 660 MW to 500 MW (100 MW spinning reserve), which is a tiny fuel penalty compared to the total amount of fuel passing through the power station. In fact it is about 1.5% - most of which has to be carried anyway.

Diesel generators

Perhaps surprising, a large number of small diesel generators are presently used within the national grid, in order to assist with sudden shortages of capacity or intermittency.

Connecting diesel generators to the National Grid

All modern diesel generators come with an electronic governor and this enables them to be safely synchronised and paralleled with each other and the mains. To do this the generator set must be operating at exactly the same frequency as the mains, and exactly in phase the circuit breaker is closed. This is all handled automatically by the electronic speed governor and synchronising system.

To meet the technical and safety standards of the electricity distribution companies, special monitors are required so that the set stays paralleled safely and is automatically disconnected if there is a fault. This G59 equipment (as it is colloquially referred to) is relatively inexpensive nowadays.

Wessex Water diesel generators

Wessex Water has a total of about 550 diesel powered generators ranging in size from 50 kW to 1200 kW, with a total capacity of over 110 MW.

32 Sets spread across 24 Sites totalling 18 MW have already been converted to Load Management, and some of the others are now being converted as well. These are at clean water sewage treatment works

The Wessex Water diesel generator control system

Located at the Wessex Water control room, this control system consists of a standard PC controlling a master Mitsubishi PLC (a small computer). Using 4 auto dialling modems and standard telephone lines the master PLC talks to a further PLC retro fitted to each of the remote generators spread over the South West.

The PC with software developed in-house is used to monitor the status of each engine (e.g. availability, running time, stopped or not, failure mode, fault type, fuel level, hours run etc). It can also be used to start and stop selected sets manually. Whenever the system is about to operate , the local treatment works operators are automatically sent an SMS message telling them that it is about to start, mainly so they are not startled by a sudden starting of the generators - no actual attendance is needed. The company is fitting a system to automatically detect low fuel and send an email to fuel suppliers to refill the sets.

The system also flags availability of sets to NGT.

Typical conversion and operating costs for a diesel generator

• Approx. £3,000 to fit the PLC to the set
• Paralleling and synchronising gear and G59 equipment (this allows grid connection) Approx £5,000
• Tidying up set (noise, larger fuel tank) Approx another £5k

As an example:

• A 1MW set…£13/kW
• 50 kW…maybe £260/kW
• Running costs - fuel 10p/kWh
• Maintenance about 0.5p/kWh

This is very cheap capacity considering power stations are about £350/KW for a CCGT. A diesel set itself is about £150/kW fully installed and connected.

The number of diesels generators in the UK

Wessex Water has 550 generators of capacity 110MW in the range 50 kWe to 1.2 MWe. Presently, it only uses 32 generators of the larger sets (greater than 250 kW) with 18MW total capacity for Load Management / Triads / Reserve Service. Many of the other smaller sets have started to be converted also.

According to EA Technology nationally there are up to 20 GW of emergency diesels. With the right financial incentives and explanations of the benefits large numbers of these could be brought into the Reserve Service type of scheme. Over 20 years this practice and associated technology will probably become standard.

Revenue earning opportunities from third parties

Perhaps surprisingly, Wessex Water along with other generator owners also has contracts with national generating and energy supply companies, who also pay to operate the diesels remotely from time to time. This is for there own balancing purposes and when they are short of capacity. Notably there was extensive third party running during the gas shortages of winter 2005-2006 when most of the CCGTs that could also switched from gas to run on liquid fuels.

NGT SRD computer

This is a separate PC provided by NGT which is physically adjacent to and linked to our PC and linked via broadband to NGT's national control centre. It enables NGT to issue an instantaneous call to the PC to run which then instructs the PLC to start dialling all the generators and instructing them to start.

Furthermore, the system gives day to day and second by second visibility of generator status and availability to NGT. For example we might have sets down for maintenance, or they might be scheduled for a Triad run (see later) and NGT is automatically aware of this and our revised totals of sets available.

During runs Wessex Water's control PC is continuously polling round all sets and reporting the total actual kW output to NGT at 1 minute intervals, soon to be reduced to instantaneous reporting with the installation of the broadband link.

Triads

Triads http://en.wikipedia.org/wiki/National_Grid_UK#Triad_Demand are both a revenue earning opportunity and a way of assessing National Grid high voltage transmission costs.

Triad Demand is measured as the average demand on the system over three half hours between November and February (inclusive) in a financial year. These three half hours comprise the half hour of system demand peak and the two other half hours of highest system demand which are separated from system demand peak and each other by at least ten days. These half hours of peak demand are usually referred to as Triads.

Triad avoidance are a further revenue earning opportunity for diesel generator set owners separate from Reserve Service, where the diesels can earn substantial amounts by reducing a site’s peak demand at national grid peak periods.

This is because the way the National Grid is paid for is by means of a capacity charge, i.e. a charge on kW not kWh. This is levied on the – Centrica, BGB, Powergen, etc. - who then pass it on to their customers in a more or less transparent way.

The charge is calculated in retrospect, by NGT looking back over each of the 17520 half hours and locating the half hours, separated by at least 10 days, of total NGT system maximum demand – which at peak might approach 60 GW. Having identified these Triad period half hours, it then charges each of the energy supply companies according to their average peak loads in GW on the national grid system during those three half hour periods.

For example at the extremity of the system, the Western Power Distribution area in the South West, the total annual transmission cost is about £21,000 per MW per year, again charged to supply companies at the average of their loads at the three Triad half hours.

So, if you can cut your load by 1 MW or start a 1 MW diesel during those periods you can save £21,000 whereas the fuel cost will be a trifling £150.

The snag is that it isn’t easy to predict exactly when the Triads are going to occur, so to ensure Triad capture, Wessex Water starts its generators about 30 times per year for about 1 hour, expending about £3,000 on fuel. Since the triads always occur at times of high power prices, further savings are obtained from avoiding the purchase of power during the same time, which might be about £3,000 which more or less offsets the cost of diesel.

Wessex Water pays for a Triad forecasting service which typically are received at 11.00am for a triad expected at around 5.30pm later that day.

Triads always occur about 5.30 pm on winter weekdays except Friday, between November 1st and March 31st.

Potential conflicts

Wessex Water clearly can’t be running for Triads and third parties when it has taken a capacity payment from NGT to keep its generators available. However Wessex Water’s contracts enable it to declare its generators unavailable during the anticipated Triad periods - so they will call someone else. Wessex Water also declares the status of it generators automatically by the control PC in real time to the third party so they also know when they can or cannot be available.

See also

• Calculating the cost of the UK Transmission network: Cost per kWh of Transmission
• Calculating the cost of back up: See Spark spread

References

Original article content based on a talk given at an Open University Seminar, Milton Keynes, England, 24th Jan 2006 http://eeru.open.ac.uk/conferences.htm#jan06