Less than the sum of its parts: Rethinking “all of the above” clean energy

Guest Post by John Morgan. John is Chief Scientist at a Sydney startup developing smart grid and grid scale energy storage technologies.  You can follow John on twitter at @JohnDPMorgan.


The fastest path to decarbonization would seem to be combining every kind of low carbon energy available – the so-called “all of the above” camp of clean energy advocacy.  The argument runs that different kinds of clean energy are complementary and we should build as much of each as we can manage.  This is not in fact the case, and I’ll show that a mix of wind and solar significantly decreases the total share of energy that all renewables can capture.  The “all of the above” approach to emissions reduction needs to be reconsidered.

In a recent essay Breakthrough Institute writers Jesse Jenkins and Alex Trembath have described a simple limit on the maximum contribution of wind and solar energy: it is increasingly difficult for the market share of variable renewable energy [VRE] sources to exceed their capacity factor.  For instance, if wind has a capacity factor of 35%, this says it is very difficult to increase wind to more than 35% of electrical energy.  Lets look at why this is so, and extend the principle to a mix of renewables.

The capacity factor (CF) is the fraction of ‘nameplate capacity’ (maximum output) a wind turbine or solar generator produces over time, due to variation in wind, or sunlight.  Wind might typically have a CF of 35%, solar a CF of 15% (and I’ll use these nominal values throughout).

Jesse and Alex’s “CF% = market share” rule arises because it marks the point in the build out of variable renewables at which the occasional full output of wind and solar generators exceeds the total demand on the grid.

At this point it gets very hard to add additional wind or solar.  If output exceeds demand, production must be curtailed, energy stored, or consumers incentivized to use the excess energy.  Curtailment is a direct economic loss to the generators. So is raising demand by lowering prices.  Energy storage is very expensive and for practical purposes technically unachievable at the scale required.  It also degrades the EROEI of these generators to unworkable levels.

Jesse and Alex make this argument in detail, backed up with real world data for fully connected grids (i.e. not limited by State boundaries), with necessary qualifications, and I urge you to read their essay.

The “CF% = market share” boundary is a real limit on growth of wind and solar.  Its not impossible to exceed it, just very difficult and expensive. Its an inflexion point; bit like peak oil, its where the easy growth ends.  And the difficulties are felt well before the threshold is crossed.  I’ve referred to this limit elsewhere as the “event horizon” of renewable energy.

So if wind is limited to say 35% of energy, and solar to 15%, can we add them together and achieve 50% share?  The Breakthrough authors seem to think so, writing that “this threshold indicates that wind and solar may be able to supply anywhere from a third to a half of all electricity needs”.  That would be a very considerable addition of low carbon energy.  But unfortunately this is not the case.

Here’s the problem with adding solar: it produces about half as much energy as wind for the same capacity.  And the capacity factor rule sets a limit on total variable renewable capacity.  So at the limit solar capacity is not additive to wind, it displaces wind, while producing less energy.  Any amount of solar lowers the share of energy that wind and solar together can acquire, and the optimal mix for decarbonization is all wind and no solar.

This is a general corollary to the capacity factor rule – adding lower capacity factor generation to the mix reduces the potential share of variable renewable energy.  It is the energy equivalent of Gresham’s Law – “Bad energy drives out good”.  Far from targeting a “mix of renewables”, we are better off targeting just the one with the highest capacity factor.  We should build wind and not solar.

You can see this dynamic in the following figure, which plots the limiting share of wind and solar energy (VRE) in the grid as a function of solar’s share of wind and solar capacity.  Adding solar capacity cannibalizes wind capacity, and reduces the total amount of low carbon energy that these sources can ultimately provide.  Solar is not additive to wind; its subtractive.

The situation becomes even clearer if we shift focus from installed capacity to energy delivered.  In the plot below, the x-axis now shows the fraction of wind and solar energy that is produced by solar.

Introducing solar energy into the mix causes a rapid drop in the maximum grid penetration of all variable renewable energy.  Wind alone could potentially achieve 35% of grid energy share.  But with 50% solar, the maximum share that wind and solar together can achieve is just 21%.

In other words, building out solar effectively robs us of a whole climate stabilization “wedge”.

It should be remarked that this capacity factor rule sets too optimistic a limit.  The Breakthrough writers cite estimates that only 55%-60% of grid energy could be replaced by variable sources, due to stability requirements.  This means VRE share will struggle to exceed 60% of capacity factor, and the limits described above will be reduced by that factor.  So while wind alone could achieve up to about 21% of all electricity, a 50-50 mix of solar and wind is practically limited to only 12%.

This is a lot to give away.

So long as we only have a small amount of solar and wind we can build as much of either as we like.  The limit only becomes apparent at higher penetration.  But this happens much more quickly if there’s a lot of solar in the mix.

There may be good reasons to build solar in the early stages of a clean energy expansion.  The rate of emissions reduction matters, and while supply chains are developing, building both solar and wind might help.  But if this trajectory is to continue we will need to shift resources to wind fairly early on, and allow solar capacity to decline.

This should prompt a rethink of the simplistic “all of the above” response to emissions reduction, and the popular notion that there should be a mix of renewables.  If it doesn’t even work for wind and solar, does it work anywhere at all?  Its time to pick some winners, and support for renewable energy at scale should increasingly favour wind over solar.

And we should also think about how to decarbonise the remaining eighty percent of the grid that variable renewables can’t touch.

91 Comments

  1. John, from reading this article I take it that this is more so a scenario for Utility scale and grid connected PV systems?

    I take it this wouldn’t be an issue for household or commercial building (i.e. warehouse) PV if there was no export into the grid? In effect it is lower daytime demand. In that instance I suppose the management of the grid due to the duck curve comes into play.

    Like

  2. @IR
    That’s exactly what I was thinking.

    A market structure where-in rooftop PV operators are rewarded in the long term for investing in storage (would need to be more versatile than Tesla’s) such that the load they place on the grid is smoothed out.

    @John
    I expect you’re aware that Carnegie reports an expected capacity factor of 40% for CETO? In principle, this would seem to mean we have to balance the benefits of wind build now with the overall maximised benefits of wave power build later plus the necessary development/investment.

    Like

  3. Irregular Commentator, the limit applies to grids, in particular to the full extent of all connections to the grid, that is, not limited by artificial boundaries like state or country borders. Jesse Jenkins & Alex Trembath describe this very well in their article.

    For a small scale system disconnected from a grid the homeowner/PV operator probably accepts in much more direct terms that they will be managing their demand very actively. That sort of commitment does not extend to the grid-connected population in aggregate.

    Like

  4. ActinideAge, yes, if there are higher capacity renewables available, then they should be favoured over wind, other things being equal. Any contribution from lower capacity factor energy reduces the total energy that can be delivered into the grid from the combination of variable sources.

    Like I said, “bad” (i.e. low capacity factor) energy drives out good.

    Like

  5. John Morgan,

    Excellent contribution, as always.

    However, I don’t agree with these statements:

    We should build wind and not solar.

    So long as we only have a small amount of solar and wind we can build as much of either as we like.

    These two statements may be technically correct, which I recognise your posts is about, but they are not correct when you consider the economics, and therefore, the viability of a policy given the real world constraints..

    The build out of wind and solar is delaying the development and build out of economically viable, low emissions intensity electricity system. Even before we get to penetration levels approaching the average capacity factor of the technology, the cost of wind and solar power – and particularly the CO2 abatement cost – are much higher than recognised (even by the economists who are doing the analyses that policies are based on). In Australia, the continuation of the RET to 2020 will delay the build of new baseload generation by decades (the 2014 Warburton Review of the RET explains this).

    Wind energy in the NEM in 2014 supplied 4.5% of electricity and avoided 3.5% of the emissions from the displaced generators. Therefore, the CO2 abatement effectiveness of wind was 78% (i.e. 3.5% / 4.5%) http://joewheatley.net/emissions-savings-from-wind-power-australia/.

    Wind is projected to supply about 15% of NEM’s electricity in 2020. The CO2 abatement effectiveness is likely to be about 60% at 15% penetration. At 60% CO2 abatement effectiveness, the CO2 abatement cost would be 67% higher than the already very high CO2 abatement cost of wind power as noted in the Warburton Review.

    The cost of abatement with wind power in Australia in 2020, under the RET, is likely to be:

    • 2 to 5 times the carbon price which was rejected by voters at the 2013 election

    • 4 to 8 times the Direct Action average price at the first auction

    • 6 to 12 times the current EU ETS price

    • 100 to 200 times the international carbon price futures to 2020

    Read more here: http://judithcurry.com/2015/04/27/wind-turbines-co2-savings-and-abatement-cost/

    I nearly choked on this bit:

    Its time to pick some winners, and support for renewable energy at scale should increasingly favour wind over solar.

    Picking winners is completely the wrong approach. Such policies do not last. They will not succeed. To succeed, policies must be economically rational.

    Mandating and subsidising wind and solar are economically irrational polices. These polices are delaying progress (by decades).

    Like

  6. Thanks John for the reply.

    Thinking this out further this would be a condition of sources dependent upon variable sources of fuel source such as weather patterns (wind and precipitation), seasons and sunshine. Whereas for others that have a fuel source that is tangible it is not the case. Reason being that a 50-65% cf for Gas isn’t a threshold, so the shorthand “cf% = market share” is really for fuel sources humans cannot control.

    I’m just trying to flesh out a shorthand rule that is easy for public consumption beyond BNC i.e. If fuel source is out of human control, then cf% = market limit?

    Like

  7. “If fuel source is out of human control, then cf% = market limit?”

    Yes, thats correct. Thats more or less the formulation Jesse and Alex used. Its well explained in their article, and its a very concise expression of the limit. What I’ve done here is look at how that limit works in the case of two or more sources with different capacity factors.

    Like

  8. This is an interesting post but I’m not quite buying the conclusions. I did an (admittedly very rough) estimate of what would happen on a daily basis in a combined German + Austrian grid that gets about 50-60% of its annual electricity from wind and solar. See here:

    http://jmkorhonen.net/2014/04/08/graphic-of-the-week-having-too-much-and-too-little-renewables-at-the-same-time/

    The key here is that wind and solar production tend to peak at different times of the day. Of course, not always, but at least to some extent.

    While oversupply is a major problem and likely to seriously hurt the profitability of such a system (and thus probably limit installations of variable REs to below the 50% level), there are still times when the supply is below total demand. If the price of electricity during these times is sufficient, it is at least theoretically possible to recoup the costs even if much of the generated electricity is given away for pennies.

    Furthermore, this assumes no changes in demand structure, which is a very doubtful assumption, and no energy storage, which is at least a debatable assumption. If both of these develop at least to some extent, the problems would be mitigated even further. And if the grid interconnections span wider geographical areas, there would be even more room for renewables.

    I’m personally skeptical about the incentive structures needed for this thing to work, but I’m not willing to rule out the possibility. Nevertheless, even this estimate shows that there is probably still a need for other low-carbon solutions.

    Like

  9. Correct me if I’m wrong, but is this limit more of a “you can’t integrate any more without requiring curtailment in time of high generation, thus increasing levellised cost of generation and decreasing EROEI” type thing, than a hard & fast “thou shalt not…” rule?

    If so, it would be interesting to see a chart of effective cost of renewable electricity plotted against market penetration, for various mixes. Has someone already done those sums?

    Like

  10. Peter Lang, I agree with most of what you write. My intention was to keep tight focus on the antagonistic interaction of wind and solar, which I think is an interesting insight not widely appreciated, given the support for a mix of renewable energy in particular and all clean energy in general.

    I have no problem with trying to guide the development of the energy system towards a chosen goal. Thats rather the point of this conversation. That doesn’t mean I’d support measures doomed to fail, but it also doesn’t mean we should abdicate our ability to influence outcomes to the invisible hand.

    Like

  11. I had to think this a bit more, but I think J. M. Korhonen has a point. Let us say we have 35% penetration for wind. It is operating close to its “horizon” and consequently costs have started to escalate. Now if we introduce PV into the mix it is (at least in principle) possible that costs are reduced at the same penetration of wind+PV since it is unlikely that both will be producing a lot at the same time. Both sources separately would be producing below their horizons and this could save in costs since wind is now clearly below horizon and causes less cost escalation. Overlapping production peaks will not happen most of the time. It would seem plausible that your argument is valid if one tries to push all the way to 15+35% penetration. However, it is not clear that things should be nicely monotonic as the PV share increases.

    Like

  12. Bern, yes, you’re right, this is not a hard boundary, but it marks a point of real difficulty. I wrote:

    Its not impossible to exceed it, just very difficult and expensive. Its an inflexion point; bit like peak oil, its where the easy growth ends.

    The first graphic in the article is in fact a set of plots of wind and solar market value vs penetration, from the BTI article (click the image to go there). The point of zero value for wind and solar is near enough to the penetration = 60% of capacity factor suggested in the text.

    Like

  13. Isn’t there another limit here in that PV and wind are not synchronous generators and would have to be curtailed significantly before supply (from PV+wind) met demand. You need a certain amount of inertial regulation of frequency using thermal or hydro generators. AEMO discusses this requirement in the 2030 and 2050 “All renawables” scenarios and as I recall they look at minimum requirements.

    Like

  14. The Breakthrough analysis and your extension look incredibly important if one assumes a relatively free energy market.

    But I’m not sure I quite understand your extension to multiple sources without consideration of how they interact.

    Germany currently has 27% from renewables … with wind, biomass and solar being 8.5%, 6.8% and 5.3% ..

    http://en.wikipedia.org/wiki/Renewable_energy_in_Germany

    the rest is hydro and odds and sods. Renewables are growing linearly, with some indications that solar is flattening. What would this theory predict as the maximum for that basket of wind/biomass and solar?

    Like

  15. J. M Korhonen, thanks for your comment and the interesting link.

    If wind and solar were fully anti-correlated then yes, they would be fully additive, and the wind+solar energy share would look exactly like their capacity share. That is, the second plot would be the same as the first – a linear interpolation.

    Of course, they’re not anti-correlated, and any overlap in output will cause that line to sag, resulting in a plot like the second chart. That is, the lower capacity factor energy always drives out higher capacity factor energy on more than a joule-for-joule basis.

    Just stepping back three days, for instance, on June 2, in France the wind peak correlates exactly with the solar peak. Then again on May 31, etc. See http://www.rte-france.com/en/eco2mix/eco2mix-mix-energetique-en

    You could change the demand structure. But on which days? In France again, today and yesterday the wind peaked, this time, at night. How far can we move the demand structure around from day to day? When we are talking about a sizeable fraction of energy in the grid (which we are, if we’re talking about climate impact of energy emissions), that demand management doesn’t mean shifting to a different daily pattern but moving around large fractions of overall energy use in different ways day-to-day. That is practically unincentivizable, i.e. very expensive.

    I don’t see energy storage doing much to improve things. Grid connected energy storage is really just being used to shift output within short (say 15 minute) windows for frequency control. Shifting over larger times is not in view, and I’ve written previously about the EROI impacts.

    Like

  16. Geoff, because biomass is despatchable, and has a very high capacity factor (if fuel supply is available) it doesn’t suffer significant limits itself. Despatchable output can be shaped to demand as required. So it doesn’t affect the limits for wind and solar, which remain the same.

    You can’t talk about limits for Germany per se as energy flows across national borders. You need to look at what I called the “fully connected grid”, which Breakthrough did in this post. Wind and solar are high in Germany but they spill over into neighbouring countries that are connected into the same network.

    Like

  17. @J. M. Korhonen

    “Furthermore, this assumes no changes in demand structure, which is a very doubtful assumption, and no energy storage, which is at least a debatable assumption.”

    Changing the demand structure to mach the wind and sun has some serious drawbacks. I wrote a blog post about this recently.

    http://ratdog-justbecause.blogspot.com/2015/06/demand-response-not-so-holy-grail-of.html

    In summery it would make products more costly and reduce the utility people receive from 24/7/365 electricity.

    Like

  18. As an economist, I don’t understand your logic. Let us start by ignoring fossil fuels and imagine a world where a demand of say 1 MW of energy during the day could be met through either wind or solar PV. Following your example, wind has a CF of 35% and solar a CF of 15%. So this demand could be met with either 2.9 MW of wind capacity or 6.6 MW of solar PV capacity. The constraint here has nothing to do with nameplate capacity or CF, The decision whether to supply using wind or solar PV depends on the levelized cost of the two alternatives (taking into account fixed and variable costs over the life of the asset discounted to present value).

    I can’t see how this is a case of bad renewables driving out good renewables. It may be perfectly rationale to go with a huge solar PV roll out as long as the levelled cost of solar PV is low. The CF has nothing to do with anything. You just build out however much nameplate capacity you need to provide the deliverable energy you need (as long as the price is right).

    Like

  19. I happily accept the conclusions, which as I understand it are in line with many such analyses pointing to the limitations of wind and sun in replacing our current energy supply system. But what impact have they had? The public still loves sun and wind energy. The public still loves the idea of a limitless solar resource that can power everything (Arnie said it again this week, on ABC television). And the public still holds the trump card – they vote (bizarrely) for the kind of energy technologies they love, despite lack of any real understanding. How can the analytical thinking in such a piece change public delusions?

    First, we need to accept that this particular kind of analysis is too hard for general consumption. But its important conclusion still needs to be promulgated. Involve those who our society relies on for everyday welfare and prosperity. I would like to see real-world power engineers asserting loudly that they simply can’t give the public what they want (including price and reliability) using only wind and sun.

    Then, the distinction between electricity and energy must always be emphasised. This article is about electricity but occasionally slips in referring to energy. Keep reminding the public that electricity accounts for only around 40% of total primary energy supply. There are no green energy solutions for most of the rest.

    The myths about renewables need to be attacked constantly. Always make the distinction between the grid and the individual domestic power supply that most consumers are familiar with. Indeed, for single domestic dwellings solar PV is attractive to many consumers, but the reasons are special and need constant reinforcement. Otherwise that attractiveness gets extrapolated to all electricity supply. Domestic solar PV can succeed because of its symbiosis with our existing system. It latches onto existing land, existing mounting structures (one roof per building, OK for detached houses, but not enough for multi-storey apartments etc.), existing infrastructure like roads and water, existing rewards for surplus power, and existing backup for providing the c. 99.99% reliability that our society demands. None of these symbiosis requirements are available for solar or wind farms in the isolated areas that favour such technologies. The isolation needed for large scale electricity generation is a huge burden on feasibility and cost.

    The public must never be allowed to forget that no-one can build solar PV installations or wind turbines without using fossil fuel energy.

    This is a propaganda war. Renewables have been winning because of their natural appeal and the propaganda skills of their advocates. Critics need to communicate with the best counter-strategies.

    Like

  20. Risk and well being,

    It’s not as simple as just the capacity factor. We need to use the minimum capacity factor, not the average. Look at the limit analysis herehttps://bravenewclimate.com/2009/08/16/solar-power-realities-supply-demand-storage-and-costs/ to explain the issue. This is not intended to be a realistic scenario for an electricity system but a simplification to help explain. It assumes all power for the NEM is generated by a fixed PV array power station at a single location in NSW (where we have 2 years of power output data at 1/2 intervals over 2 years). The minimum capacity factor wqs 0.7% on some days in winter. We need around 30 days of energy storage to power the NEM. The analysis gives the total cost of the PV and energy storage system with pumped hydro or NaS batteries and the area that would have to be innundated with pumped hydro reservoirs to achieve that amount of storage.

    Like

  21. @Well-Being

    This is not an economics problem it’s an engineering problem. The electric system as a whole isn’t a market it’s a machine. One really big complicated machine. Knowing the cost of individual parts doesn’t tell you much about the costs of the system as a whole if you don’t understand how the machine as a whole functions. I wrote a blog post that might help you understand the electric grid better.
    https://www.blogger.com/blogger.g?blogID=7315386763198461000#editor/target=post;postID=8759119799864529532;onPublishedMenu=posts;onClosedMenu=posts;postNum=31;src=postname

    Like

  22. “bad” (i.e. low capacity factor) energy drives out good.

    If we extend this to all electricity sources then we have an interesting outcome:

    We should only use high capacity factor sources such as gas (CCGT), coal, geothermal, biomass (depending on fuel supply) and, of course, nuclear power as all these become “good” sources.

    But we are concerned about greenhouse gas emissions.

    So that rules out coal and gas leaving us with geothermal, biomass and nuclear.

    Both geothermal and biomass are energy source constrained depending on geographic location. In Australia we have very little accessible geothermal. We have limited available arable land to grow sufficient trees, given that it takes 10-20 years before the trees are mature enough to log and chip.

    That leaves us with the only practical solution ….. nuclear power. (Sorry Helen).

    Like

  23. Gene Preston,

    Your linked article says:

    Yes, it’s possible to have 100%
    renewable energy, but it takes a lot of effort to make it work reliably. I also have other scenarios with
    varying amounts of renewables, gas, storage, and nuclear. The next step is to estimate the capital costs
    for these different scenarios.

    I look forward to that. It will be a great contribution. I hope you will publish it in a Journal.

    Like

  24. quikka1,

    Isn’t there another limit here in that PV and wind are not synchronous generators and would have to be curtailed significantly before supply (from PV+wind) met demand.

    Absolutely. This is in part a factor in the system stability reserve mentioned here:

    It should be remarked that this capacity factor rule sets too optimistic a limit. The Breakthrough writers cite estimates that only 55%-60% of grid energy could be replaced by variable sources, due to stability requirements. This means VRE share will struggle to exceed 60% of capacity factor, and the limits described above will be reduced by that factor.

    If the ceiling for introduction of variable renewables is, say, 60%, then the same logic says the penetration is limited to 60% the VRE capacity factor. That sets a maximum penetration of VRE in the fully connected grid in the case of a 100% wind mix to something in the vicinity of about 20%.

    Graham Palmer’s book “Energy In Australia” has a great discussion of the role of synchronous generators in stabilising the grid. I reviewed it here and discussed some of that content.

    Like

  25. Pingback: Less than the sum of its parts: Rethinking “all of the above” clean energy | Damn the Matrix

  26. Morning Risk and Well-Being,

    Alice recently updated a post/critique on the difficulties of managing society with large percentages of RE that cover some of the technical difficulties: http://energyskeptic.com/2015/critiques-of-mark-jacobsons-ideas-to-run-the-world-on-renewable-energy

    CA is moving towards higher percentages of RE to supply our electric grid. CASIO (the state grid operator) has indicated this:

    http://www.caiso.com/Documents/May8_2015_DeterministicStudies_nocurtailment_ExistingTrajectory_40percentRPS_R13-12-010.pdf

    “…..Conclusion
    With no curtailment of renewable resources, the CAISO identified upward and downward reserve
    and load following shortfalls and unsolved over‐generation in both the Trajectory and 40% RPS in 2024
    scenarios. The unsolved over‐generation is significant in the 40% RPS in 2024 scenario. Simply adding
    more flexible generation resources cannot solve the problem. The frequency and magnitude of the
    reserve shortfalls and unsolved over‐generation reflect conditions that do not support reliable grid
    operations. As a result alternative options must be explored, including:……”

    John’s 2+2= 3 thoughts are spot on (from a CA grid perspective) as adding in a lot more PV is leading to diminishing returns. In a virtual world (where distance doesn’t make a difference and say new cell towers aren’t needed- make that very expensive high voltage transmission lines) CA could just change its’ loading order to favor wind from other parts of the country and forgo adding more physical resources (utility and/or residential PV) that are leading to 2+2= < 3.

    I have had PV for 9 years now, and it is technically speaking operating just fine (20% average capacity factor, but as noted above this is a somewhat misleading way to look at the capability of my system to meet my demand usage). I happen to have enough property that it would be very easy technically speaking to add more PV, the question is should I if I know we already are going to have a problem managing instantaneous supply with instantaneous demand.

    Like

  27. The scenario requires that the grid takes all RE produced, adjusting the variation with dispatchable power – no storage.

    Rapid adaption to crashing from 100% RE to 0% RE would require all power to come from the rapidly responding Open Cycle Gas plants. Although in this scenario, RE reduces the gas-fuelled power to two thirds of electricity production, it requires the use of the technology which is only two thirds as efficient as a more sluggish Combined Cycle Gas plant. That is, there has been no reduction in emissions achieved by the RE.

    Sufficient grid storage to provide five minutes of full power would allow slower-responding steam driven systems (such as CCGT or nuclear) to kick in instead and provide a net carbon gain. Even so, they would have to be on standby with all systems hot and idling.

    Like

  28. Er, no, not “and nuclear”. Commercial slow neutron reactors can load-follow, but not cycling up and down that fast — each period of high power accumulates the neutron poison, xenon, which takes time to decay and burn off. The scenario quoted belongs to gas, both Open and Combined Cycle.

    Like

  29. I support the analysis of Martikainen, I think the curve is not monotonic and at low penetration of solar, there is some gain of having the two.
    It would take a hour by hour analysis over the year to have a better estimates.

    Recently a 100% renewable scenario for 2050 has been published in France which is much better done any other I’ve seen before, including hour by hour simulation of the outcome over the data of 7 years of renewable availability and demand data. It has some major weak point like the estimates about how cheap wind and solar would become, the availability and cost of massive power to gas, uses intense demand management without caring about incentivization (a dictator just displaces the demand as much as physically possible, people have no say and no compensation), but it brings some very useful data.

    One of them is that the actual optimum with a climate like France (highest consumption in winter, where wind produces most and solar produces mostly in summer) is a capacity for solar that is one third of the one of wind. And I expect simulation would show it partially holds true, even without the optimistic assumption of that scenario about storage and demand displacement.

    Like

  30. John, thanks for the reply.

    Perhaps the chart could be adjusted by taking into account measured correlation between wind and solar power. Were the two perfectly anti-correlated (correlation coefficient 0), the two would of course be additive; were they perfectly correlated (coefficient 1), they would be competitive. The realistic case seems to lie somewhere in between.

    Furthermore, the geographic dispersal of particularly wind power causes some anti-correlation on its own. That seems to help a bit with reaching higher penetrations.

    BTW, I hear – haven’t checked myself – that Ireland is already operating wind power at about 60-65% of total electricity supply and they see no problems increasing the share to 75%. Someone should probably look at that case…

    Like

  31. And wind was just 53% effective at avoiding GHG emissions http://joewheatley.net/category/wind-energy/page/2/ .

    Therefore, the CO2 abatement cost with wind was nearly double what the standards estimates say, because the estimates (incorrectlky) assume 1 MWh of wind displaces the average emissions intensity of the gird in the absence of wind:

    In the NEM, wind penetration was 4.5% in 2014, and it was 78% effective at avoiding emissions http://joewheatley.net/wp-content/uploads/2015/05/sub348_Wheatley.pdf . It is likely to be about 60% at 15% penetration, which is the likely situation in 2020 under the RET. At that rate, the CO2 abatement cost will be 67% higher than the estimates that do not take CO2 abatement effectiveness into account.

    Like

  32. @Korhonen : Portugal is a better example of very high wind penetration, however it has massive electric connexion to Spain which allow it to simply export every surplus there, so it’s not very representative.

    @Clifton : It assumes IIRC around 16 GW of power to gas to convert excess of electricity to hydrogen, and then methane, and then will burn the methane back to electricity when needed. End to end efficiency is only around 33%, but as the scenario shows, this is needed or else things just won’t work, or would need investment in storage capacity that would be used only once in months, so end up horribly expensive even with very cheap storage assumption.
    The scenario is really smart in some ways, but insane in many others, including the fact the end solution would be incredibly fragile, looking more like juggling with chainsaws than any reasonable way of handling a grid. At the end however it doesn’t work, as methane generation opposite to what they believe is not carbon free since it requires a large carbon source that it only captures for a short while. Also reading carefully the scenario, it relies on strong imports form neighboring countries(“compensated” by exports at other times), and it does actually admit a major part of it will come from fossil fuel, even if the neighbors are 80% renewables as assumed.

    Here is that report in French : http://www.ademe.fr/sites/default/files/assets/documents/rapport100enr_comite.pdf

    Like

  33. I’m not sure I buy the statement “we should build wind, not solar.”

    My personal opinion is that we should build neither, since neither qualify as “sustainable energy.” They are in fact, environmental disasters, perhaps not on the scale that dangerous fossil fuels are a disaster, but any attempt to scale them to meaningful levels will help them approach dangerous fossil fuels as a disaster.

    Low rumblings about this fact are growing louder in the scientific literature, as I realize in going through it during the project on which I’m working.

    We are wasting tremendous resources – resources we really can’t spare – on the “renewable energy” fantasy. It’s not renewable, and as energy, well it’s insignificant, particularly when one looks into the money that’s been spent on it.

    Like

  34. The Guardian has a piece trying to spin some tale about the “massive growth” of PV in Europe. They link to an industry report:

    http://www.solarpowereurope.org/fileadmin/user_upload/documents/Publications/Global_Market_Outlook_2015_-2019_lr_v23.pdf

    Figure 7 is an eye opener for me. I was aware that PV new installations had declined in Europe but I did not know how much they had declined. 2014 was less than one third of 2011. So much for “exponential growth”.

    Like

  35. “Solar plus storage will be less than gas plus carbon tax. What’s the CF of storage?”

    All right, I’ll bite.

    Solar plus how much storage? And how much solar – a power station’s worth, or enough for the whole country? Which solar – PV, or prohibitively expensive CST?

    A gas plant (CCGT) will supply absolutely reliable power barring breakdowns or polar vortices, plus grid inertia from the turbines. Adding a carbon tax does not diminish the value of the service being provided. In contrast,
    a PV farm of size necessary to supply an equivalent average annual output (~4x nameplate capacity of CCGT) does not provide this service. It provides DC current in a rough bell curve inversely proportional to thickness of cloud each day, for say around 6 hours. Assuming the batteries are perfect and charge from the excess in that curve, we are still short on overnight backup because many days will be cloudy. All we can do is add more panels and battery cells, but where does that end? To cover an entire day of storm, we must at least double each – batteries to backup, and PV to charge them in advance. 2 days? A week?

    Furthermore, if these grid connected utility scale batteries – extremely expensive despite the initial assertion – are underutilised due to low sun conditions, why wouldn’t the operators charge them up from the grid overnight at cheap low demand and then discharge them the next day at peak demand times to maximise ROI and avoid the cost of idle plant – exactly like storage is currently used?

    What’s the CF of storage? To start with, by definition it doesn’t produce energy, so there’s little use applying the term, but generally the CF depends on the reliability of the generator used to charge it up. An intermittent source that may go days without producing much will depress that CF (no input=no output). A baseload, dispatchable source which potentially provides excess during low demand will mean the storage will always be charged for meeting peak demand the next day, maximising CF.

    Like

  36. I am unconvinced by this article.

    This is an important graph from the original source on the breakthrough site:

    So the peak price got cut by nearly half and somehow that is supposed to be a bad thing?

    Like

  37. the peak price got cut by nearly half and somehow that is supposed to be a bad thing?

    In other words, the marginal return of any addition to PV has been cut by nearly half.  That’s supposed to be a good thing?  Continuing this trend will eliminate any real profit that PV is supposed to make; only continued FITs could keep it in play.

    The graph shows an incipient “duck-belly curve”, or as much of one as you can see from a graph of price rather than actual demand.  You have a local minimum at about 1600h, and a secondary peak at 2000h.  If you force more PV into the system the ramp in net demand over the 1600-2000h period becomes steeper and steeper.  What this does is force the replacement of efficient CCGTs by less-efficient open-cycle GTs, increasing net carbon emissions despite the higher fraction of “renewable” generation.

    The only way around this is with storage, and not even Elon Musk plans to sell daily-cycling storage systems to the consumer.  Musk’s initial offering of PowerWall will be for backup, not load-levelling.  The problem will not be addressed until the errors in the market design are fixed, and that will put baseload power back into play in a huge way.

    Like

  38. “In other words, the marginal return of any addition to PV has been cut by nearly half. ”

    No. The market price is cut by half. we can still decide, to give only half of that to customers and the rest to the solar PV producers.

    the “ramp up” is only interesting, if you want to protect the fossil fuel industry. These types of arguments are taking sides with big power companies. Why should their interests be more important than those of the customer?

    Why is the inflexibility of coal and nuclear plants a problem of solar power?

    Apart from that, the Tesla house battery is designed for daily cycles over a 10 year warranty.

    http://rameznaam.com/2015/04/30/tesla-powerwall-battery-economics-almost-there/

    And we should also not forget about east/west distribution via better grids. solar PV in multiple countries will flatten out that “duck” again, as will storage.

    Like

  39. “The problem will not be addressed until the errors in the market design are fixed, and that will put baseload power back into play in a huge way.”

    Just looking at Australia, it seems like base load coal is pushed out first, which obviously is a good thing.

    In direct contradiction of the main article above, SA also seems to run on 35% wind alone, with added solar PV.

    http://www.businessspectator.com.au/article/2015/6/11/energy-markets/sa-coal-killed-wind-and-solar-origin-ceo-says-agl-should-worry

    Main problem, like in Germany, is a carbon price that would punish the dirty brown coal.

    But apart from that, that graph looks great:

    Like

  40. The delivered cost of energy includes not only cost of production but also the cost of distribution, the power lines. There are areas where this cost exceeds the cost of production several times. They are typically served by diesel generators. Wind, solar plus storage is a useful way to provide energy cutting fossil fuel use and SPM pollution.
    A change of technology will be useful there. The wind energy can be use to store compressed air which can be used even without conversion to electric power.
    Solar/photovoltaic with storage can work the electronic uses including radio, TV and computers. It could be used for lighting by CFL or LED lighting.
    For the remaining, nuclear remains the best bet but uranium rich Australia should use a different path.
    A fast reactor of Russian SVBR type using enriched uranium fuel. It can use local uranium production.
    A molten chloride fast reactor using plutonium created by enriched fuel fast reactor. It will close the cycle for the IFR. It is more synergetic with pyroprocessing.

    Like

  41. Why is the inflexibility of coal and nuclear plants a problem of solar power?

    This is neither stated nor implied in the article.

    The problem I describe arises because of production in excess of demand. It has nothing to do with the flexibility or otherwise of the generation mix. The CF limit would apply if the balance of grid were hydro and open cycle gas turbine.

    In direct contradiction of the main article above, SA also seems to run on 35% wind alone, with added solar PV.

    This is in no way in contradiction to the article. South Australia is not a grid, South Australia is a state. The fully connected grid which extends in part over South Australia and which draws some part of its generation from within South Australia is not yet close to the capacity factor limit on share of wind power. State borders are political boundaries, not grid boundaries. The limit applies to share of energy on the fully connected grid, not share of energy produced within an arbitrary geographical bound within a larger grid.

    I described the capacity factor limit as applying to

    fully connected grids (i.e. not limited by State boundaries)

    but I admit I could have emphasized that point a bit more.

    Like

  42. John Morgan’s point about the electricity grid and transfers across state boundaries is very well explained in “CO2 emissions savings from wind power in the NEM” http://joewheatley.net/wp-content/uploads/2015/05/sub348_Wheatley.pdf

    This analysis shows which thermal generators are being displaced by wind power. It turns out about 12 thermal generators are doing most of the back up; mostly they are OCGT and black coal in NSW. Brown coal in Victoria is not being displaced by wind. This is an excellent analysis.

    Like

  43. “The problem I describe arises because of production in excess of demand.”

    Sorry, but i read both breakthrough articles and i assumed that you use their arguments about curtailment as well.

    If you are really only looking at “production in excess of demand”, capacity factor gets a very bad predictor. of the amount of production being spilled.
    If you use the economic (or CO2) argument brought forward by the breakthrough article, you are not only looking at 100% solar output, but at a high range, in which solar output is causing disruptions. These does dramatically increase te number of days on which there is a problem.

    Do you have any numbers for any of the grids you are thinking about?

    ” fully connected grids (i.e. not limited by State boundaries) ”

    John, i did understand that. But i think you underestimate a problem of your argument, when looking at such huge grids.
    (i use the breakthrough article examples again)

    If a grid spans multiple time zones (Australia beyond south australia is such an example) , solar will get spread over time and will rarely, if ever reach a 100% output, even if we install 20% (at 15% CF) for example.

    The breakthrough article is also using the Nordic Synchronized Area as an example, which i think is plain out false.

    http://thebreakthrough.org/index.php/voices/energetics/wind-and-solar-how-far-weve-come

    The idea, that a grid with so much hydro storage is curtailed at 15% solar power is strange at best. Why not use even a tiny amount of that storage to store for example the 16th percentage point of PV output?

    Like

  44. Further, South Australia imported approximately 6 times as much supply over 2013-2014 from the rest of the NEM than it exported. The rest of the NEM, large hydro capacity not-withstanding, runs on something like 3/4 coal at any given time. Geographically correlated wind output will dictate continued large imports of mostly coal-fired supply, except conceivably in a situation of dramatic wind over-capacity, which seems to be what some are hoping will happen and the rest of the NEM can just deal with it.

    Like

  45. “Brown coal in Victoria is not being displaced by wind. ”

    This is an economic problem and has no connection to capacity factor.

    The pdf from your link is very interesting, but it does not contradict my position. The best example is Tasmania. Wind there does displace hydro, not fossil fuels on the main land. (page 16 of the pdf)

    http://joewheatley.net/wp-content/uploads/2015/05/sub348_Wheatley.pdf

    This is a price effect, as hydro is just plain out more expensive than brown coal. As i wrote above, we need a carbon tax to displace brown coal. but this has NOTHING to do with renewables. new nuclear plants also could not displace brown coal, simply because it is much cheaper!

    Like

  46. No. The market price is cut by half. we can still decide, to give only half of that to customers and the rest to the solar PV producers.

    Your little game is exposed here.  You want PV owners to be paid a high FIT to encourage them to buy and install systems, but industrial customers to pay a far lower real-time market price for e.g. aluminum production (which requires 24/7 power to keep the cells from freezing up; not at all suited for DSM for balancing).  Who pays the difference, plus the cost of transmission?  Consumers, in “environmental fees”.  There’s a limit to how much you can gouge consumers before they go broke, or just throw out the Greens.

    the “ramp up” is only interesting, if you want to protect the fossil fuel industry.

    You’re not even coherent here.  Your own graph shows that PV has fallen to roughly zero output before the evening demand peak starts ramping down.  You have to have SOMETHING to serve that demand; you can’t just force people not to cook dinner.  You’re the one protecting the FF industry; I’d rather have nuclear power with heat batteries for peaking.

    These types of arguments are taking sides with big power companies. Why should their interests be more important than those of the customer?

    Ah, finger-pointing.  I am forced to remind you, AGAIN, that the “big power companies” exist for the purpose of serving the customers.  The customers require generation, reactive power, spinning reserve, regulation and a bunch of other things that the “big power companies” mostly spare them from having to know anything about.  You certainly know nothing about them, but that doesn’t mean that the grid will stay up if they’re not handled.  Doing that requires equipment, which costs money.  Piling a bunch of your favorite hardware onto the supply side requires more equipment which costs more money.  Even if you put the existing power companies out of business, you either put in that equipment (which someone has to buy) or your grid goes down.

    Why is the inflexibility of coal and nuclear plants a problem of solar power?

    Because solar power makes the remaining load spiky and unpredictable with much lower minima, which is much more expensive to deal with than a smooth, predictable load curve with a well-defined minimum.  Someone has got to pay to deal with the consequences, and fairness says it should be those who create them.

    the Tesla house battery is designed for daily cycles over a 10 year warranty.

    The 7 kWh unit is, but it’s not on sale yet.  The 10 kWh unit is designed for weekly cycling only.

    And we should also not forget about east/west distribution via better grids.

    Who gets to pay for these “better grids” again?  Should not the additional grid build-out required for solar be billed as a cost of solar?

    You’re all about special pleading.  Someone needs to keep you honest, as hard as that is.

    Like

  47. John Morgan –

    At this point it gets very hard to add additional wind or solar. If output exceeds demand, production must be curtailed, energy stored, or consumers incentivized to use the excess energy. Curtailment is a direct economic loss to the generators. So is raising demand by lowering prices. Energy storage is very expensive and for practical purposes technically unachievable at the scale required. It also degrades the EROEI of these generators to unworkable levels.

    Just because something is a direct loss to generators doesn’t mean we have to avoid it in all circumstances. Solar PV and wind reaching peak output at the same time is rare. When it happens, curtailing wind power production isn’t difficult, and if it’s generating all the energy we can use or store, the wholesale electricity price would drop to zero so curtailing the wind power output would not financially disadvantage the wind turbine owners.

    And it’s rather disappointing to see you linking to your Catch22 article despite your having withdrawn from that discussion for three months when I pointed out to you the flaw in your argument there!

    Like

  48. sod, the idea that “hydro is just plain out more expensive than brown coal” is just silly – brown coal has fuel costs, hydro doesn’t.

    If you look at page 22 of the pdf you linked to, you’ll see:

    There is no coincident emissions reduction when wind displaces hydro generation. However displaced hydro energy is stored and is therefore available to displace thermal plant at some later time.

    Like

  49. “If you look at page 22 of the pdf you linked to, you’ll see:”

    Yes, but the article later concludes (via transmission lines to main land) that wind does indeed displace hydro instead of brown coal.

    I don t know anything about the island, but i assume that this is peak hydro and it is just cheaper to keep the coal plant running.

    This kind of thing really needs government action.

    http://joewheatley.net/wp-content/uploads/2015/05/sub348_Wheatley.pdf

    Like

  50. For readers from outside Oz, the island of Tasmania has 68,000 km2 and 500,000 people. Unlike mainland Oz, it would have plenty of hydropower compared to local demand. However it is linked to the mainland grid by 300 km of submarine DC cable, <a href="Basslink“>Basslink, rated at 500 MW. Consequently it can only balance a maximum of 500 MW of wind-and-solar.

    That is, while there is enough water above the dams. Tasmania’s catchments, snowpack and dammable valleys are small by international standards. Providing only 2% of Oz supply, Tassie’s hydro is reserved for storage, not baseload.

    Like

  51. Yes, but the article later concludes (via transmission lines to main land) that wind does indeed displace hydro instead of brown coal.

    No it doesn’t – you misunderstand it. What it’s saying is that wind power displaces some of the hydro power generated in Tasmania, and therefore, since the limiting factor is the amount of water available, it allows more Tasmanian hydro power to be generated at other times when it can be exported to the mainland more profitably.

    Most of Tasmania’s power comes from Hydro. There’s also wind, rooftop solar, gas (though that’s rarely used) and imports (though Tasmania’s a net exporter).

    There is a relatively small amount of hydro power generated on the mainland, in NSW, Queensland and Victoria. It’s used mainly for peaking, and the Victorian one also has pumped storage.

    Wheatley’s main point seems to be that it is more likely to be gas than coal that is ultimately displaced. However his conclusion may not hold as large amounts of wind power make coal less economic. A few days ago it was announced that generation of electricity from coal in SA would soon cease.

    Like

  52. I’d urge followers of this discussion to read Wheatley’s report for themselves rather than believe the interpretations being written in comments here and elsewhere. Wheatley explains clearly what is happening with the interconnectors between states including the Tasmanian interconnector and the effect of it its capacity limits. He also takes into account all plant outages throughout the year. It is important to note, the study is of the empirical data for NEM for 2014. It’s not crystal ball gazing about what may happen in the future, or in 2015. (Although it does point out that the empirical data suggests the CO2 abatement effectiveness in the NEM is likely to decrease to about 70% for a doubling of wind energy penetration, from 4.5% in 2014 to 9%).

    The main message from his report (and his analysis for Ireland, and many other analyses around the world) is that wind generation is much less effective at reducing emissions than most people have realised. Therefore, the CO2 abatement cost with wind is much higher than the current authoritative estimates.

    This is not the case for nuclear, because nuclear is dispatchable and displaces baseload coal. The conclusion that should be drawn is that nuclear’s CO2 abatement cost advantage over wind power may be some 50% more than recognised, including in most of my previous comparisons such as this one for about 90% reduction in CO2 emissions intensity of the NEM (See Figure 6): http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.363.7838&rep=rep1&type=pdf

    Like

  53. Aidan Stanger,

    Just because something is a direct loss to generators doesn’t mean we have to avoid it in all circumstances.

    Indeed not. But the more frequently it occurs the harder it is to justify addition of such generators, either by project developers if generators bear the cost, or taxpayers if subsidised, or electricity customers if the cost of idle capital passes through to them on their bills, etc. And the greater the penetration beyond this capacity factor threshold, the more frequently it does occur.

    At some point the cost becomes too much to bear, by whomever it falls upon. Whether it happens before CF%, at, or somewhat above depends on the distribution function of generation of the particular class of VRE. A highly variable and a moderately variable generator could each have exactly the same capacity factor, but will approach the CF% penetration limit differently – the economics of the more highly variable source will be impacted at lower penetration than the less variable source, because more of its generation will have to be stored/spilled/curtailed.

    This is why I called this limit an inflexion point and explained it was the point at which things become much more difficult. We may choose to bear some curtailment, but a little bit more penetration comes at rapidly increasing cost. Whether its a direct loss to generators, to taxpayers, or to energy customers, at some point related to the CF% horizon, building the next generator becomes unviable.

    Solar PV and wind reaching peak output at the same time is rare.

    Nope. Wind frequently runs at consistent high output for days at a time. These days have middays, where the solar peak is frequently found.

    When it happens, curtailing wind power production isn’t difficult, and if it’s generating all the energy we can use or store, the wholesale electricity price would drop to zero so curtailing the wind power output would not financially disadvantage the wind turbine owners.

    And here you’ve identified the problem, though you don’t seem to have recognised it. Yes, if the electricity price is zero, you don’t bring on more generation. If this happens frequently, then you don’t build more generation capacity. It happens frequently as you approach CF% penetration. So CF% marks a boundary whose influence arrests the buildout of the variable generator.

    And curtailing wind because the price is zero doesn’t financially disadvantage the operator? Really? Selling into a market that prices your product at zero, or suffering lost revenue by curtailing output, is not a financial disadvantage? I suppose if the turbines etc. were free to build, or free to finance, or free to maintain, or had infinite service life, or investors that don’t expect a return you’d be right.

    And it’s rather disappointing to see you linking to your Catch22 article despite your having withdrawn from that discussion for three months when I pointed out to you the flaw in your argument there!

    Between BNC and The Energy Collective and elsewhere there have been nearly 800 comments on that article in which it has been very robustly tested, and with little qualification the thesis has stood up to scrutiny. I’m quite comfortable referencing it. I’m sorry if I didn’t address your particular comment. Possibly it found the flaw that everyone else missed.

    Like

  54. “This is why I called this limit an inflexion point and explained it was the point at which things become much more difficult. ”

    Sorry, but the title is “less than the sum of the parts” and CF is a pretty central point of your article.

    It makes a HUGE difference, if the real problems start at double CF (or even only 1.5 times CF).

    I asked above: do you have any data supporting your claim about CF being even some sort of a limiting point?!?

    “Nope. Wind frequently runs at consistent high output for days at a time. These days have middays, where the solar peak is frequently found.”

    This is totally unsupported by daily data graphs, for example from Germany.

    http://www.agora-energiewende.de/service/aktuelle-stromdaten/?tx_agoragraphs_agoragraphs%5BinitialGraph%5D=powerGeneration&tx_agoragraphs_agoragraphs%5Bcontroller%5D=Graph

    I do not see a single big solar spike over a wind spike over the whole of the last year.

    Windy days do very rarely show much sun.

    And this is a central point of your article (“less than the sum”).

    So where is the data showing this important aspect?

    Like

  55. I found good statistical data of the combination of wind and solar:

    http://www.ise.fraunhofer.de/en/downloads-englisch/pdf-files-englisch/data-nivc-/electricity-production-from-solar-and-wind-in-germany-2014.pdf

    Page 45 of the pdf shows that 30 GW of each will rarely generate over 30 GW in total and nearly never over 35 GW.

    For the moment and without data showing the opposite i would assume its bigger (just plain out more reliable) than the sum of the parts, not less.

    Like

  56. In order to provide sufficient space for the wind power in the grid, nuclear power plants have reduced their base load generation by about 10%, lignite plants by about 30%.

    Not a single mention of cost or price in the whole article. Not a mention of the economic consequences of the government interventions to incentivise renewables and to effectively mandate the sort of economic damage resulting from the effects stated in the quote.

    If you ignore the economic costs, all that follows is irrelevant.

    Like

  57. sod,

    I asked above: do you have any data supporting your claim about CF being even some sort of a limiting point?!?

    There is quite a bit of data to this point in the Breakthrough article, looking at curtailment and pricing as a function of penetration of wind and solar, with literature references. The paper by Hirth for instance shows the cross correlation, with solar value being forced down as wind penetration increases, just as you’d expect from my article.

    This is totally unsupported by daily data graphs, for example from Germany.

    And it is totally supported by daily data graphs, for example from South Australia. Here’s average hourly wind generation for SA in 2013(from AEMO:

    That’s pretty flat generation through the day. There will obviously be differences around the world. But it won’t take you long to find days in say France in the last month with wind and solar peaks overlapping http://www.rte-france.com/en/eco2mix/eco2mix-mix-energetique-en

    Like

  58. Peter sceptically quoted someone saying – In order to provide sufficient space for the wind power in the grid, nuclear power plants have reduced their base load generation by about 10%, lignite plants by about 30%.

    Wind cannot displace baseload, unless it is wind-plus-gas. It cannot even displace gas baseload (CCGT) because wind must be balanced by fast-responding gas turbines (OCGT), without a steam stage.

    In the scenario of the article it seems that insertion of 35 units of average wind power requires a balancing insertion of 65 units of OCGT. The lower gas efficiency of OCGT (~37%) compared to CCGT (~60%) means that the net gas consumption has if anything, increased slightly (0.65*60/37=1.05).

    The scenario does seem to leave us to infer that the insertion of wind is a pointless exercise. Of course it is a simplification of reality for the sake of argument. Rather than all of the response being covered by OCGT, I would be interested to hear how much of the rise time (required by steam turbines to pick up the load of lost wind) could be covered by the spinning reserve of all turbines/generators in the grid. That may be another limit to the penetration by wind.

    Like

  59. “. Link to chart of average South Australian hourly wind generation is here:”

    Very interesting graph (though i think the different scales on left/right axis give a misleading impression.).

    But it does not say anything about wind plus solar.

    Typically high wind days are called “storm” and do not come with high solar output.

    The graph i linked above (pdf p 45) shows, that 30GW of wind and 30 GW of solar basically NEVER add up to anything like 60 GW combined.

    this does challenge you “less than the sum” claim!

    Like

  60. “The lower gas efficiency of OCGT (~37%) compared to CCGT (~60%) means that the net gas consumption has if anything, increased slightly (0.65*60/37=1.05).”

    Sorry, but this is not “simplification” but plain out false.

    The main problem at the moment are old coal plants and gas can not replace those, because coal is cheaper to run.

    your scenario assumes, that coal had been replaced with the most modern combined cycle gas plants and that we would then replace those COMPLETELY with gas plants with a higher capacity to follow load. That will not happen.

    We are talking about a supergrid,, spanning half of Australia here.

    In such a grid, up to a extreme penetration of wind/solar both gas plants will work simultaniously. We just do not need a 100% fast following back up.

    Even the wind report linked by Peter Lang above (that report is extremely sceptical of wind power!)

    gives a 70% CO2 reduction rate at double of today wind penetration (reduced from 78% today)

    http://joewheatley.net/wp-content/uploads/2015/05/sub348_Wheatley.pdf

    (pdf page 6)

    Like

  61. Even the wind report linked by Peter Lang above (that report is extremely sceptical of wind power!)

    The analysis is objective. honest and impartial. If you find any errors you should says so, not make assertions that imply it is less than objective and impartial.

    gives a 70% CO2 reduction rate at double of today wind penetration (reduced from 78% today)

    Yes. And linear projection and all else equal gives 60% effectiveness at 15% penetration (i.e which is what will need to be achieved by 2020 to meet the RET as currently legislated). That’s nearly half the effectiveness that the RE advocates claim is the effectiveness of intermittent renewables.

    The effectiveness will continue to decrease as penetration increases. As effectiveness decreases the CO2 abatement cost increases (inverse proportion). At 50% effectiveness, the CO2 abatement cost is twice the claims made by the proponents (where the CO2 abatement effectiveness was not take into account in the analysis of CO2 abatement cost – which is normally the case).

    However, nuclear does replace coal and does avoid all the emissions. Nuclear’s effectiveness is 100%. If it replaced brown coal instead of black coal, it’s effectiveness would be well above 100%. If it replaced Hazelwood, it would be about 150% effective.

    I hope I’ve explained this significant and important difference clearly.

    All we need to do to achieve this is to allow nuclear to be cheaper than fossil fuels. And all we need to do to achieve that is to encourage the IAEA, US and EU to remove the impediments that are preventing nuclear from being low cost. Here’s my recipe on how to do that: “How to make nuclear cheaper” http://judithcurry.com/2015/05/23/week-in-review-policy-and-politics-edition-6/#comment-705879

    Like

  62. John Morgan,

    Indeed not. But the more frequently it occurs the harder it is to justify addition of such generators, either by project developers if generators bear the cost, or taxpayers if subsidised, or electricity customers if the cost of idle capital passes through to them on their bills, etc. And the greater the penetration beyond this capacity factor threshold, the more frequently it does occur.

    So it invalidates your original argument. Once the requirement to never ever exceed capacity is removed, even if it replaced by one to minimise the amount of time that occurs for, it’s better to use solar as well as wind because it’s far more likely to be windy when it’s windy than it is to be sunny when it’s windy. And you’re greatly overestimating the latter situation’s occurrence. It’s rarely windy when the barometric pressure is high, but that’s usually when it’s sunniest. Cloudless windy conditions are rare. You’re treating something that may happen 1% of the time as a dealbreaker!

    When the electricity price is zero, by definition curtailing the output loses the operator zero dollars. The money is made at other times. Of course the more power renewables supply, the more they will depress prices, and while this will create great opportunities for heavy industry, it will make it less profitable to fund in the conventional way, and alternatives (such as subsidising it by taxing household electricity consumption) will have to be considered. A very high reliance on nuclear could (though not necessarily would) have a similar effect.

    Between BNC and The Energy Collective and elsewhere there have been nearly 800 comments on that article in which it has been very robustly tested, and with little qualification the thesis has stood up to scrutiny. I’m quite comfortable referencing it. I’m sorry if I didn’t address your particular comment. Possibly it found the flaw that everyone else missed.

    I wouldn’t call your Catch22 argument a thesis; for it is based on a fundamental misunderstanding of what our needs actually are.

    There is no plausible reason why an EROEI of at least 7 could be a requirement for an advanced society. All the claims otherwise assume some other limiting factor, for instance your example assumed a limited amount of oil. I’m happy to discuss this fully if you are, at whichever site you like.
    BNC MODERATOR
    I have deleted a personal comment which contravenes the Comments Policy. Your comment is full of assumptions for which you provide no references, also a breach of the Comments Policy. Repeated questioning of other commenters, when the answer has already been provided, is deemed trolling. I am placing you in the moderation queue so I can vet your responses for non-adherence to our policies. I suggest you read the BNC Comments Policy before posting again.

    Like

  63. sod and Aidan,

    this does challenge you “less than the sum” claim!

    So it invalidates your original argument. Once the requirement to never ever exceed capacity is removed …

    The original argument is not invalidated, because that is not what I’m saying and not what I wrote. What I said is that generation in excess of capacity comes at a cost: “Its not impossible to exceed it, just very difficult and expensive.”

    When the electricity price is zero, by definition curtailing the output loses the operator zero dollars. The money is made at other times.

    But overall less money is made per unit of capacity, so the whole fleet is less economic, with proportionately less revenue to service a proportionately greater fixed cost base. When the electricity sale price is zero that doesn’t mean curtailment is free, it means your business is suffering a loss of income. That can only go so far before the next wind farm developer decides his project is unviable and the rollout stops.

    The effect can be seen in economic data from Hirth 2013. Hirth models the “wind value factor” (value relative to the average price of all energy in the market) with and without matching solar present. His plot with my annotations in red:

    Without solar, 10% wind has a value factor of 84%. With equal solar penetration, 10% wind has a value factor of 74%. So wind has lost value when solar is present.

    Unfortunately data for solar is not presented. But we can say that the best case is that the value for solar is not diminished, because it is certainly not increased. In that case the value of wind and solar together is less than the value of each, at 10% penetration. Much more realistic is that solar similarly loses value as wind is added, and so the sum of the parts is reduced even further than the indicated lost value from wind.

    Hirth’s modelling shows wind loses about 50% of its value at 35% penetration and solar about 55% of its value at 15% penetration. Both wind and solar currently require economic support to expand, even when operating at their full value at low penetration. This huge loss of value at penetration ~ CF% will put the brakes on further share expansion.

    I wouldn’t call your Catch22 argument a thesis

    I would, and so does the dictionary:

    “a proposition stated or put forward for consideration, especially one to be discussed and proved or to be maintained against objections”

    And it certainly was that.

    Like

  64. John Morgan,

    “The original argument is not invalidated, because that is not what I’m saying and not what I wrote”
    I find that claim to be rather disingenuous, as you wrote “The “CF% = market share” boundary is a real limit on growth of wind and solar” then devoted most of the rest of your argument to what it would take to avoid reaching it completely, and even claimed it to be too optimistic a limit!

    You’re right that “generation in excess of capacity comes at a cost”. But the claim that “Its not impossible to exceed it, just very difficult and expensive” is false; it’s easy and relatively cheap, as the losses occur at the times when electricity is least profitable.

    It’s true that adding more renewable energy depresses prices. It suggests that it is a good idea to subsidise renewable energy, but we can discuss that some other time. Right now, I suggest you have another look at the graph you got from Hirth: it clearly shows the combination of wind and solar depressing prices less than wind only at the same overall market share. So if Hirth is right, surely that proves that your original argument is wrong?

    You certainly didn’t maintain your Catch22 argument against my objections. As I said, if you’d like to try then you’re very welcome to at a site of your choosing – though preferably not this one, as the mods wouldn’t even let me state what I thought was a more accurate description of it than a thesis (wrongly mistaking it for a personal attack). I believe it would be strongly in your professional interest to do so, for what I think are obvious reasons.

    Like

  65. A very sober introduction to problem with Wind and Solar grid integration.

    Unless there is amble dispatchable hydropower readily available there will be a limit to Wind and Solar integration.

    The only open road without hydropower is to create power dumps where excess power can be utilized.

    Opportunities include Synfuel, agriculture with artificial lighting, ocean mining etc.

    Each of these applications will only be economically rational at a fraction of the current cost of energy from Wind and solar.

    Like

  66. Opportunities…?

    Jens, what industrial equipment can run for an unpredictable minority of the time and still make a profit? It would have to be profitable enough to pay for all the extra wind and solar plant over the capacity of the backup (ie the predictable fraction).

    I suspect it would be less lossy to let the wind and solar lie idle at all times of excess, as any industry would prefer to work from predictable baseload. Said like that, wind and solar alone could never power a nation’s industry.

    Like

  67. Roger, whether something can make a profit depends on the cost of finance – few things would be profitable at 10% interest, but many would at 1%.

    It wouldn’t have to be very profitable, as its electricity costs would be very low.

    And using solar thermal power, wind and soar alone could power a nation’s industry, though whether you’d wan them to is another matter.

    Like

  68. Aidan,

    Cost of financing a project depends on the financial risk. For policy analysis we commonly use LCOE to compare the viability of projects. The all inclusive weighted average cost of capital is the discount rate (I am simplifying here). Mining projects commonly have discount rates of 15% and higher. EPRI used discount rates of 13% in about 2010 in comparing the cost of electricity technologies for Australia. The AETA reports, 2012 and 2013 Update, use 10% discount rates to compare new electricity technologies for Australia: http://industry.gov.au/Office-of-the-Chief-Economist/Publications/Documents/aeta/australian_energy_technology_assessment.pdf

    However, if the private sector was to invest in solar plants without government assistance or guarantees, the financing charges would be very high – I expect around 20% or more. It would be extremely difficult to get private finance for solar projects without government mandated incentives or loan guarantees. Solar is nowhere near to being financially viable (except for off grid and small grids) in the foreseeable future, if ever.

    Like

  69. Roger Clifton

    Your question is valid and fortunately the development in capacity factor for both solar and wind is positive.

    Especially the capacity factor for wind power is predicted to grow significantly from present day average in new US installations at 35% to 65% within a foreseeable future. http://apps2.eere.energy.gov/wind/windexchange/windmaps/resource_potential.asp#states

    The list over possible uses for cheap excess power is long but in temperate climates the most useful power dump would probably be for heating either direct or via heat pumps. Other possible candidates for usage of cheap power could be industrial processes.

    John Morgan has however himself delivered a very interesting analysis of the potential for producing Synfuel based upon electricity and seawater. https://bravenewclimate.com/2013/01/16/zero-emission-synfuel-from-seawater/

    This phd thesis from 2010 is looking on the potential in more detail and concludes that electricity below $0,02/kWh would be competitive but with caution about the Synfuel plant utilization factor. http://orbit.dtu.dk/fedora/objects/orbit:83036/datastreams/file_5193307/content

    At the moment the US wind PPA contracts are still too expensive but with the significant price drop in recent time there is reason to expect that the relevant price points can be achieved.

    Like

  70. Peter Lang

    Any sane house owner will increase their house loan and get very cheap long term financing of their solar system.

    In many cases the return on investment is achieved directly when the solar system is roof integrated and therefore partly substitute purchase of roofing materials because the net cost of installation is match by the property value increase.

    As for utility scale solar you are correct in assuming that the ITC is instrumental for the investment decision. The ITC will however not be continued after 2017 so by that time you will be able to see what will happen with utility scale solar. I expect that solar expansion will continue much as today because the anticipation in the business is that LCOE drop will make up for the reduced government funding.

    A possible way to make utility solar more competitive would be to use dual axis trackers and allow utilization of the land for grazing or the like, which would lower the cost of land occupied and create solar with significantly higher capacity factor.

    Like

  71. Pingback: The capacity factor of wind | Brave New Climate

  72. well you’d better tell Scotland because they are already at 100% RE (net) and planning to go to 300% RE which will certainly have them 100% RE 24×7 especially with the UK to Norway HVDC interlink going in to make use of fjord located Pumped Hydro Energy Storage which is already in existence and you claim is unworkable and undermines ROI for RE generation assets.
    BNC MODERATOR
    As per BNC Comments Policy, please supply refs to back up your assertions. Thank you.

    Like

  73. The Scotland data up to 2012 seem to support exactly what John is saying … electricity (not all energy, just the electric bit) is 40% renewable and but they already have an excess that
    needs exporting … about 25 percent. The ability to export your excess will vanish as the destinations expand their wind+solar.

    https://en.wikipedia.org/wiki/Renewable_energy_in_Scotland

    Far from 100%, they are aiming for 50% of electricity as renewable by 2020 …

    https://en.wikipedia.org/wiki/Wind_power_in_Scotland

    They can exceed their wind CF because they have 1300 MW of hydro … useful baseload … on top of the 5328 MW of wind as of March 2015.

    The wikipedia sources are given, but I confess to not checking right back to the primary data!

    Like

  74. John,

    I wonder if you could estimate the GWh of storage capacity and GW of wing generation capacity that would be needed to provide constant power from wind power alone using data you’ve already extracted for NEM?

    Roger Andrews looked at scenarios to supply a constant 25 GW with wind power assuming UK’s February 2013 wind power generation (‘a typical winter month’) http://euanmearns.com/estimating-storage-requirements-at-high-levels-of-wind-penetration/ . He assumed 100% efficiency of energy storage. Of the five scenarios he considered, the cheapest option is 700 GWh energy storage and 200 GW wind power capacity. For convenience I’ve converted from 25 GW constant power to a W constant power. To supply a constant W with wind power in UK in February 2013, with no energy storage losses, the least cost option would be 28 Wh of energy storage and 8 W wind power capacity. If we assume the energy storage is 80% efficient we’d need 10 W wind power capacity.

    At $3/Wh for energy storage and $2.2/W for wind power, the total capital cost would be $84 + $22 = $106/W average power delivered. Compare this with nuclear at $6.5/W (i.e. $5.5/W at 85% CF, from EIA https://www.eia.gov/analysis/studies/powerplants/capitalcost/ )

    This analysis suggests to supply constant power from 100% wind power would cost sixteen times more than with nuclear power. (This is assuming optimistic costs for wind energy and energy storage at high penetration; Furthermore, February 2013 is ‘an average winter month’, not a worst case, so the costs would be higher to cover for the worst case month.)

    Would you be able to do an analysis of the storage capacity and wind power capacity for the NEM equivalent to what Roger Andrews did for the UK?

    Like

  75. Correction: Assumed energy storage cost should be $0.3/W, not $3/W. The paragraph should read:
    “At $0.3/Wh for energy storage and $2.2/W for wind power, the total capital cost would be $8.4 + $22 = $30.4W average power delivered. Compare this with nuclear at $6.5/W (i.e. $5.5/W at 85% CF, from EIA https://www.eia.gov/analysis/studies/powerplants/capitalcost/ ). This analysis suggests to supply constant power from 100% wind power would cost about 5 times more than with nuclear power. (assuming optimistic costs for wind energy and energy storage at high penetration; and assuming Feb 2013 represents the worst case wind profile).

    Like

  76. Pingback: Capacity Factors And Coffee Shops: A Beginner's Guide To Understanding The Challenges Facing Wind Farms - New Matilda

  77. Pingback: Wind Industry Loses Support of Lunatic Fringe: Green-Left Blog ‘New Matilda’ Turns Against the Wind Power Fraud – STOP THESE THINGS

Leave a Reply (Markdown is enabled)

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s