|
Post by Barry Brook on Apr 1, 2013 21:35:33 GMT 9.5
A new guest post by Graham Palmer has been published on BraveNewClimate. Link here: bravenewclimate.com/household-pv-primary-le-powerWith declining system costs and assuming a short energy payback period, photovoltaics (PV) should, at face value, be able to make a meaningful contribution to reducing the emission intensity of Australia’s electricity system. But will it? Graham Palmer takes a critical look at this key question. This BNC Discussion Forum thread is for the comments related to this BNC post.
|
|
|
Post by Martin Burkle on Apr 2, 2013 0:10:36 GMT 9.5
"Peak usage happens after work and basically due to air conditioning" this statement brings two thoughts to mind:
1. There is a kind of air conditioning that freezes water at night and cools by thawing the water in the day time. I wonder what the EROI is on such a system?
2. Where I live we change time twice a year because everyone else changes time twice a year. It's called "day light savings time" but I don't know what purpose it serves these days. Could we have an "Electric peak" kind of energy savings time that could shift peak usage closer to solar output (everybody goes to work earlier)?
|
|
|
Post by edireland on Apr 2, 2013 0:11:53 GMT 9.5
Daylight Saving Time in the UK is primarily because otherwise we could end up with children in the north of the country going to school in the dark in Winter otherwise.
|
|
|
Post by anonposter on Apr 2, 2013 1:35:16 GMT 9.5
Daylight Savings Time would reduce lighting energy consumption but increase air conditioning energy consumption, or at least that's the explanation for recent adoptions of it not saving energy when past switches to it did.
On the whole though I think we'd be better off if we realised that Daylight Savings Time was a joke, not something that should actually be implemented and if you're worried about children going to school in the dark just start school later rather than screw around with the clocks.
|
|
|
Post by engineerpoet on Apr 2, 2013 4:09:02 GMT 9.5
1. There is a kind of air conditioning that freezes water at night and cools by thawing the water in the day time. I wonder what the EROI is on such a system? You mean, like Ice Energy's Ice Bear? They may increase efficiency slightly by operating the A/C at night when the air temperature is cooler and system COP will be a bit higher, but mostly they avoid the need for peaking plants. The peakers are usually built for low capital cost and efficiency is sacrificed. How much... hard to say. On the other hand, the arbitrage between on-peak and off-peak rates appears to pay for the storage system without other subsidies.
|
|
|
Post by edireland on Apr 2, 2013 4:23:58 GMT 9.5
On the whole though I think we'd be better off if we realised that Daylight Savings Time was a joke, not something that should actually be implemented and if you're worried about children going to school in the dark just start school later rather than screw around with the clocks. But then what do you do with the children before they go to school, after the parents have gone to work? They still have to transit somewhere in the dark because if you wait till 10am or whatever, the parents will already be in work. And you still run into trouble during the depths of winter because if you are not careful because they will be coming back from school in the dark instead. Even in Edinburgh (so south of Scotland)... the day in the depths of winter is only ~7hrs long.
|
|
|
Post by sod on Apr 2, 2013 5:30:08 GMT 9.5
|
|
|
Post by sod on Apr 2, 2013 5:36:47 GMT 9.5
This german site shows expected and real power production, including solar. www.transparency.eex.com/de/You can see how solar power reduces the peak and replaces conventional power.
|
|
|
Post by edireland on Apr 2, 2013 7:03:07 GMT 9.5
The fact they still need a subsidy of 30p/kWh or more in Britain is rather telling.
|
|
|
Post by Graham Palmer on Apr 2, 2013 10:04:03 GMT 9.5
The real data from SA is showing a much more optimistic situation than the simulated data used in the article: Sod, the AEMO news article from Giles’ article states that falling industrial and mining demand, along with the uptake of renewables, has reduced energy demand - it doesn’t quantify the reduction in annual peak demand due to solar nor indicate that it has been significant. Giles also refers to Sandiford’s article, and I quote from Sandiford “With solar PV biting into the daytime demand but barely shaving peak demand, the unit cost of distribution will inevitably rise.” The point of Sandiford’s article (which I cite in the paper) is that PV is reducing the spot price, which will have long term consequences, which I discuss in the paper. This german site shows expected and real power production, including solar. Germany, like the UK and other countries have their annual peak during winter, hence PV provides no reduction in annual peak demand. Sandiford, Watt, Myers, Denholm, IPART, the Productivity Commission etc all come to the same conclusions: PV’s diurnal profile works in generally well with daytime demand, but PV does not meaningfully displace annual peak demand. This doesn’t mean that there isn’t a single instance in which PV will provide some support, but that in most large grids, particularly those with industrial demand, PV cannot meaningfully provide the guaranteed supply when it is demanded. I discuss the role of storage and actually highlight network support as PV’s key strength in the abstract, conclusions and several pages, but it requires a small amount of storage to capture this benefit. I think you need to read the whole paper to get a sense of my main arguments.
|
|
|
Post by Martin Burkle on Apr 2, 2013 14:05:32 GMT 9.5
Peak usage and the peaking power are just a little different.
Peak usage is the maximum power demanded by the system normally about 6PM on a very hot day. The generation system (all the generators combined) must supply 100% of the maximum power needed.
Peaking power generally is used and paid for at the highest rate during a time span (like 2 to 7). Solar PV is really good peaking power between 2 and 4 but starts to fall off as the sun gets lower in the western sky.
Two seemingly contradictory statements can both be true. Solar really does supply a lot of the peeking power needed by the system. Solar really does not affect the peak usage very much at all.
Without storage a grid designer can not count on PV solar to reduce the amount of non-solar generation needed to power a system.
If you need more detail about the above statements, please reread the article we are talking about.
|
|
|
Post by sod on Apr 2, 2013 16:54:04 GMT 9.5
Sod, the AEMO news article from Giles’ article states that falling industrial and mining demand, along with the uptake of renewables, has reduced energy demand - it doesn’t quantify the reduction in annual peak demand due to solar nor indicate that it has been significant. sorry, but i have to disagree with this explanation. In contrast, the article i quotes says: and: the article gives very strong indications for solar PV being the reason for that serious drop in peak demand! reneweconomy.com.au/2013/rooftop-solar-reshapes-energy-market-in-south-australia-18272Giles also refers to Sandiford’s article, and I quote from Sandiford “With solar PV biting into the daytime demand but barely shaving peak demand, the unit cost of distribution will inevitably rise.” The point of Sandiford’s article (which I cite in the paper) is that PV is reducing the spot price, which will have long term consequences, which I discuss in the paper. sorry, but the graph shows a different picture. the peak has been moved from afternoon 16 to evening 19 hours. reneweconomy.com.au/2013/how-rooftop-solar-is-reshaping-energy-market-in-s-australia-18272/sa_2008_td_time_of_day_averageI discuss the role of storage and actually highlight network support as PV’s key strength in the abstract, conclusions and several pages, but it requires a small amount of storage to capture this benefit. I think you need to read the whole paper to get a sense of my main arguments. I will have a more detailed look, as soon as i got time. The article is interesting!
|
|
|
Post by Graham Palmer on Apr 2, 2013 17:55:45 GMT 9.5
sorry, but i have to disagree with this explanation. Sod, I'd recommend you read the paper in full to get a fuller understanding of the subject; this is a topic that a lot of researchers have devoted a lot of time to and I'd recommend you refer to some of the authoritative sources I've cited, eg. Myers et al, Denholm and Margolis, Nelson et al, the IPART review on FIT's received a large number of submissions, Denholm and Hand, Watt et al, etc.
|
|
|
Post by edireland on Apr 2, 2013 22:01:47 GMT 9.5
In britain, during winter, the demand spike starts before sunrise, let alone the sun clearing trees/buildings.
So the demand would spike and then drop again leaving us with an OCGT suited peak in the morning and again in the evening.
|
|
|
Post by sod on Apr 3, 2013 4:01:02 GMT 9.5
|
|
|
Post by sod on Apr 3, 2013 4:05:20 GMT 9.5
In britain, during winter, the demand spike starts before sunrise, let alone the sun clearing trees/buildings. So the demand would spike and then drop again leaving us with an OCGT suited peak in the morning and again in the evening. i notice that quite often when discussing solar power, i find myself debating Britain morning power demand peak or night time power. Why not focus on the part, on which PV solar obviously is a working solution?m like it is in south australia?!?
|
|
|
Post by edireland on Apr 3, 2013 4:08:28 GMT 9.5
You may not have noticed, but most of the areas where Solar PV could Potentially be considered a working solution..... has very few people in it.
Nevada or most of Australia is a desert for a reason. The fraction of the world's population living in appropriate areas is tiny.
|
|
|
Post by sod on Apr 3, 2013 4:46:59 GMT 9.5
You may not have noticed, but most of the areas where Solar PV could Potentially be considered a working solution..... has very few people in it. Nevada or most of Australia is a desert for a reason. The fraction of the world's population living in appropriate areas is tiny. Sorry, but this is plain out nonsense! India jumps to mind: en.wikipedia.org/wiki/Solar_power_in_India
|
|
|
Post by edireland on Apr 3, 2013 6:31:11 GMT 9.5
" Government support and ample solar resources have also helped to increase solar adoption, but perhaps the biggest factor has been need. India, "as a growing economy with a surging middle class, is now facing a severe electricity deficit that often runs between 10 and 13 percent of daily need"" That is the telling statement. The only reason solar is making any progress is the fact that the Indian economy is still in the geometric energy use increase phase of development. The amount of solar added is simply because nothing else is available and there is a glut of solar panels caused by the collapse in demand for them everywhere else. In the long term it is probably no more viable than it is in Southern Europe. And what happens to the power demand during the Monsoon?
|
|
|
Post by cyrilr on Apr 3, 2013 17:32:24 GMT 9.5
In figure 4, the solar generation is bloated by a factor of 8 over the substation load.
Why? I would expect this from pro-solar advocacy groups, not from a scientific paper.
If 50% of the housholds have that 1.5 kWp system, and the PV generation is displayed fairly (ie on the same scale as substation load) then the solar generation appears as a tiny speck not a giant mountain. I understand that the graph must show the shape of PV versus the load curve, but that doesn't justify chartjunk.
To get that mountain peak, you'd need 6 kWp systems on 100% of the homes. Most homes actually couldn't fit that on their roofs, and it would still only generate a fraction of total load and 1/3 has to be dumped since it exceeds load on the full light of day.
|
|
|
Post by sod on Apr 3, 2013 18:27:49 GMT 9.5
That is the telling statement. The only reason solar is making any progress is the fact that the Indian economy is still in the geometric energy use increase phase of development. The amount of solar added is simply because nothing else is available and there is a glut of solar panels caused by the collapse in demand for them everywhere else. In the long term it is probably no more viable than it is in Southern Europe. And what happens to the power demand during the Monsoon? you missed the point. India has perfect conditions for roof top solar PV. en.wikipedia.org/wiki/File:Solar_Resource_Map_of_India.pngAnd India will move into massive deployment of air conditioning systems very soon. PS: With California and India now on your list of good places for PV solar, you should think about changing your "99.999% not suited for PV" number...
|
|
|
Post by sod on Apr 3, 2013 18:36:37 GMT 9.5
In figure 4, the solar generation is bloated by a factor of 8 over the substation load. Why? I would expect this from pro-solar advocacy groups, not from a scientific paper. The graph comes from another paper and I have seen it multiple times. We have even already discussed that graph here. The real problem with the graph is a completely different distortion. (people can just look at the right hand axis and understand that PV is on a different scale) The graph is using households as (nearly) the sole demand. This is moving peak demand later into the day (people get home and leave their factories and workplaces, which now use LESS power). It is a cheap trick to make PV solar look more off-peak than it really is. There is a similar problem with table 2 on the next page (pdf p. 8). The table is used to give the impression that on some days PV does little to reduce peak demand. But the values show, that top day peak demand can drop from 9858 to 9051 (nearly 10%) and that is with a very crude and unoptimised PV solar allocation!
|
|
|
Post by cyrilr on Apr 3, 2013 21:03:24 GMT 9.5
Actually if you include industrial and commercial demand, it gets a lot worse on the whole: this is because those demands are much more baseload than homes demand. So PV's mismatch with demand would further deteriorate, even though the peaking portion may improve slightly, the situation on the whole looks far worse. For example, most industries I work with have 6000 hour/year workload. They work 3 shifts, and each shift produces about the same amount. In fact in some cases the night shift has more production because of excess grid capacity available. Obviously the amount of solar available at night is zero, not useful for industry.
This isn't really about peaking demands, this is about how will we power a modern economy with lots of industry without fossil fuels. As the paper from Palmer suggests, PV is not very useful and in fact may risk fossil lock in due to it's fickle nature requiring loads of fossil backup. Actually providing 10% of your power with PV and most of the remaining 90% with gas and coal is not fossil backup, it's downright greenwashing.
I don't feel that the paper has made this explicit enough.
|
|
|
Post by edireland on Apr 3, 2013 22:05:13 GMT 9.5
So according some days in Melbourne, the grid will save 1200MW of conventional demand thanks to 2000MW of solar.....
But on other days there is no demand benefit at all, so assuming the Table 2 data is accurate.... you only save 807MW of plant (reduction in absolute peak demand from ~9858MW to ~9051MW).
So, lets just do some back-of-the-envelope calculations:
Which is cheaper to build out and operate? 2000MW of solar, or 807MW of gas or nuclear?
Well even at the low low prices being reported for large scale installations (so hardly the "democratisation of power" hailed by solar advocates) of $2200/kW, this would give an installation price of $4.4bn US dollars.
Which means the price of the nuclear build, assuming a 900MW unit, would have to be below $4880/kW in order to be even on capital cost grounds. That would be challenging in the first world if you are only going to build a single unit or a small fleet of them (you should easily be able to beat that in the west if you go all out for nuclear rollout).
But there is something that is not mentioned in these comparisons, while solar would have no fuel costs and nuclear does have manning costs of about 1cent/kWh (US), the plant will be able to turn out power at 1-2cents/kWh all night and for significant portions of daylight hours.
This power could be used to run a large desalination plant during off peak hours, water being something you will probably need relatively soon in Australia and large parts of India. As the cost of the nuclear plant has been defrayed the electricity is so cheap that the relatively low capacity factor (~70%) for the desalination equipment becomes effectively meaningless.
So no, nuclear has solar beaten in this application as well.
|
|
|
Post by sod on Apr 3, 2013 22:07:44 GMT 9.5
Actually if you include industrial and commercial demand, it gets a lot worse on the whole: this is because those demands are much more baseload than homes demand. So PV's mismatch with demand would further deteriorate, even though the peaking portion may improve slightly, the situation on the whole looks far worse. For example, most industries I work with have 6000 hour/year workload. They work 3 shifts, and each shift produces about the same amount. In fact in some cases the night shift has more production because of excess grid capacity available. Obviously the amount of solar available at night is zero, not useful for industry. This is false. Household power demand peaks in the early evening because people cook food at that time and many people only arrive home. (they use absolutely no power at home before) the graph is used above shows what is true for the majority of demand curves in hot countries: reneweconomy.com.au/2013/how-rooftop-solar-is-reshaping-energy-market-in-s-australia-18272/sa_2008_td_time_of_day_averagePeak power demand (without vsolar PV drwaing it down) happens in the late afternoon, not early evening as pure household data suggests. the graph in the Palmer paper is giving a false impression. (solar being completely out of sync with peak power demand) This isn't really about peaking demands, this is about how will we power a modern economy with lots of industry without fossil fuels. As the paper from Palmer suggests, PV is not very useful and in fact may risk fossil lock in due to it's fickle nature requiring loads of fossil backup. Actually providing 10% of your power with PV and most of the remaining 90% with gas and coal is not fossil backup, it's downright greenwashing. I don't feel that the paper has made this explicit enough. wow, 10% of PV is a pretty strong number! the reality today is, that multiple countries are moving into 20% to 30% alternative (mostly wind and PV) regions. So your 90% number is already wrong today.
|
|
|
Post by cyrilr on Apr 3, 2013 22:31:10 GMT 9.5
Sod's rebuttal is that 70-80% fossil is acceptable. Whereas the climate scientists make it clear than 80% reduction is needed. Considering economic and population growth, we're talking about 90% reduction. Reality is that Germany is the world's biggest user of brown coal, the dirtiest type of coal. Oops! Never mind, be happy, look at the pretty solar panels. Denmark, the biggest champion of wind in Europe, gets a bigger percentage of its power from coal than almost any other European country. Oops!
Sod claims that my assertions are false, but provides no evidence or even claim that shows this is the case (or even that Sod has read my comment at all).
Sod further asserts that I am wrong about industrial demand; apparently claiming that industries don't use power at night. Sod completely misses the point, as usual. So I will repeat: the issue is not peaking, it is total energy supply. PV and wind combined cannot supply the needed 90% reduction, therefore they are dangerous distractions. People like Sod use wind and solar as an excuse to not build nuclear plants, which is a recipe for failing on all climate targets.
The best synergy so far that I can imagine is for solar to provide airconditioning in hot sunny regions, reducing peak demand, possibly with the help of some ice making air conditioning as buffer, and then the remaining demand is much more baseload so can be serviced with nuclear. Airconditioning is typically less than 5% of total grid demand, so a 5% solar, 90% nuclear, 5% fossil (preferably natgas) is probably a realistic option.
|
|
|
Post by anonposter on Apr 4, 2013 1:46:20 GMT 9.5
wow, 10% of PV is a pretty strong number! I suspect it to be beyond what is sensible. the reality today is, that multiple countries are moving into 20% to 30% alternative (mostly wind and PV) regions. So your 90% number is already wrong today. But when we look at those countries we see: - High electricity prices
- High specific CO2 emissions from their electricity sector
Not exactly a winning combination. The best synergy so far that I can imagine is for solar to provide airconditioning in hot sunny regions, reducing peak demand, possibly with the help of some ice making air conditioning as buffer, and then the remaining demand is much more baseload so can be serviced with nuclear. An even better synergy would be hot water, in fact that seems to be about all solar is really good for. Hydro where it can be built is great at peaking and LFTRs also load follow well, in all likelihood we're going to have to load follow nuclear power plants no matter what as there just won't be a carbon neutral alternative in many places.
|
|
|
Post by geoffrussell on Apr 4, 2013 6:53:21 GMT 9.5
Congratulations on a great paper Graham. I'm still digesting!
|
|
|
Post by cyrilr on Apr 4, 2013 17:50:52 GMT 9.5
Solar hot water, I have my doubts about this. Usually hot water demand is fairly stable seasonally. Not a good match for solar. In fact, I need a lot more hot water in winter due primarily to lower feedwater temperature from the drinking water supply system (ie requires more heating to get to the required 35-55 degree Celsius that I need). This puts solar hot water at a disadvantage in my mind.
At least with airconditioning in hot arid areas like much of Australia, the output dovetails the demand closely enough. I'd expect ~5% PV would be sensible for this, but it would certainly require chilled water or ice storage to deal with the afternoon/evening aircon peak. Not a big deal I would say, maybe 3 hours of low temp thermal store. Ice is more compact, but chilled water would be more efficient and practical and seems... sensible... pardon the pun.
|
|
|
Post by anonposter on Apr 4, 2013 17:56:40 GMT 9.5
Solar hot water, I have my doubts about this. Usually hot water demand is fairly stable seasonally. Not a good match for solar. In fact, I need a lot more hot water in winter due primarily to lower feedwater temperature from the drinking water supply system (ie requires more heating to get to the required 35-55 degree Celsius that I need). This puts solar hot water at a disadvantage in my mind. The main advantage is how simple it is, in fact I understand that the power saved by not needing to heat as much water already makes it worthwhile in reasonably sunny places.
|
|