Value chain · CLM-1487
In plain terms
In Britain, the amount of fast back-up power kept ready depends mainly on the biggest single block of power that could be lost at once; the rules do not set it by fuel. So a space solar station would need about the same fast back-up as a gas power station whose biggest single loss is the same size. A 2 GW space plant would be a bigger single loss than the 1,800 MW Britain now plans for, if one fault could take all of it. A satellite in geostationary orbit also passes through the Earth's shadow each night for about six weeks around each spring and autumn equinox, for up to about 70 minutes. Those times can be predicted exactly, and other power stations would have to cover them.
The claim as recorded
Great Britain's transmission standard requires the system to be run so that the fault outage of a single generating unit, a single power park module or the most onerous loss of power infeed causes no unacceptable frequency conditions, with the system operator's Frequency Risk and Control Report deciding which events this covers, and the system operator's 2025 frequency report says response holdings are adjusted "to ensure that the largest infeed loss can be accommodated". The OECD Nuclear Energy Agency's 2019 review of system costs states that for dispatchable plants "the amount and thus the cost of operating reserves are generally given by the largest contingency in terms of the largest unit (or the two largest units) connected to the grid", and it puts nuclear's small balancing cost down to nuclear plants being the largest on the system. On these pages the fast reserve for a sudden loss is set mainly by the size of the largest infeed loss, and none of them sets it by fuel; on that rule a space solar plant and a gas station with the same largest infeed loss would need about the same fast reserve (our reading of the rule, not a comparison any source makes). The standard caps the infeed that one fault on a generation connection may disconnect at 1800 MW, and the system operator plans for a largest loss of 1800 MW, so a 2 GW plant would be the larger single loss if one fault could remove all its output. A geostationary plant also has a scheduled loss that a gas station does not have: up to about 70 minutes a night for about six weeks around each equinox. The US Electric Power Research Institute, reviewing the 5 GW reference design, said in 1980 that this outage is "exactly predictable" and that during it the satellite's 5 GW "must be replaced by reserve units".
Evidence behind this: T3, graded at the weakest cited source: SRC-242 (the system operator's report), SRC-796 (an OECD agency's review of system-cost studies), SRC-449 (programme proceedings) and SRC-549 (a NASA centre's in-house report). SRC-238 is T1, the GB standard read for its own requirement. That the same fast reserve follows for the same size of loss is a reading of the rule's form; no held page compares space solar with gas.
Scope and limits
The standard and the system operator's report do not mention space solar or gas by name; the reading that equal loss sizes need about equal fast reserve is ours, from the form of the rule. The system operator also weighs how often each loss happens against the cost of securing it, and no held source gives a trip rate for a rectenna or a satellite, or for a GB gas station, so "about the same" is the most the sources allow. A 2 GW plant is above both 1800 MW figures. The standard's cap applies to a fault on the grid connection, and splitting the connection meets it (CLM-0780). It does not cover a fault on the satellite or its beam, and held records describe faults that remove the whole beam at once (CLM-1081, CLM-1478). Whether a space plant's largest infeed loss is its whole output is therefore open. No held source gives the unit sizes of a GB gas station. The FRCR sentence comes from a case study of one event in 2025 and states practice, not a rule. The Nuclear Energy Agency states its rule for dispatchable plants and for unforeseen outages; treating a space solar plant as such a plant is our reading. Its statements are a review of other studies, written by an agency whose remit is nuclear energy; its balancing cost for nuclear is printed in a unit per MWh of variable renewable output, so it is not used here as a figure. The 1980 statements concern the 5 GW US reference design and US utility practice of that time; PG&E's 6,000 MW is its own result for its own three criteria. The eclipse figures are geometry (CLM-0895). NASA's 1977 grid study judged, for the case where the outage falls near midnight, that other plant could carry it because demand is then about half its peak. It put the outage at fifteen minutes at most, against up to about 70 minutes by geometry, it limited the number of rectennas each regional grid could take by how far its other plant could throttle, and it wrote for US grids. The 1980 utility research paper says the midnight argument fails if load management flattens daily demand. It also says an SPS unit's reliability was then unknown, and that SPS in place of large synchronous machines leaves less system inertia; no held page works out what that does to the reserve. No held source gives a capacity-market de-rating for space solar (gas: CLM-1470), or shows a beam moved to another site to cover an outage (CLM-1474, CLM-1475 (low confidence)). Shutdown and restart times would lengthen each eclipse outage (CLM-1081).
What it assumes
That a rectenna connected to the GB transmission system is a power park module or other generation connection under the standard, with no special treatment (as CLM-0780 assumes). That the plant delivers to one ground site, so its eclipse is not covered by another satellite.
Where it comes from
PDF p.28 (printed p.28), clause 5.1, read at a zoom 3 crop: "The onshore transmission system shall be operated under prevailing system conditions so that for the secured event of a fault outage on the onshore transmission system of any of the following:" with item 5.1.2 "a single generation circuit, a single generating unit (or several generating units sharing a common circuit breaker), a single power park module, or a single DC converter; or" and item 5.1.3 "the most onerous loss of power infeed; or"; and, among the outcomes that there shall not be, 5.1.7 "unacceptable frequency conditions;" (text layer). PDF p.31 (printed p.31), clause 5.8: "The ISOP shall use the latest version of the Frequency Risk and Control Report as consulted on and approved by the Authority to determine the events for which unacceptable frequency conditions shall not occur." Clause 5.11.2 allows exceptions to 5.1.7 "based on the outcome of an assessment conducted in accordance with the Frequency Risk and Control Report". PDF p.79, definition of Unacceptable Frequency Conditions: "The Frequency Risk and Control Report will define what is considered reasonable, infrequent and tolerable for each of these criteria for transient frequency deviations." PDF p.102, Appendix H.2: the report "will set out which of these frequency risks the system should be secured against by the ISOP in their operation of the system to allow a balance to be struck between the consideration of risks, the benefit of avoiding these risks materialising, and the economic and efficient costs that will be incurred in ensuring the safe and secure operation of the system to do so." PDF p.15, clause 2.6.3: "following a fault outage of any single generation circuit or single section of busbar or mesh corner, the loss of power infeed shall not exceed the infrequent infeed loss risk;" PDF p.62, Infrequent Infeed Loss Risk: "From April 1st 2014, this is a loss of power infeed risk of 1800MW." PDF p.64, Loss of Power Infeed: "For the purpose of the operational criteria: i) the loss of power infeed includes the output of a single generating unit, CCGT Module, boiler, nuclear reactor or import from an external system via a HVDC Link." SRC-242 PDF p.42 (printed p.42), appendix 9.1.1, read at a zoom 3 crop: "Response holdings would be adjusted according to the inertia to ensure that the largest infeed loss can be accommodated." SRC-796 PDF p.19 (printed p.17), executive summary, read at a zoom 3 crop: "In the case of dispatchable plants, the amount and thus the cost of operating reserves are generally given by the largest contingency in terms of the largest unit (or the two largest units) connected to the grid." PDF p.65 (printed p.63), literature review, read at a zoom 3 crop: "Electricity systems must always maintain cycling reserves according to the N-1 criterion, which means that the system must be able to continue supplying the full load, even if one plant trips. Logically, these cycling reserves are calibrated on the largest plant in the system, which happens to be nuclear." The same paragraph says these balancing costs for nuclear are below EUR 1 per MWh (the page prints the unit as per MWh of variable renewable output). SRC-449 PDF p.600 (printed 579), read at a zoom 3 crop: "The use of an SPS unit as baseload generation suffers from one major drawback involving the outage during eclipse periods whose time duration and time of occurrence during the year are exactly predictable. During these periods, a major block of baseload generation (5 GW) is unavailable and must be replaced by reserve units." Earlier on the same page: "The actual reliability of an SPS unit is clearly unknown at this point in time, but let us assume for the moment that its reliability is very high." Later on the same page, of the argument that eclipses fall near local midnight when demand is low: "This argument loses its validity if one envisions a future scenario in which widespread usage of load management programs has flattened out daily demand profiles compared to their historical form." PDF p.601 (printed 580): "Furthermore, the addition of SPS units to a national power network in lieu of large conventional synchronous machines will result in less system inertia associated with the rotating mass of conventional machines and further complicate the stability problem." SRC-449 PDF p.608 (printed 587), read at a zoom 3 crop: "For the three planning reliability criteria of: 1 day in 10 year Loss of Load Probability, 12% of monthly peak load, and two largest risks, a 5 Gigawatt (Gw) SPS in conjunction with another 1,000 Megawatt (Mw) unit could require 6,000 Mw of reserve capacity, which would predominate until monthly peak load reached 42 Gw." SRC-549 PDF p.153 (printed IV-D-13), read at a zoom 3 crop, of the case stated as "In one case, the outage occurs around mid-night for a 45-day period around each equinox.": "An outage at this time of day has a minimum impact on grid power management. The reason being, when this outage occurs the grid demand is only about 50 percent of its peak demand. For the period of the outage (fifteen minutes maximum), the demand could be handled by other capacity on the grid." (the text layer prints "mintmum"; the crop shows "minimum"). CLM-0895: up to about 67.5 minutes in full shadow and 71.7 counting partial shadow, in seasons of about 44 to 47 days. CLM-0793: the system operator's minimum inertia figure rests on a future largest loss of 1800 MW.
Basis: MW of a single loss of power infeed; no cost figure is used
Related records
- Answers question
- QST-021 not published in this release
Cite this
energyaperture.com/c/clm-1487
CLM-1487, Energy Aperture, as of 2026-05.
Connected records
Other published records linked to this one. A list, not a judgement.
- Also from SRC-238
- CLM-0780The UK rules for connecting power stations set a hard limit on how much generation one fault may disconnect at once: 1800 MW for a fault on...CLM-0782UK rules limit the overhead line between a generating unit and the grid to 5 km for units producing at least 2000 gigawatt-hours a year and...CLM-0786In December 2025 the UK system operator changed its own connection standard to allow a single large direct-current converter to carry more...CLM-1211Britain's main transmission grid runs at 400 kV and 275 kV, and parts of the 132 kV network in Scotland. On the 400 kV network the voltage...
- Also from SRC-242
- CLM-0793Britain's system operator plans to run the grid with less spinning inertia, and it sets that limit from the size of the biggest single...CLM-0795Britain's system operator says that at the lower inertia level it is moving to, losing 2000 megawatts at once would be enough to start...
- Also from SRC-449
- CLM-0964To build two of these satellites a year, about 300 tonnes of material had to arrive at the orbital construction site every day. Getting it...CLM-0974The 1978 reference design shows every loss between sunlight and the power grid as a chain of numbered boxes, and the boxes multiply out...CLM-0975Depending on which published assumptions you take, a satellite needs between about 2 and 6 square kilometres of solar blanket for every...CLM-1058The often-repeated claim that the magnetron is 188 decibels quieter than its carrier traces to a conference abstract in which the tube...CLM-1059By spring 1980 the engineers, the reviewing panel and the ionosphere experimenters were all saying the beam-intensity ceiling the whole...CLM-1060The 1980 review did not agree with itself on how much of a satellite's cost was rockets: one paper said over a third, the programme's own...CLM-1061In one 1980 volume the satellite cost anywhere from about 2,300 to 3,800 dollars a kilowatt depending on who added what - and the programme...CLM-1062In 1980 the launch vehicle that everything was costed on was said by its own designers to be arbitrarily big, and they proposed a rocket a...More on the source page.
- Also from SRC-549
- CLM-1089The higher industrial band would have shrunk the receiving field to a fifth or a third of its size, at the price of a satellite that could...CLM-1409A 1977 NASA study of how to hold up the receiving station's panels was told the ground underneath had to stay farmable, with a metal mesh...CLM-1486A 1977 NASA study of how receiving stations would join the US power grid looked at two kinds of shadow. For about six weeks around each...
- Cited in articles