Solar power correlating with peak demand - a myth?

[LancedDendrite]

One talking point that I often hear from pro-renewables folk (especially pro-solar) is that:

Peak output for electricity is during the day in summer, when everyone is running AC [air conditioning]. Using solar is actually ideal for this, as solar is usually especially effective when you most need ac, when it's really sunny.

Now, I happen to live in Australia - a particularly sunny country, as it would so happen. So I decided to have a look at some hard data - demand statistics from the Australian Energy Market Operator, which is the ISO for the National Electricity Market.

For the purposes of this analysis I looked at the load curves of certain days to see if peak electricity consumption did coincide with high temperature days or have much seasonal variation. I'm going to assume that the type of solar power in question is solar PV installed on residences, as it is the most popular and discussed type of solar power.

Now, the NEM is separated into regions (states), so I decided tp have a particular look at the demand statistics for my home state, Victoria. Some notes about the demand profile of this state:

• High penetration of natural gas usage for heating (due to state encouragement to use domestic natural gas from Bass Strait), so presumably there is less demand for electricity for heating.
  
• Local generation is mostly baseload brown coal - has had problems with responding to high-demand periods in the past
  
• Due to being the southernmost mainland state in Australia, it has some of the lowest temperatures on the mainland but is still capable of reaching very high temperatures during summer.
  

Firstly, I looked at a semi-random sample of consumption data throughout a day-long period in both winter and summer. For this, I picked the second day of July (winter) and December (summer):


(sorry about the graph size, I don't know how to resize the images using the forum software)

The average temperatures during those days were 13 °C (9 to 16 °C) for Saturday 2nd of July and 16 °C (11 to 20 °C) for Friday 2nd of December. Note the day of the week that each one was on - even on a weekend, demand was still higher during the day in winter!

So, these data sets were fairly close together for temperature, so I decided to look at extreme differences in temperature - the hottest and coldest days for 2011 in terms of peak temperature. These turned out to be 41 °C on 31st of December 2010 (close enough to 2011 to count, I reckon) and 3 °C on 27th of July 2011.
So, what did the load curves look like for those days?



The peak load was quite close - within 200MW for each of those extreme days! And look at that massive change in load profile during that extremely low temperature day.

So, high temperatures do induce extra demand, and the highest demand on that day is almost 100% of the lowest demand on that particular day for this data set. However, so do really low temperature days, and I can guarantee that in Melbourne, those days will not only be bloody cold but also fairly overcast.

This seems to suggest that the use of domestic solar PV as a demand reduction technology is a highly specific one, unless you have plenty of storage and backup. Of course, you could put your solar panels/solar thermal towers in another part of the country that gets more sun on those days, but then you start running into transmission losses and so on.

In addition, if you assume near-constant insolation on the high temperature day (this is fairly reasonable for Melbourne on those sorts of days) then the solar PV won't be able to fully cope unless it's overbuilt to handle that peak. So you'll still need a fair bit of flexible/dispatchable generation to handle that demand curve, especially in winter.

Now, the big flaw that I'll admit to in this analysis is that I don't have insolation figures for the days in question - I've merely made an assumption that in Melbourne, high temperature days are sunny and low temperature days are cloudy (from my anecdotal experience). If anyone knows where I can find those, I'd like to know.

So in conclusion, the use of solar power as a demand reduction method is only applicable in very specific circumstances and is not very useful in actual peak demand periods, which are during the winter months. And if it can't really handle these peak periods on its own, how will it deal with supplying baseload power without an overbuild or a lot of independent backup?

So, questions/comments/accusations of cherry-picking data? :-)

[sod]

the situation is not trivial.

the correlation between temperature and power demand shows a "U"-shape.

ars.els-cdn.com/content/image/1-s2.0-S0360544205000393-gr2.jpg

the specific shape of that curve is different in different countries. solar is mostly a solution to the "upwards" part of that curve.
(you get many different versions with a google picture search for *correlation temperature power demand*)

better buildings and small changes, like LED lights will help with winter demand.

whether there is a good fit between solar production and summer afternoon demand peak, also depends on the country.

for Germany, the fit is next to perfect: (real daily numbers!)

www.transparency.eex.com/de/

in warmer countries, peak demand tends to happen after peak solar supply. this was discussed in the "100% alternatives" topic. there is an interesting paper with a nice graph at the end:

www.eap-journal.com/archive/forthcoming/forthcoming_Tim_Nelson_Paul_Simshauser_and_James_Nelson.pdf

but even there, pure solar will have a massive impact and only a few hours of storage will help to make another perfect fit. (electric car should be home, half loaded (via the grid) and parked, when its peak demand time...

---------------------

if we take summer peak demand as double of summer base (and similar winter numbers), we see that solar could supply up to 25%.

it also eases the strain on the grid, as base load plants can run at best level. (this is the opposite of what is constantly said about alternative energy, but look at the graph above for Germany!)

[sod]

i forgot, why i think that solar is an important solution and why the summer peak is more important than the winter one:

growing power demand will be mostly in warm countries and a significant part of that growth will be caused by air conditioning. (just an example from Hong Kong in the past: scroll down to "Fig. 3. Annual electricity use for air conditioning from 1971 to 1996.")

www.sciencedirect.com/science/article/pii/S0196890400000182

[Graham Palmer]

Muriel Watt (UNSW) looked at this about 7 years ago,

www.ergo.ee.unsw.edu.au/value%20of%20PV%20in%20summer%20peaks.pdf

More recently, IPART considered the issue:

www.ipart.nsw.gov.au/Home/Industries/Electricity/Reviews/Retail_Pricing/Solar_feed-in_tariffs/14_Mar_2012_-_Final_Report/Final_Report_-_Solar_feed-in_tariffs_-_March_2012

The conclusions are always the same, for example:

Page 68 and 69:

Based on this analysis, Ausgrid concluded that “there appears to be no case within Ausgrid’s network where it is economically feasible to defer network investment due to the presence of embedded small-scale solar generation”

Also of interest

In areas with high penetration of PV the networks are experiencing issues in maintaining appropriate voltage levels, and this is likely to have design and cost implications.

The conclusion is simple: the diurnal cycle of solar provides a generally good match with demand however solar cannot be guaranteed to be available during critical peak demand periods, therefore has little or no value in deferring network investment. Closer examination of the critical peak demand periods shows that peak demand on the hottest days occurs in the early evening. In other words, solar PV can reduce fuel load and therefore emissions while it is sunny but without battery storage, cannot displace fossil fuel generators.

[LancedDendrite]

Jun 11, 2012 5:19:38 GMT 9.5 @sod said:

the situation is not trivial.

the correlation between temperature and power demand shows a "U"-shape.

ars.els-cdn.com/content/image/1-s2.0-S0360544205000393-gr2.jpg

the specific shape of that curve is different in different countries. solar is mostly a solution to the "upwards" part of that curve.

If you follow that curve, demand should be much higher during a heatwave day like the 41 °C day I plotted the load curve for. However, the lowest temperature day was also a high demand day of a similar scale. So hot days are only a subset of the high demand days on the electricity grid. That's what I'm saying - it's a bad investment because it costs too much to address a scenario that is not as bad compared to others.

better buildings and small changes, like LED lights will help with winter demand.

Better buildings can work, yes. But LED lights? They reduce the instantaneous demand of a light bulb from something on the order of 80W for incandescent lights to around 9W. There's even less savings if you've already got fluorescent lighting installed - you'd go from 18W to 9W. Those are nice percentage reductions in demand. However, do they compare to heating demand?

Now, air conditioning systems can be rated from as low as 2.5kW to up to 18kW! That's two orders of magnitude more demanding as a light bulb - if you look at something like a 3 bedroom single-storey house, you'll have perhaps up to 12 lights in the place. That means that you're consuming at most around 1 kW using incandescent bulbs, and if you switch from fluoro tubes to LEDs then you're going from around 220W down to 110W.

So changes in lighting efficiency aren't going to make much difference compared to changes in heating and air conditioning.

it also eases the strain on the grid, as base load plants can run at best level. (this is the opposite of what is constantly said about alternative energy, but look at the graph above for Germany!)

If you look at that demand curve, it really doesn't help baseload power unless you start storing or curtailing the solar PV. You'll need flexible generation to steadily ramp up over that ~10 hour period. The links that Graham Palmer provided below say as much - solar PV doesn't help out at peak times of network congestion without some storage being added. And remember, storage is an added cost and quite expensive at the moment.

[anonposter]

The BOM does have daily global solar exposure figures on their web site which give (at the Botanical gardens site) on July 2, 2011 a total of 10.2 MJ m^-2 and on December 2, 2011 a total of 33.6 MJ m^-2 (both those days appear to have been pretty sunny for the season). The lowest temperature day of July 27, 2011 got 10.9 MJ m^-2 (pretty sunny for winter, though the coldest days do tend to have no clouds to insulate) while December 31, 2010 (which actually shouldn't count) got 27.3 MJ m^-2. The variation last year was from a highest value of 35.0 MJ m^-2 on January 4 to a low of 2.5 MJ m^-2 on July 30.

[proteos]

Here are a few homemade graphs to show how solar PV works in Germany. Solar is orange, Wind yellow, the rest blue, green is the trade position.




For the record, nuclear power with 10GW through the day produced more energy the 25th and 26th of may, which were record days for solar PV (22GW peak). Of course it's also the case for lignite, the worst fossil fuel (13GW). And Germany will exit nuclear before it exits lignite: it's a suboptimal choice.

So back to the subject: solar production is roughly correlated to the increase in demand on late spring--summer--early fall days. the correlation is not perfect though.
a) it's only valid today when the sun shines over the whole of Germany
b) the peak of demand occurs before the solar peak. It is reflected in the trade balance: Germany exports after noon
c) solar power does not take off early enough. Again the trade balance is the clue: Germany imports early in the morning.

Here are a few thoughts:

• The summer intraday peak is not the winter peak: it's lower, and does not occur at the same time in the day. It's around 11am in summer and 7-8pm in winter. Solar's no help in winter, yet the winter peak is much sharper in Northern Europe
  
• If PV installation continues at the same pace, Germany will have problems with excess power. First it could be solved by turning off baseload plants. But after sometime it will exceed consumption and export capacity. What will happen then, with no storage in sight?
  
• From where will the juice come at night? As everyone agrees, solar PV does little to decarbonize baseload.
  
• Solar production also has an impact on spot prices: it lowers them (because of the 0 marginal cost of power). How solar PV is to pay for itself if the value tends to zero when the production is maximal?
  

[sod]

thanks for those graphs, cool work!

i think it is obvious that solar will have little effect on winter peaks and basically none on night time. but the topic of this post is: "Solar power correlating with peak demand - a myth?"

and this can be definitely answered: in Germany, the correlation is close to perfect in summer.

ps: nuclear generation might have been higher over all day on the 15th/26th of May, but solar was producing at the time when electricity is most needed and most expensive, while a significant portion of the nuclear power was cheap night time tariffs. 8and demand at night will even fall more, when peak prices come down...)

[anonposter]

Solar can only really work in the summer (in Melbourne the typical winter insolation is only around a third of summer (10 MJ m^-2 compared to 30 MJ m^-2)) yet we need power in the winter, a peaker which only worked in the day wouldn't be too big a deal given that the peak is during the day, but not working in winter (or working at only one third capacity (best case)) really isn't good enough because then you'd need something else to handle winter and there isn't much that works better in winter than summer (thermal plants (at least of the non-solar variety) do work a bit better in winter but not 3 times better).

Things worse than nuclear power also has an analysis on what has been happening in Germany.

[david jones]

A major part of supply infrastructure cost is associated with peak demand. By peak I mean the absolute maximum demand which might occur rather than typical daily maximums. If you compare the average daily demand profiles for solstice months June and December in the Victorian region there is no great difference. (I used 2009 data since I had that to hand)The June average maximum was about 7.3 GW, minimum 5.0 GW and December values 6.4 GW and 4.5 GW respectively. However, the yearly peak was 10.4 GW, at noon on January 29. On January 29, the temperature reached 44.3 degrees and demand remained above 10 GW from 10am until after 4:30pm. On January 29, horizontal insolation was 30.6 MJ/m2.
Supply infrastructure must be sized to satisfy this peak air conditioning load, which might only occur for ten hours in a year. This infrastructure would certainly benefit from having some local solar generation.
In a full renewables grid where solar provided (say) 50% of total energy; this energy would be acquired over a roughly 12 hour period and the peak output would need to be somewhat more than average demand. For Victoria this might be (say) 6 GW. If half of this generation was co-located with demand (i.e. in Melbourne) then it would be generating at least 3 GW on any hot summer day where the air conditioning load was high. This would mean that residual demand would be less than 7 GW and transmission infrastructure could be down-sized somewhat for this lower peak and would better match normal demand levels.
The other half of the solar generation could be located in a more favourable climate. Coober Pedy not only has almost twice the winter insolation that Melbourne receives, it is almost an hour “behind” Sydney and so its solar peak is time shifted by an hour to better match demand profile in our population centres.

[proteos]

Jun 12, 2012 6:54:00 GMT 9.5 @sod said:

ps: nuclear generation might have been higher over all day on the 15th/26th of May, but solar was producing at the time when electricity is most needed and most expensive, while a significant portion of the nuclear power was cheap night time tariffs. 8and demand at night will even fall more, when peak prices come down...)

I doubt that nuclear have been displaced by solar. In the merit order, nuclear comes before lignite, if I remember correctly. So lignite should have gone first. I'd rather bet on maintenance and planned closures.

I'd like to know the real importance of night-time tariffs in Germany, notably compared with France, where they are common because of electrical heating. We should not forget the big industrial users, with facilities running 24/7. Here lowering the demand would mean these industries have moved elsewhere, perhaps to places with little regard for climate change issues.

A consequence of what you say about solar PV in Germany is that subsidies should stop now: solar PV is doing what we can expect from it in Germany with today's technology status.

[sod]

i agree, Solar will not directly replace nuclear. (you are right about merit order effect, it is also a myth that germany was importing french or chech nuclear power after the switch of, as merit order makes this very unlikely)

www.oeko.de/oekodoc/1130/2011-015-de.pdf
(page 24 shows no increase in French nuclear power, and the power would have been sold anyway..)

i couldn t locate good data on night-time tariffs, but this paper mentions 2 million electric heating systems (households i guess), with 20TWh and about 13% of household electricity use.

several big companies have been advertising "heating electricity" at cheap prices and now are increasing prices on customers stuck with the heating systems..

Merit order and nighttime tariffs (and spot market..) means, that solar is driving down peak prices. this is good news for industry, as they don t really pay for the FiT and so have double profit from the changes.

there should be a change to solar FiT. But the drastic change brought up got stopped by even conservative (CDU) state leaders. and abrupt end of FiT causes bubble implosions.
what is needed is a reliable mechanism reducing the tariffs.

ps: germany produced 10% solar in May. not bad for a "myth", isn t it?

thinkprogress.org/climate/2012/06/12/497984/solar-provides-10-of-germanys-electricity-in-may/

[Graham Palmer]

Jun 12, 2012 20:14:13 GMT 9.5 @david jones said:

A major part of supply infrastructure cost is associated with peak demand. By peak I mean the absolute maximum demand which might occur rather than typical daily maximums.

Excellent point

Supply infrastructure must be sized to satisfy this peak air conditioning load, which might only occur for ten hours in a year. This infrastructure would certainly benefit from having some local solar generation.

This is covered in my two links above: the critical peak demand periods often occur early evening when there is no solar output, hence there is no benefit in having local solar generation. In Melbourne, there are also occasions when afternoon cloud cover and increased humidity occurs while air-conditioning demand stays high.

In a full renewables grid where solar provided (say) 50% of total energy; this energy would be acquired over a roughly 12 hour period and the peak output would need to be somewhat more than average demand.

I gather you are assuming the use of battery storage. The average home uses about 20 kWh a day and to store, say 1.5 days of storage would require a battery like this (assume max depth-of-discharge of 50%,48V, 1600 Ah, $19,920, 10 year life) :

www.allnaturalenergy.com.au/index.php?page=shop.product_details&flypage=flypage.tpl&product_id=39&category_id=12&option=com_virtuemart&Itemid=2

The other half of the solar generation could be located in a more favourable climate. Coober Pedy not only has almost twice the winter insolation that Melbourne receives, it is almost an hour “behind” Sydney and so its solar peak is time shifted by an hour to better match demand profile in our population centres.

Theoretical musings such as this are interesting, but in practice Vaclav Smil provides a good overview of some of the challenges (see page 6).

www.vaclavsmil.com/wp-content/uploads/docs/smil-article-2011-AMSCI.11.pdf

[anonposter]

Jun 13, 2012 6:58:56 GMT 9.5 @sod said:

Merit order and nighttime tariffs (and spot market..) means, that solar is driving down peak prices. this is good news for industry, as they don t really pay for the FiT and so have double profit from the changes.

No, it's residential users who pay for the FiT (i.e. the employees of those factories) and especially the poorer residential users who don't own their own home.

Of course those industrial users will pay later on if the reduction in peak prices means there isn't enough generating capacity because no one could make a case to invest in ensuring it would exist.

Jun 13, 2012 6:58:56 GMT 9.5 @sod said:

and abrupt end of FiT causes bubble implosions.

That's fine by me, at least compared to the alternatives (and more bubble implosions could have the useful effect of discouraging investment in useless renewable energy).

Jun 13, 2012 6:58:56 GMT 9.5 @sod said:

what is needed is a reliable mechanism reducing the tariffs.

If you don't reduce them fast enough then you end up wasting a lot of money (and hurting the poor more) while if you do reduce them fast enough you get a bubble implosion.

[david jones]

Graham Palmer, unfortunately the two reports you cite are not terribly relevant. The first is rather old now and the peak loads continue to become more extreme. The second report is considering exported power from PV which is very different in scale from actual displaced generation. Even so, this report acknowledges the reduced network costs and reduced losses which the exported power is responsible for but it cannot readily value it so it effectively ignores it. The beneficial impact for the full displaced load is much much greater.
To put this in perspective; assume you are running your major city at a yearly peak load of 10 GW with 10% losses in transmission and distribution. Now substitute just 5% of that load with locally generated power (solar or otherwise) – you now need to generate about 750 MW less at the power stations because the load reduction of 500 MW is accompanied by a significant reduction in transmission and distribution losses for the remaining 9.5 GW. At less extreme demand levels the benefit is very much reduced.
I disagree with your comment about peaks occurring in early evening when there is no solar input. In Melbourne, solar continues to generate useful output until 6pm (7pm daylight saving time) in summer, which is well after demand peak. I have attached (I hope) a plot of notional impact of a 3 GW solar generation on the Jan 29 data I cited previously. The residual peak demand would be in the late afternoon (as you suggest) but is about 1.5 GW lower than for the no solar case.
Oh – I certainly do not advocate batteries for grid storage.
[/img]
I give up on the image attaching!

[sod]

Anon, i was replying to proteos, who was worried about industry being driven away by FiT. In reality, under the current German FiT scheme the industry does profit twice: they don t pay the FiT scheme and they pay lower peak prices.

i am also surprised that you are worried about a lack of base load plants. With nuclear being so cheap, it should be profitable even without peak prices?!?

you are also wrong about the bubble. what happens is this: you announce the end of the FiT period and people build MORE before that date. In the end you pay the same amount of FiT money, but you cause a lot of chaos. and some companies will end up with stockpiles of modules that didn t make it onto roofs in time, forcing them into bankruptcy.

----------------------------

on the topic of peak solar output and peak demand:

In Germany (and other temperate regions) the fit is close to perfect.

www.transparency.eex.com/de/
(also look at the graphs provided by proteos, which shows it even better with import/export added)

in warmer countries, demand peak is later. But this will shift, with air conditioning and a few hours of storage will also help...

[Graham Palmer]

Jun 13, 2012 23:18:51 GMT 9.5 @david jones said:

I disagree with your comment about peaks occurring in early evening when there is no solar input. In Melbourne, solar continues to generate useful output until 6pm (7pm daylight saving time) in summer, which is well after demand peak.

David, I agree that there will be days when the solar does in fact contribute as you suggest. Here are two graphs from the highest demand day in 2010 which occurred on 11 Jan , temp was 43.6. A correctly installed north facing solar panel will drop off rapidly from 6.30pm, but demand continued to be strong until past 9pm. Note that the sun sets in the south-west in summer, so north facing panels lose output sooner than one might intuitively expect.

The first has is demand in MW with the maximum of 9843 MW at 2pm, but still 8948 MW at 8pm, and a scaled PV output (I don't have the data for that day, but this uses generic Melbourne data for the same day of year, clear skies).

The second graph has the difference between the demand and scaled PV. As can be seen, your comments about the timing of the peak and solar is valid (ie : there is plenty of solar at the time of the actual peak in demand), but note that the demand remains high while PV drops rapidly, leaving a 9000 MW shortfall - in other words, on the hottest day, when solar is also at its best, a theoretical 10,000 MW of solar provides effectively 1,000 MW of avoided demand on this day. So yes, solar has provided some benefit - is 10% of peak output a cost-effective peak demand strategy just past the summer solstice?

The situation is worse than this because the rapid decline in solar output between 6.30 and 7.00pm of 4,000MW in our theoretical scenario actually accentuates the peak demand problem. Despite the increasing peakiness due to air-conditioning, at least this is predictable and relatively gradual during the day, but the loss of supply due to solar is an order of magnitude greater in ramp rate.

(hopefully the graphs work)

[anonposter]

Jun 14, 2012 1:20:54 GMT 9.5 @sod said:

Anon, i was replying to proteos, who was worried about industry being driven away by FiT. In reality, under the current German FiT scheme the industry does profit twice: they don t pay the FiT scheme and they pay lower peak prices.

But their employees pay for the FiT (and if their cost of living increases they may start asking for a raise to help pay for the increased electricity bills).

Jun 14, 2012 1:20:54 GMT 9.5 @sod said:

i am also surprised that you are worried about a lack of base load plants. With nuclear being so cheap, it should be profitable even without peak prices?!?

Peak prices help though (and whilst paid off nuclear plants are cheap to operate, building new ones is what needs to happen, not merely keep the old ones running (which would happen on its own without government intervention to shut them down pretty much regardless of any FiT)).

Jun 14, 2012 1:20:54 GMT 9.5 @sod said:

you are also wrong about the bubble. what happens is this: you announce the end of the FiT period and people build MORE before that date.

So what is the way to end it? I mean people are going to do that to get in before the deadline as soon as it's announced no matter how you announce it so the best approach to avoid spending so much money is to announce it as quickly as possible so that the installing frenzy causes as little trouble as possible.

Jun 14, 2012 1:20:54 GMT 9.5 @sod said:

In the end you pay the same amount of FiT money,

No, you pay a lot less (even though there is an installing surge towards the end).

Jun 14, 2012 1:20:54 GMT 9.5 @sod said:

but you cause a lot of chaos. and some companies will end up with stockpiles of modules that didn t make it onto roofs in time, forcing them into bankruptcy.

Which is exactly what should be happening (and it's actually good for those who have a legitimate need for solar panels as they can then pick them up nice and cheaply).

If there are too many companies installing solar panels then some of those companies are going to have to go under, if you remove the FiT then no matter what happens there will be too many companies installing roof decorations.

[cyrilr]

Jun 12, 2012 6:54:00 GMT 9.5 @sod said:

thanks for those graphs, cool work!

i think it is obvious that solar will have little effect on winter peaks and basically none on night time. but the topic of this post is: "Solar power correlating with peak demand - a myth?"

and this can be definitely answered: in Germany, the correlation is close to perfect in summer.

Such an answer would be disingenious at best, as Germany's demand peaks in winter, not in summer. As such it is very much ANTI-correlated with demand. In the future with heat pumps providing winter heating (to eliminate natural gas and oil heating systems) this anti-correlation will be further enforced.

[Graham Palmer]

CSIRO just released its report on solar intermittency:
www.csiro.au/science/Solar-Intermittency-Report

The report provides an excellent overview of the literature, focuses on distributed PV rather than CSP, and the input from network operators lifts the value of the report. However, the report is also a strong sales pitch for the CSIRO - the report sets out a number of problems which the CSIRO is perfectly placed to research!

I'm not sure whether others reading the report will see the report in the same terms as I do, but the single-minded pursuit of solar as a desirable outcome, regardless of costs or challenges is striking. It almost appears as though there is a view that increasing the penetration of solar is ipso facto desirable and that supporting solar should be a primary objective of grid operators. For example:

The literature indicates that cheaper less flexible plants will need to be replaced with more flexible expensive plants to accommodate high penetrations of solar generation. Otherwise, a significantly larger amount of ancillary services or additional generation would be required to manage PV power output fluctuations.(pg 172)

Strengthening the electricity network so that intermittency affects are not as localised (pg 11)

Controlling loads in response to network requirements (pg 11)

The report also lays out other solutions to intermittency such as:

Using short-term energy storage systems (pg 11)

Another method is to control the solar generation output under intermittent cloud coverage during periods of peak
system demand when the network has fewer generating units on standby and less on-line regulating capacity. These
corrective measures, however, may cause the system to deviate from its optimal operating conditions, thus adversely
affecting the economics of solar generation. (pg 173)

The question the rest of us need to answer is whether we are prepared to pay to upgrade the grid to ensure that solar PV is more viable, or take a more holistic approach and ask what functions rooftop solar is best suited to performing. In my mind, this is summer peak demand management, but to perform this role will require distributed battery storage. If we are not prepared to pay for battery storage, why would we want to pay for grid upgrades which will be largely ineffective without storage?

The report focuses on PV since CSP is already available with storage and there is nothing to prevent developers from connecting a CSP generator to the grid and collecting the RECS, but evidently CSP requires a much larger subsidy than wind.

Like most of these reports, it makes its share of motherhood statements such as:

controlling loads in response to network requirements (pg 11)

What the report doesn't state is that this means persuading householders to turn their air-conditioners down on the hottest days. My experience is that many people, particularly people on modest incomes or the elderly, only purchase an air-conditioner so that they have relief on the hottest days, and particularly during the evening. It is obviously possible to use differential pricing to force these people to turn their air-conditioner down or off while the rest of us accept the higher cost, but is this politically, socially, or morally feasible?