Bishop Hill Overoptimistic
Nov 18, 2014 Energy: grid Here's an interesting little detail from the National Grid report on capacity margins that I wrote about a few weeks back.
It seems we have windfarms with a nameplate capacity of 7.6GW. National Grid obviously then have to derate this capacity for planning purposes. As we all know, it's perfectly normal for the whole of the UK to simultaneously experience very low windspeeds (or no wind at all), and this has been known to happen even in the depths of winter, for example the very cold winters of 2009/10 and 2010/11. I had therefore assumed that the Grid would have to plan on the basis that they might get nothing from windfarms at all, but in fact they do nothing of the sort. According to Table 16, the grid assumes that they will get 23% of nameplate capacity, or some 1.7GW.
The explanation in the text makes no sense to me at all:
173. For wind power stations, Equivalent Firm Capacity (EFC) is used. This is a measure of how much 100% reliable generation would be required to replace the installed wind generation whilst maintaining the same level of system security.
174. This wind modelling approach combines the risk from wind variability with the other conventional risks, e.g. high demand or low availability of conventional generation. When the installed wind is a small proportion of total generation the main risk to system security comes from the conventional risks and the EFC is very close to the mean load factor for wind generation. In the coming years, as the installed wind capacity grows as a proportion of total generation, the risk that the variability of the wind will affect system security grows as well as wind’s contribution to system security, which will have the affect of decreasing EFC.
This footnote also appears:
Equivalent Firm Capacity is a model output from a Monte Carlo simulation. For further information explaining the theory behind the EFC, see paragraph 173 and http://pio.sagepub.com/content/226/1/33.
Perhaps there is something in that paper that can justify what appears on the face of it to be a rather reckless assumption, but the abstract doesn't fill me with hope:
This paper describes a new probability theory of the capacity value of additional generation in electrical power systems. A closed-form expression for the effective load carrying capability or equivalent firm capacity of a small additional capacity is derived. This depends on the mean and variance of the distribution of available additional generation capacity, and the shape of the distribution of the difference between available existing capacity and demand, near zero margin. The theory extends naturally to the case where the pre-existing background and additional resource are not statistically independent.
The theory may be used to explain and confirm the generality of various well-known properties of capacity value results, as is illustrated using Great Britain examples. Of particular note is the common observation that if the distribution for demand is shifted so as to increase the calculated risk, then the capacity value of additional generation increases. The new theory demonstrates that this is not true in general, but rather is a consequence of the shape near zero margin of the probability distribution of the margin of existing generating capacity over demand.
Incorporating an assumption that we would get 1.7GW from windfarms, the net forecast margin was 2.3GW. Without it then, the margin would be 0.6GW, or 1%. And as noted at the time the National Grid report was published, the figures that were published excluded the loss of the Heysham and Hartlepool nuclear reactors. In other words the true margin may have been negative.
Put that in your pipe and smoke it.
I do hope they are unmothballing a lot of power plants.
Reader Comments (89)
Adam G
"no mention of STOR"
I've certainly yet to hear the BBC mention it - every reference to extra capacity I've heard refers to some mysterious (and unquestioned) 'reserves'. I'm looking forward to seeing (e.g.) Evan Davis's head explode when an interviewee mentions the diesel generators required to cover becalmed windmills.
Does anyone know if the same 'Monte Carlo modelling' applies to solar farms, for those awkward periods between sunset and sunrise?
One can foresee that future weather forecasts will include blackout warnings. No wind tomorrow. All place names beginning with the letters S H I and T will be without electricity. Enjoy.
It's very simple really.
Consider the following self evident propositions:-
1. A reliable efficient Electrical Grid is absolutely fundamental to the very existence of our civilisation.
2. Wind is incapable of being utilised to deliver any despatchable power. Utterly incapable.
I can't remember us having built a power station in twenty years so our despatchable generating capacity is now completely unable to meet the full range of demands placed upon it. Davey's plans to shed our industrial capacity to meet periods of insufficient supply are no use whatsoever.
I predict a major disaster. When I do not know - but there will be a major disaster. Unfortunately I don't have any confidence that the idiots responsible for the mess will be punished.
I debated this at some length with Richard North some three weeks ago, who remained obtuse and insulting on having the main assertion of his post shown to be incorrect (I no longer tolerate his abuse and censorship and have given up even monitoring his blog).
My initial post was this:
(p. 40)i.e. no consideration is given to the possibility of extra demand because of 3GW of unmetered wind (and more other capacity) contributing zero to supply.
There is also an assumption that 1.2GW of demand will be unmet on peak days through Customer Demand Management - essentially contracts for interruptible supply. (para 126, p41) This is additional to the DSBR 0.3GW - a fairly trivial amount which will, if invoked trigger substantial compensation payments.
The assumptions about interconnector flows make no reference to the supply problems in Belgium that are expected to result in rolling blackouts there: we could be in a bidding war and find ourselves obligated to supply the Continent to make good their shortfalls. Zero imports from France and Holland is by no means a pessimistic case for this winter. Figure 20 shows clearly that we experienced strong price incentives for reverse flows last December, when there was no loss of Belgian capacity. Gridwatch data shows this resulted in UK net exports of 3.2GW at the peak via interconnectors.
Table 16 (p 55) shows that wind is assumed to produce at 23% of capacity - an assumption which is maintained in the "arduous forecast" (para 181). This seems to be assuming 2.5GW that would not be there in a winter high cold snap.
The nub of the matter is revealed here:
North extended his abuse and denial:
The scare is being perpetrated by the media, pandering to the Bishop Hillites.
I think I was the only commenter I recognised on the thread as commenting here, so I guess I'm tarred with that worthy brush.
@Dave Ward
"Much of the railways would have been OK, but they are being electrified!"
Even a "Heritage" Deltic or steam loco like Tornado is going to be in trouble if the signaling system doesn't have a reliable back up.
" And who needs radio, radar and landing lights at airports?"
Airports will have back up generators to power essential services, but how many of the approach lights will be included I don't know...
You can't get fuel into the tanks of a diesel locomotive or a plane without electricity. You can't work the points without electricity. And I don't think you can get people on or off an aeroplane without electricity.
In many cases it would be theoretically possible to drive pumps, aircraft docking platforms, signal-boxes or control towers by using generators. But you would probably need modifications to the supply services for them, which would need documentation (from computers) and on-site engineers/mechanics, in a world where transport and communications had suddenly become very difficult.
And generators are a short-term solution. Fuel supply and maintenance parts would rapidly run out if they were to be used 24/7.
It is worse than Andrew suggests, for two reasons.
First, in England, in winter, there is a negative correlation between cold and wind, and a positive correlation between cold and power demand. It is therefore more likely than not that demand will peak on a day without much wind.
Second, demand will be near-peak almost every weekday in December and January. While there will be wind on most of these days, there will be days without wind too.
John Lyon wrote:
"And of course on these cold still days there will be an electrical demand from some of the wind farms to maintain various functions and stop them freezing solid, has a figure for this type of demand ever been disclosed?"
It's not just freezing. Some of the bigger turbines can't be left static for too long as it will induce distortion in the blades, shafts and bearings.
Apparently large ships face the same problem so the propellers are constantly slowly rotated, even when docked in port.
RobertChristopher
Yes, it seems feasible.
But not quite..The proof is that it does not come automatically from private industry.
First of all, 3% capital cost return requirement in a 1% inflatio context, compounded, adds another 100%
Nobody is going to build something to have in 20y exactly the same money as you started with.
One can build 100s of ingenious devices in the fossil fuel economy whereby it turns out that you can just break even to "make energy". And then do not get me started on those contraptions that are "nearly" breaking even.
You could hang miniature windmills behind every fat arse in the country and trap the energy with a light lithium battery. when the fatarsed comes neer a plug it is then wirelessly fed into the grid.
The big point is that you have to
lapogus, sherlock1:
It must have been Oct. 29th that you heard Dr Capell Aris of the Adam Smith Institute say on BBC Radio : "the report found that wind farms generate below 20% of their supposed output for 29 weeks a year, and only exceed 90% of their rated output for 17 hours a year."
The report is here.
Seconded. I wonder if they get their meteorological inputs from the the Met Office?
Postscript to my post above at 2:30 (just missed the editing window) -
The "below 20% .. for 29 weeks a year" appears to be incorrect. The report says it's below 20% for 20 weeks a year, and below 10% for 9 weeks a year. But the way in which it's stated strongly implies that the 9 weeks are part of the 20 weeks figure, and can't be added to it.
Paul asked earlier if the gridwatch templar site was accurate - I believe that it is as Gridwatch simply takes the data from the NETA web site and shows it in graphical form. What is often not fully understood is that NETA, and hence Gridwatch, only shows part of the wind story/capacity/output.
The power distribution system is made up of the long distance high voltage lines of the National Grid (400 kV and 275 kV) and then these are connected to the District Networks of 132 kV or less – down through 33 kV. 11 kV to domestic voltage. There are 14 of these networks operated by different DNO companies. The National Grid can only meter anything which feeds directly to the high voltage system (plus the 132 kV lines in Scotland which for historical reasons are still NG operated).
Anything which happens within the DNs is only seen by National Grid (NG) as a change in load at grid feeder points. Wind generation within the DNs is reflected on the NG data as a reduction in demand. Almost all wind farms in England and Wales feed to the District networks – referred to as “embedded” – and they cannot be “seen” by National Grid instrumentation and hence don't appear on NETA/Gridwatch.
The 8.4GW capacity currently shown on the NETA (largely from Scottish wind turbines) is thus only part of the overall wind capacity - I believe that there is probably another 1.5 - 2.0 GW embedded in the DNs making the overall wind capacity somewhere over 10 GW
According to RenewableUK there is currently 7,774MW onshore wind and 4,409MW offshore wind. Total wind capacity (metered and embedded) is thus currently 12,183MW (12.183GW).
Since NETA gives metered wind capacity as 8,403MW and all the offshore wind farms will be metered, it can be assumed that there is about 3,780 metered onshore wind (big wind farms)and about 4,000MW of embedded onshore (small wind farms and individual turbines).
@Dave Ward: "Much of the railways would have been OK, but they are being electrified!"
Apart from the problems of refuelling locomotives, signalling and working the points, as mentioned by others; points also have a tendency to become iced-up in cold weather, and most of the UK network uses mains-powered electric heaters to prevent this happening.
The way I see it, the only way that wind turbines can make a useful contribution to electricity production is if they are recycled and the fibreglass/plastic chips produced used to power converted power stations
I agree that the wording here is strange. It seems to say that they need 1.7GW of 100% reliable (non-wind or other intermittent generation) to cover for the wind farms. So, what they are saying is that they only 'expect' to get 1.7GW on a regular basis from wind and that they need to have something to cover for this if the wind stops blowing.
I first read it as saying that 1.7GW was what they guaranteed from wind and as many people have pointed out - that is cra... - er unrealistic.
On the whole, it shows quite a low expectation for "base-load" wind generation - or maybe just realism!
Re ZT
PopOffsets is unique — it is the only project in the world that helps individuals and organizations to offset their carbon footprint by improving family planning provision in both developed and developing countries.
The problem is when it is cold, dark and no TV, most couples go to bed. That will cause a problem in the family planning provision.
@ jamesp Nov 18, 2014 at 1:25 PM
Monte Carlo analysis is repeated simulation of a system with varying initial conditions. Indeed it can be used for a collection of solar farms and the analysis output will be somewhere in the neighbourhood of 10% of nameplate capacity, same as the average energy obtained over the year. If the simulation is done with initial conditions valid for the winter season October up to and including February it will be about 2.5%, due to a lot less sun during the whole period.
For wind energy the same is valid, the MC analysis output of 23% of nameplate power is close enough to the capacity factor/average power output of wind energy over the year.
Of course, this value is not a guarantee for every time interval. It is very clear from sites like gridwatch and bm.reports that there are long periods of days and even weeks with wind power delivering less than 5% of nameplate power.
A question would be how long would it take to restore full operation again after a possible cessation in wind?
Often if things are forced to shutdown it takes longer to restart. If wind is intermittent for a period of days a situation may occur where full power restoration is impossible as power never reaches a steady state so to speak? Or is this too worst case?
If not it only takes one major dropout of wind or another supply source to produce a prolonged effect.
I think the gist of the first piece is that, since wind power makes such a pitiful contribution to our power supply, any failings on the part of wind will, by definition, be tiny, since a fraction of what is already trivial will be even more trivial. That's logical, in a manner of speaking. As a justification of future reliance on wind turbines, it doesn't work all that well.
As for this, from the second extract...:
"The theory may be used to explain and confirm the generality of various well-known properties of capacity value results."
Since when could a "theory" "confirm" anything? How do you confirm "generality"? On the other hand, if "various properties" are already "well-known" (supposedly), why do they need to be confirmed by theories and generalities?
I recall a software programme which used to be able to generate management gibberish on demand. Did the geniuses who gave us the ever-conflicting climate models also create an "Auto-twaddle" app, to provide easy input for the likes of Geoffrey Lean?
@ It doesn't add up.
...I debated this at some length with Richard North some three weeks ago, who remained obtuse and insulting on having the main assertion of his post shown to be incorrect (I no longer tolerate his abuse and censorship and have given up even monitoring his blog)...
I also! In particular after a comment that I put in, largely agreeing with a piece of his, was greeted with the single line informative response:
"You don't know what the f**k you are talking about".
Re: Philip Bratby,
> 4,000MW of embedded onshore (small wind farms and individual turbines).
I obtained some figures from DEFRA about how much electricity they generate from renewable sources.
They have 9 x 15kW wind turbines located in various places and they only get around 5% of the nameplate:
Lion House NE66 2PF, 31,347.4kWh from 4 x 15kW turbines = 5.96%
Draycott, 5970kWh from 1 x 15kWh turbine = 4.54%
Teesdale Houset, 8337kWh from 1 x 15kWh turbine = 6.34%
Tolgus Depot, 4613kWh from 1 x 15kWh turbine = 3.51%
Launceston Depot, 7826kWh from 1 x 15kWh turbine = 5.95%
Pevensey, 3975.54kWh from 1 x 15kWh turbine = 3.02%
If that 4000MW is made up of small onshore turbines then they will be lucky to get 200MW from it.
The ECF is a fallacy, it is certainly not a firm capacity like baseload. Monte Carlo analysis does not bring anything useful, the output will be almost equal to the average wind power over the year.
Much more useful would be to predict when the wind energy will be dangerously low. I think that a form of Kalman filtering could be used for this. As a young engineer I was fascinated by this technology. After Kalman filtering the course of a zigzagging ship for some time it was possible to point a torpedo at it and have a significantly higher probability of a hit than with just randomly pointing somewhere close.
If the National Grid, by Kalman filtering of wind speeds, could predict to some extent when the wind might be insufficient for a significant period, there might be time to adopt useful measures to prevent the lights going out.
Philip Bratby You make the statement "Since NETA gives metered wind capacity as 8,403MW and all the offshore wind farms will be metered, it can be assumed that there is about 3,780 metered onshore wind (big wind farms)and about 4,000MW of embedded onshore (small wind farms and individual turbines)." I don't believe that you can assume that all offshore wind turbines will be metered - it depends on the grid voltage level at which they are connected to the grid - some could well be "embedded" in DNs and not "seen" by the grid
When blackouts occur due to the lack of generating capacity caused by all this ridiculousness, it is only fair to refer to them as greenouts.
John de Melle,
"The problem is when it is cold, dark and no TV, most couples go to bed. That will cause a problem in the family planning provision."
Perhaps this is why the government is trying to encourage gay marriage.
I was thinking Black Swan event but that is an unexpected event, seems that the powers that be have created a certainty that the UK will have a major outage event this year.
Just count yourselves lucky that the UK does not have minus 20 to minus 40 weather. I went through 2 total power outages at a hydrocarbon facility.The first time it took 3 months to get everything up and running, second time took 7 months, the crash shutdown resulted in equipment failure and fires.
Monte-carlo, stochastic, simulations are fine but the output must be considered very carefully, the tails of the probability curve can really sting.
If they aren't using zero output for wind as their baseline, they are totally bonkers.
Here is a graph of the last 7 days from the Bonneville Power Administration in the Oregon/Washington area of the US. They have 4500 MW wind capacity installed. Since the 14th of this month they haven't generated more than 300 MW at any time. Currently looks to be zero. Green trace is wind production.
http://i58.tinypic.com/mn019y.png
"I do hope they are unmothballing a lot of power plants."
I hope they don't.
A short sharp electrical shock is needed to bring them to their senses.
If any reader is intending to send Ed Davey a Christmas present I would suggest
"When will the lights go out?" by Derek Birkett (retired grid control engineer)
Published by Independent Minds in mid 2010 ! !
Uh - oh - it is indeed worse than the sterilization teams feared - this could be a disaster for the planet: “It makes perfect sense,” Dr. Licht said, “that during a period of a prolonged loss of electricity, people take comfort in creating their own.”
I discuss this in my paper published by the Adam Smith Institute recently:
http://www.adamsmith.org/wp-content/uploads/2014/10/Assessment7.pdf
I agree with the figure 0f 23 %, and also show how this figure alters as the size of the wind fleet changes.
But I go on to demonstrate how typical winter intermittancy experienced over the last nine years suggests that the firm capacity may, in fact, be as low as 2 % of nameplate.
What NGT say in public and practice in private may well be different things.
This note explains the Grid's operational metered capacity and how it fits in to the overall picture:
https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/65923/6487-nat-grid-metering-data-et-article-sep12.pdf
Details of precisely which wind farms are included are available here:
http://www.bmreports.com/bsp/staticdata/PowerParkModules.xls
A number of the offshore farms are metered.
The abstract for the paper used to model "equivalent firm capacity" reads:
(Bolding additional) i.e. they have used Itô's lemma to find a solution to the stochastic p.d.e. they develop. You can be sure that there is a can of worms in the idea that the pre-existing background and additional resource are not statistically independent. That is, the wind doesn't blow when it's very cold, with very high correlation, but the variation of wind speed with temperature otherwise is less clear cut - now try modelling that with a single correlation coefficient. Orr, as I put it at RIchard Norths blog:
ECF is not firm capacity that can be relied on for planning purposes and the Monte Carlo simulation produces a figure that is closer to capacity factor than any measure of dependable cpapcity during peak load periods. If the wind-turbine capacity factor is in the order of 30% ( which few achieve in practice) then the capacity that can be relied on for planning purposes is less than 10% and could be as low as 2%.
It wasn't till I attempted to read the National Grid report abstract that I realised that they are actually tapping into a new energy source - wordpower. Certainly, you could argue that this example is still fairly obscure, and hence not very powerful, but you get a hint of the possibilities.
Thanks Capell: I did skim your excellent ASI paper when it was published and I filed it away. I shall now read it thoroughly. Thanks very much for all the hard work that must have gone into it.
I first replied to this thread late last night, so I didn't have time to discuss the capacity credit issue in detail.
In my paper I arrived at a figure of 23 % of nameplate output for the capacity credit using the same method that the CEGB used to use.. The last instance of their calculations that I know about is 1985. I describe the method in chapter 9 of my report; put simply, it's the intersection product of two normal distributions for predicted load and predicted generation capacity. (It's easier to understand with a diagram!) The calculation was usually used to assess capacity margin with a time horizon of 25 years.
Then I later go on to explore wind intermittancy and discover that in each and every winter of the nine year study there are periods (in some cases, longer than a week), when wind output falls far below that 23 % figure. In all winters, (and all seasons, for that matter) there are instances when wind production drops away for periods for as long as a week.
Clearly, the calculated capacity credit seems wrong. So why is that?
The problem is that wind does not have a normal distribution of likely production output. The CEGB method relies on both load and generation capacity following normal distributions. In wind generation, it's common knowledge that wind speeds are not normally distributed: they follow a Weibull distribution (a fancy was of calling a positive, skewed distribution). If that distribution of wind speeds is supplied to a wind turbine generation curve you get an energy production curve that has both negative and positive skewness! But averaging power production over the whole country, the upper skewness portion disappears, and then wind production remains heavily skewed towards lower outputs.
Skewed distributions are dreadful to handle. Whereas a normal distribution can be described simply by knowing the mean and standard deviation and the mean, mode and median are identical, with skewed distributiions mean, mode and median are all different.
So one explanation of the disparity between capacity credit and intermittancy is quite simply that the old CEGB method is really invalid for calculating plant margins wat a time horizon of 25 years when the plant mix includes high proportions of wind generation.
Wind power production in the UK has a mode at about 8 % of nameplate capacity. I tested this to see if that was a more realistic capacity credit, but that also fails. If you assume 8 % as a guaranteed output for wind, the intermittancy is such that this will fail. Wind can realistically only carry about 4 % of its nameplate capacity as a guaranteed supply.
Of course, there is a weasel way of twisting these results. If you drop your planning horizon to a few weeks then our confidence in predicting wind production increases. In fact, we can then probably assign a normal distribution to wind production and the capacity credit will become higher. Trouble is, you can't build even a new gas plant in a few weeks!
Being my area even just large scale spikes and drops in electricity could have big knock on effect to IT. Most will probably run under UPS but a key one failing could have big knock on effects.
a couple of weeks ago we had a blip in central York. All the ticket machines rebooted. Not sure how people having tickets printed got handled?
Being my area even just large scale spikes and drops in electricity could have big knock on effect to IT. Most will probably run under UPS but a key one failing could have big knock on effects.
a couple of weeks ago we had a blip in central York. All the ticket machines rebooted. Not sure how people having tickets printed got handled?
Re Dodgy Geezer 10:14
"@omnologos
...Or in other words...do they imply there is an optimum amount of wind power to install, and that to install more of it would actually make things worse?..
I don't know whether they are implying it or not, but it is perfectly true, and has been recognised in some quarters since at least 2007."
FYI
http://joannenova.com.au/2011/07/lessons-in-wasting-money-use-more-wind-and-solar-and-emit-just-as-much-co2/