Monday, July 27, 2009

Wouldn't it be better? No, I don't think so! (Hot Air Oscar nomination)

The website encourages people to be aware of the carbon intensity of the grid, saying "Wouldn't it be better if we could use power when it's greenest?". I am pretty sure that the answer to this question is No!
Imagine, for the sake of simple discussion, that we have a country in which on average half the electricity comes from baseload nuclear power (intensity, 20 g/kWh) and half from demand-following gas (470 g/kWh). And that at night, demand is 60% of the average, and 83% of the electricity comes from nuclear. And that in the day, demand is 140% of the average, and 36% comes from nuclear.
Under these assumptions, the nighttime grid intensity is 95 g/kWh, and the daytime grid intensity is 310 g/kWh.
People using the RealTimeCarbon service will be advised by the red flashing "carbon alert" icon to avoid using electricity during the day, and will be rewarded with feelings of green smugness if they go to great lengths to use electricity at night instead. They may delude themselves into claiming that they have reduced their carbon footprint. You could even imagine them selling carbon offsets based on this sort of electricity-consuming time-travel. But, in the cartoon world that I have just described, the time at which you use electricity makes no difference at all to the carbon emissions! Imagine that 1000 people all earnestly follow the RealTimeCarbon guidance and turn on their 1kW toasters in the middle of the night instead of during the day. What happens? Well, in response to the increase in demand, an extra 1MW of electricity is generated while their toasters are on; and this electricity (in my cartoon world) comes from the gas power stations being turned up just a little bit, whether they turn their toasters on at night or in the day. The true marginal impact of their consumption is 470 g per kWh, whenever they consume.
Now, I am not saying that this cartoon is a faithful representation of what's going on in the UK. Maybe in the UK there are some times of day that are "good" times to use electricity, and others that are "bad". But I think this cartoon proves that "knowing the grid average" doesn't tell you anything useful about that. And I think that the cartoon is a fairly good cartoon of the UK, since in the UK much of the really low-carbon electricity is wind and nuclear, both of which are (at present) not demand-following.
Moreover, I think that if people go to great trouble to check RealTimeCarbon for guidance on "when it is ok to consume", the end result may be a worsening of the UK carbon footprint! Here's two arguments why:
(1) I can imagine people inconveniencing themselves in order to switch on their equipment at night - their inconvenienced lifestyle may well use more energy (for example, when they wait up late for the RealTimeCarbon to go from red to green, they may keep the lights on for longer at night!);
(2) If people think that their electricity is "green" they may give themselves permission to consume more of it. (I know some people argue, for example, that "their electric car is powered by wind, therefore they can drive as much as they want, and it doesn't do any harm to the planet".)
I therefore nominate RealTimeCarbon for a Hot Air Oscar for "Best intentioned but most useless consumer-engagement".

Thanks to Kim West for pointing me to the website and asking questions.

PS - I posted a message on the realtimecarbon forum 3 days ago, querying another aspect of their methodology, and there has been no response.


sourcerror said...

Sorry, but I have to disagree with you. More uniform energy consumption levels are "greener" because a higher percentage of it can be produced with nuclear. And even if we used a bit more of it, it's still carbon neutral.

But anyway, I wouldn't change my daily routine because of that.

DanH said...

What if nuclear power stations are associated with pumped storage of imperfect efficiency? That way, nuclear is partly demand-following, but its whole-system efficiency depends on the temporal distribution of demand. I don't have any kind of quantitative handle on how big the effect is, though.

Damon said...


I did think about this quite a bit while constructing my page and I think that the electricity is indisputably 'greener' if *both* the current intensity is low *and* no (pumped) storage is being used to meet demand which implies losses.

Unless that storage is filled from entirely zero-carbon sources that implies extra emissions.



Joshua said...

I think the oscar nomination is harsh. demand shifting could save money and potentially carbon by creating a closer peak:base match - might not deliver that result immediately, but understanding temporal shifts in grid carbon intensity is still an important component.

in fact, it's also good bullshit detection, since it should finally stop people quoting spuriously high electricity offset emissions factors for things like CHP... for example a few years back, all the emissions savings calculated for Woking Borough Council assumed that every unit of power generated from CHP would offset a unit of coal-fired generation from the grid at 860gCO2/kwh.

gavin said...

Hi David

Firstly I wanted to say thanks for writing - one key reason for launching this service was to stimuate exactly this kind of debate.

I think your "oscar" analogy is funny (if unwarranted) but the purpose of the project is linked to your own talks and writings - how do we increase our knowledge and understanding of the space.

I'm not going to comment here on the impact of demand-response vs grid-load vs peak/trough vs engagement. That's a longer post.

Government policy and commercial investment and is being driven into the "cleantech" space. This has an extremely long way to go before it becomes meaningful. The CRC is a great example.

Raising the bar on transparency, bringing more data to market, fuelling debate are all crucial if we're going to make any sensible headway at all. Helping this spread into the public sphere is also important if we want to bring about change.

That's the point of

The realtimecarbon guys are working up responses to your questions, which I have no doubt will include more questions: which is why realtimecarbon has a forum(!)

Please help us and the community at large engage in this important debate.

I'd also be happy to host a session in London next time you're down with a range of opinions and expertise - would you like to get involved?

(Finally one niggle: complaining about not getting a response until Monday - to a message you posted at 5:30pm on a Friday - is a little precious, but no matter.)

Jamie said...

Hi David,

I am one of the people behind Whilst we're pleased that you've engaged in the debate we aimed to stimulate, it seems that there has been a misunderstanding over our objectives for the project.

It is very much a first stab at tackling a complex subject, and we are not attempting to be any authority advising people on how and when to use electricity. Sorry that we didn't respond to your forum post immediately, but we do plan to ensure that it houses the full spectrum of analysis as the project moves forward and we will include our below response there when we get the time (probably next week).

Right now in the UK, your argument is valid. However, our project is attempting to look to the future as renewables (hopefully) propogate on a significant scale. As we get more uncontrollable (i.e. renewable) sources of generation on the system it will be increasingly important to have flexible load that can be shifted to times of maximum availability and away from times of high demand.

All gas stations are not equal - 470g/kwh is an average value - there will be times when the least efficient gas stations are being turned up, or coal stations are, or even worse oil stations are being drafted in to catch the peak - this will become more obvious in winter when we are using more energy overall and therefore more marginal plant come into service. As we improve our methodology we want to be able to differentiate the values further using plant efficiency data.

We created the website to stimulate debate and make it clear that responding to demand (and at a national level, shaving peaks so that baseload supply can be cleaner) is an important concept to start thinking about. We certainly won't be linking the figure to offsetting or otherwise engaging in practices that dupe customers into thinking they are being greener if they are not.

The ultimate target market for this data is not individuals 'waiting up to switch the toaster on' but products and energy processes that can react automatically to the signal. The 'customers will use more' argument is also missing the point that this will become automated over time - machines and building management systems will be controlling demand for us.

It's also important to note that coal generation is the primary method that is used in the UK to match supply to demand, more so than pumped storage. Having said that, we agree that a different value for pumped storage is needed. Now that we have the ability to average the carbon intensity of the previous night's electricity we are working on ways of incorporating this into the algorithms driving the feed.

Please do continue to write about the site as it evolves, and we will do our bit in keeping the debate going.

Kind regards,

Jamie Andrews

Damon said...

Jamie: this may be of interest to you though the data only goes back to last November...

The variation by month in intensity mean/spread is clear, though much of it is a consequence of having a smallish near-fixed zero-carbon baseload of nukes at the moment of course:

As you say, dynamic reaction to availability of intermittent sources will be more important as more wind comes on line: nominally more than summer minimum demand...



dave said...

David (MacKay) - you bodly assert that when extra demand is added at night, a CGGT power station is 'turned up'... but can you provide a link or evidence for that? Clearly crucial to this whole discussion is the answer to precisely this question, viz. "what meets EXTRA demand when 'carbon intensity' seems to be low?". This is a question of how the grid is controlled. I imagine if the extra demand really is tiny (less than a few kW) then it is simply met by spinning reserve (rotational inertia of generators / a very very tiny speed change). If more, then some kind of feedback control system will kick in and turn a generator up or turn a load off. Perhaps someone can enlighten us? I can only see a 'carbon benefit' if my extra demand is met by either a controllable low carbon source (eg waste CHP plant) being turned up, or a controllable load (eg pumped storage that is in 'pumping' mode) being turned down/off. Otherwise your argument would seem to hold.

dave said...

anyone willing to engage with my comment?

Felix said...

I am a great fan of Without Hot Air and its philosophy. However, the example given in this post is arbitrary since it assumes that the marginal carbon intensity is constant. Clearly, that is not the case. Power plants are diverse and with the correct incentives in place, power companys should use their most efficient plants the most and their least efficient plants the least, i.e. only as peaking plants. To my knowledge, as a general pattern this is what they tend to do in practice (although imperfectly since there are some adverse incentives, too). Hence marginal carbon intensity in peak hours is higher.

So: it most certainly does make a difference whether your use of the toaster implies firing up a dirty coal peaking plant at about 1000g/kWh or a 500g/kWh gas plant when the grid is less busy.
Unfortunately, I'm not an energy expert and do not have any actual data on marginal carbon intensities in relation to demand - maybe MacKay knows?

Admittedly, those considerations probably wouldn't save you much carbon today but, as previously commented, load management will become more important as the fraction of less reliable renewable power increases along with increased power consumption by heat pumps and electric cars which are more flexible about the time at which they suck power from the grid.

The main problem with, I believe is that it calculates average, and not marginal carbon intensity (which I guess would be more difficult, too). However, unlike what MacKay's example suggests, I'm pretty sure that the two are usually correlated.

Adjusting one's consumption patterns to realtimecarbon currently wouldn't really make a difference but giving the Hot Air Oscar to those people instead of one of the thousands of bullshitters out there seems a little harsh.

PhilJ said...

I agree that a hot air oscar for this attempt to get to grips with a complex and important issue is way off beam. If you choose a cartoon example where on- and off-peak marginal generation both come from gas then, yes, there is no benefit to shifting demand, but if coal generation is marginal on-peak and gas is marginal off-peak then there is a carbon benefit from shifting. And, as stated, there could definitely be benefits to demand shifting in future to best utilise renewable and nuclear generation. Realtimecarbon is a great effort at trying to start thinking about these issues - the designers deserve to be engaged with in a more constructive way please Mr. CSA.

matic pajnik said...

Don' forget that the nuclear power plants are switched on to 100% all the time, and that you can't regulate by the moment, It needs time. That means that the toaster is using gas in any case!

sourcerror said...

@matik pajnik:
"That means that the toaster is using gas in any case!"

Excess electricity is used to heat water in the boiler in your bathroom (at reduced prices). On that wire you only get electricity when there's a surplus. (At least that's how it works in Hungary.)

James Toal said...

Dave, you asked for techie answer to what happens when demand rises. As you said, it's all a matter of scale. If it's a single toaster, the GB system frequency will drop ever so slightly (rotors will slow down, give up some kinetic energy, and somebody's toast will be a nice golden brown). AKA: GB frequency control is an open-loop system
But if lots of people log on to and respond as intended to its info, the results will depend on how predictable that is (the grid operator can log in too). If (s)he antcipates their repsonse, (s)he could disaptch a low carbon ngenerator (assuming it's available and capable). But if it's not, or not expected, I'm afraid cutting back on p/storage units pumping, or instructing fast-response o/cycle gas turbine is about all that's left to play with. Everything else is either non-dispatchable (wind, tidal), at ceiling (wind, nukes, all renewables) or too slow to respond before the next blackout(all rankine-cylcle thermals, biogas, waste-burners, etc).