David Bidstrup: Solar nightmares

A report on 8 February told of a plan to build a 44 MW solar system “to deliver more reliable power to the grid”. The cost would be $90 million.

I have done some research into solar systems, so thought I would look at this proposal from a state wide power consumption viewpoint.

Using AEMO numbers the daily average power consumption in SA for 2017 was 31,600 MWh.

Using a solar calculator that takes information from stations that record Typical Meteorological Year (TMY) data on solar insolation, cloudiness, temperature and whatever, the 44 MW solar system might generate 69,800 MWh in a year.

Averages are misleading and do not reflect the reality that solar system output declines significantly throughout the year as the sun angle changes in the sky. When the output is shown as monthly figures the picture becomes a bit clearer.

In summer, (December, January and February) average daily output is 252 MWh. In autumn, (March, April, and May) it is 168. In winter, (June, July and August) it is 123 MWh and in spring, (September, October and November) it is 224 MWh.

The table below shows the comparison between daily average consumption and the contribution from the solar farm.

Season State average daily consumption MWh Solar contribution MWh Solar contribution as % of daily consumption
Summer 31,806 252 0.79%
Autumn 31,740 168 0.53%
Winter 35,043 123 0.35%
Spring 27,983 224 0.80%

 

The output never gets to 1% of the daily consumption and when we consider the times that output is delivered it gets worse.

In January the system operates for about 11 hours but output does not become significant until around 9 a.m. and declines after 3 p.m. 90% of the daily output is between those hours but is not a constant amount ramping up from 10% at 9 a.m., reaching 13% at 12 noon and declining to 9% at 3 p.m.

In June nothing happens until 8 a.m. when output is 8%. It ramps up to 20% at noon and then falls off the cliff at 2 p.m.

How this contribution gives “more reliable power to the grid” is beyond me. The proponents also promise a “21 MW battery” which is a nonsense statement – how long will that battery be able to supply energy (MWh)? Probably 5 minutes.

People often confuse power and energy and think that “adding 44 MW” to the system means something rather than looking at things in the context of the contribution to consumption, (MWh). In this case, that contribution is negligible, variable, and intermittent and of no real use except for the proponents to garner renewable energy subsidies, without which there would be no business case.

This entry was posted in Guest Post. Bookmark the permalink.

28 Responses to David Bidstrup: Solar nightmares

  1. Roberto

    But it’s a ‘report’. Written by someone wearing a white coat. It must be right.

  2. Cynic of Ayr

    So, in short, 90 million dollars buys 0.80% of daily power consumed, at best. Let’s round it up to 1%.
    (Use a yearly average of 31,000 MWh per day, so we don’t get bogged down in seasonal changes.)
    The maths then is 100/1 = 100. (Natch!) We need panels 100 times as big.
    100 x 90 million = 9,000 million dollars for a 100% solar farm that gives 31,000 MWh per day.
    Cheap! What’s the hold up? Must be viable, surely? We could ask the Greens?

    Oh? It only works for about a third of a 24 hour day you say? Not to worry! Batteries!
    All that is needed is storage for the 2/3 of the day the solar panels aren’t working.
    2/3 of 31,000 = 20,600 MWh needs to be stored in batteries.

    Info varies, but the SA one is 130Mhw for 50 million bucks. Soooo, 31,000 MWh should only cost 31,000 / 130 x 50 million. Lesseee…. that’s 11,923 million bucks! Say 10,000 million, with savings in scale.
    Still cheap. DiNatale and Wetherill say so!

    Ah crap!! The steeenking solar farm only produces for 1/3 of the day, so it’s actual output in that 1/3 has to be increased by 3, to charge the batteries during that 1/3 of a day, as well as usual consumption.
    So, the 31,000 MWH farm now needs to be 31,000 x 3 = 93,000 MW. 9,000 million dollars for 31,000 MWh now becomes 9,000 x 3 = 27,000 million dollars.
    So, add ’em together, 10,000 million for batteries plus 27,000 million for the solar farm,= 37,000 million for 100% solar power.
    Hmmm. Getting a bit outside a decent Lotto win!

    OK, my argument is full of holes – big holes – what with MW, MWh, scale, and all that, but the end result is the same. It can’t be paid for.
    The real question is, why the hell can’t the bright people figure this out, when dummies like me can?

  3. W Hogg

    Cynic, you forgot that solar isn’t binary – it tails off outside lunch time.

    And battery losses.

    And that batteries last 8 years so this is a recurring expenditure.

    And that the actual power could have been exceeded 50x over by keeping the coal plant open. It didn’t need to be built, or connected to the grid. It had a marginal cost of zero.

  4. Dr Faustus

    People often confuse power and energy and think that “adding 44 MW” to the system means something rather than looking at things in the context of the contribution to consumption, (MWh).

    The renewable rent-seeking industry relies on that very phenomenon. For example:

    Rooftop solar provides 48% of South Australia power, pushing grid demand to record low

    That is a phenomenal share of 47.8 per cent of the state’s electricity demand being met by rooftop solar (compares with 36 per cent in the previous record last week) and is clearly a record for South Australia, and for that matter in any large grid anywhere in the world.

    The fact is this phenomenal event happened for one 6-minute period (on a cool, Spring Sunday). When you look at the 24-hour graph you notice that rooftop solar actually contributed about 7% of the total daily consumption.
    And that was on a good day…

  5. Bruce of Newcastle

    the 44 MW solar system might generate 69,800 MWh in a year

    Dividing this by 8,760 hrs per year gives 8 MW not 44.

    So you would have to build 100 of these at a cost of $9 billion to equal one smallish 900 MW coal firepower station (ie 89% availability). Then he has to have 900 MW of low efficiency back up gas to cope with the swings as clouds come over and night falls. That’s another billion.

    Oh and the batteries all have to be replaced every 8 years or so, which is about two thirds of the original cost. Like knocking down two thirds of your power station and rebuilding it every 8 years.

    Crazy.

  6. H B Bear

    At a system level 44MW is a rounding error.

  7. Cynic of Ayr

    W Hogg.
    Sure. I said it was full of big holes.
    Your first sentence is merely a detail. 1/3 over a day, roughly covers that. I wasn’t going to attempt to give figures minute by minute.
    Battery losses. Again, a detail. You yourself haven’t mentioned the cost of toilet paper for the workers at the plant.
    You mentioned Coal. I didn’t. It was an article on Solar, not Coal.
    Your corrections, although somewhat valid, contribute what to the article?
    And, do your submissions make my figures worse or better?
    I’d have really liked your input on my last question.

  8. RobK

    The 21MW battery in this case has a buffer or dampener function. Most people think of solar output under clear conditions and visualize a sine or bell curve. There are many days when there is intermittent cloud cover. On such days solar panel output is a series of spikes, many of the troughs approach zero output. (Solar thermal behaves a little bit better). These spikes and troughs send crazy surges through out the grid. In my view a 44MW solar farm with a 21MW accumulator will only ease the slope of the peaks and troughs sufficiently for another supplier to pick it up more reliably. It will ease the surges, not eliminate them. Often the alternative supply will be in an other location possibly far away. So it will still test the grid control equipment every few miniutes throughout the day when certain conditions prevail, just not quite as severely in one sense. Still more power conditioning equipment is required because large random surges circulating about the grid will cause a changing impedence and capacitence effects on the power factor (P.F., also known as reactive power). Finkel mentioned this in his report, siting power conditioners or rotary synchronous capacitors (these are essentially idling synchronous generators reprocessing wayward P.F.). Finkel didn’t say who should pay, but someone will. If P.F. strays too far from unity, transmission becomes increasingly inefficient and difficult to control. There is a lot more redesign of the grid to come yet. These cost havent been taken into account. The implied inefficiencies mean even less of the solar power is a useable commodity, more cost is incurred.

  9. RobK

    My comment is in addition to the legitimate problems illustrated by the post and other comments above. They are all on the right track.

  10. JohnA

    W Hogg, we get your point:

    And that the actual power could have been exceeded 50x over by keeping the coal plant open. It didn’t need to be built, or connected to the grid. It had a marginal cost of zero.

    but didn’t you really mean to say that its marginal contribution is zero?

  11. W Hogg

    And, do your submissions make my figures worse or better?
    I’d have really liked your input on my last question.

    Makes them much, much worse. Solar tends to run sub 20% of capacity not a third (unless tracking the sun) and battery losses around 30% so between them probably double your numbers again.

    The only tiny benefit for solar is that peak load in DPRSA is summer lunchtime. So if you fill a low-rain hot city with solar panels it will eventually have a slight load levelling benefit and tend to take out some of the $14000 spikes.

  12. RobK

    With increasing input from Renewables, fault current discrimination and ground current issues will increase, complicating instrumentation and control of the grid. Along with various demand management schemes mooted, the complexity and cost of the grid will increase whilst its reliability an resilience will decrease with the increased proportion of renewables.

  13. Dr Faustus

    With increasing input from Renewables, fault current discrimination and ground current issues will increase, complicating instrumentation and control of the grid.

    Pretty sure that was the technical cause of the recent Victorian blackouts.

  14. Chris M

    Hopefully in five or so years time batteries will be much cheaper and better as they aren’t financially viable at the moment. Then enough of them could largely provide for the late afternoon / early evening demand that remains a problem for solar.

    I suspect domestic solar with batteries are going to become almost essential in SA simply because we cant / wont be able to afford the rates being charged. Medium business and shopping centres will have their own systems with generators. Self sufficiency by compulsion.

    Personally I want a mini-nuclear generator that lasts as long as your house.

  15. egg_

    At a system level 44MW is a rounding error.

    Yup, it’s the vibe that counts.

  16. RobK

    Personally I want a mini-nuclear generator that lasts as long as your house.

    Realistically, you would be better off with a centeralised grid as we had, all be it nuclear if CO2 abatement is your thing.
    Up to now industrial demand for energy has underwritten cheap domestic energy. That model is being wrecked. Costs will rise.

  17. RobK

     In this case, that contribution is negligible, variable, and intermittent and of no real use except for the proponents to garner renewable energy subsidies, without which there would be no business case.

    …..and in doing so ruins the business model of a centralized system that worked well, is well understood and easy to adapt. Again, any amounts of renewables over around 15% will see costs escalating in transmission, conditioning, storage and backup. We are being sold a pup.

  18. manalive

    Hopefully in five or so years time batteries will be much cheaper and better as they aren’t financially viable at the moment …

    That comment highlights the underlying folly of the adoption intermittent renewables, it’s all based on wishful thinking.
    There is no Moore’s Law for batteries: “… significant improvement in battery capacity can only be made by changing to a different chemistry …”.

  19. duncanm

    People often confuse power and energy

    Worse – they don’t even understand the orders or magnitude involved.

    Mega — must be big.

  20. Chris

    s/or/of/

    vi vi vi vi vi
    So dance the samba, so dance the samba
    vi

  21. Tel

    There is no Moore’s Law for batteries: “… significant improvement in battery capacity can only be made by changing to a different chemistry …”.

    Hmmm, that article looks like crap to me, especially this:

    For example, the maximum speed of cars, planes, or ships does not increase exponentially; maximum speed barely increases at all.

    Really? I just quickly looked up the fastest F1 lap time: Juan Pablo Montoya, 2004 Italian Grand Prix 262 kph. Then I checked some historic lap times: Melbourne 1997 average speed 211 kph, 1961 Monaco Grand Prix (Stirling Moss) average lap speed 114 kph. Looks like there’s been significant improvement. This “maximum speed barely increases at all” is phooey. Imagine a modern F-16 vs a WWII Spitfire… don’t yank my crank of course there’s been huge increases in maximum speed. If you really want to look at maximum speed take a look at what the Russians are doing with hypersonic cruise missiles (but no one is crazy enough to stick a pilot in one of those, not even the Russians).

    The thing is that different technologies improve over different timescales depending on where the money gets invested, and what options we discover. In the CPU industry nothing fundamental has changed since the idea of the silicon transistor and the Von-Neumann processor architecture… but we have make a heck of a lot of incremental improvements along the way.

    I don’t believe for a moment there’s anything preventing batteries from significantly improving. Indeed, existing Lithium chemistry has perfectly reasonable energy density (not as high as petrol, but still plenty of everyday non-military use). The problem with the Lithium batteries is nothing to do with energy density, these are the problems:
    [1] Too expensive.
    [2] Slow to charge.
    [3] Very heat sensitive, and easily catch fire.
    [4] Severely limited number of deep charge cycles before they fall apart.

    Well none of those require a fundamental new chemistry, they require better construction and more efficient processes. Plenty of room for incremental improvement. We have already had improvement, when I were a lad it was either led acid or nickle cadmium and that was the end of that.

  22. Bruce of Newcastle

    Tel – The problem with batteries in cars is that the energy in the battery has to carry the battery around.

    Same goes for a petrol tank, except it is a tenth the mass of a battery and offers twice to three times the range. Which is why the EV’s have such miserable range – the battery is already too big for the frame.

    There is nothing to stop centralized production of methanol or similar fuels, which can be done quite greenly and less expensively than Li ion batteries right now. Then the mass savings the EV’s have developed to get the most out of their batteries can be used to get the most out of the green fuel.

    As for grid storage batteries why use lithium? That is just as religious a decision as pushing EV’s is. Neither application stacks up, except that lithum is holy to Gaia or something.

    The battery system which should be used for grid applications is sodium-sulphur. Which can’t be used for EV’s because of the operating temperature. It’s what grid batteries hitherto have been made of – big ones, none of this namby-pamby chainlinked powerwall idiocity. Sodium and sulphur are so common there can’t possibly be a shortage, whereas lithium and cobalt are skyrocketing in price because of the production bottlenecks, and the rarity of cobalt.

  23. manalive

    The thing is that different technologies improve over different timescales depending on where the money gets invested, and what options we discover …

    Maybe maybe not, intermittent ‘renewables’ are useless for large scale electricity generation without the advanced storage technology that’s allegedly ‘just around the corner’, the cart is being put before the horse.

  24. RobK

    BoN,
    I agree with what you said, especially regarding EVs. Even to the proponents of Hydrogen fuel, I suggest to them; “why not hang the H on a backbone of C and sell it as a known liquid through existing distribution network.
    Regarding industrial scale storage batteries: yes they are available for niche applications now at high expense. They may well become more economic in time. If so, the system of power generation to benefit most is baseload with the appropriate storage to displace the role of pumped storage (which is very site specific). Baseload and storage cycles diurnally and its returns are predictable. Renewables cycle all over the place including weeks, seasons and variations from year to year, making a ROI more difficult to estimate.

  25. Cynic of Ayr

    Tel.
    Really? I just quickly looked up the fastest F1 lap time: Juan Pablo Montoya, 2004 Italian Grand Prix 262 kph. Then I checked some historic lap times: Melbourne 1997 average speed 211 kph, 1961 Monaco Grand Prix (Stirling Moss) average lap speed 114 kph. Looks like there’s been significant improvement.
    Just a thought, but if you could take a glance at your statement and see the classic error?
    Hint. Bold.
    Now, perhaps you could go back to your source, and quote some times and speeds for the same circuit over, say, 20 years. Even then you’d be in trouble, as circuits change for safety reasons.
    Like you said. Really?

  26. Tel

    Cynic, admittedly only a thought experiment, but suppose I can put Stirling Moss, at the peak of his career, in his Lotus built with 1961 technology and you can choose any of the top drivers and top cars from 2004 with their twin turbos and you can also choose any track ever used in the history of F1… if real money was at stake and you have to bet on the 1961 driver, would you go for it?

  27. Tel

    Same goes for a petrol tank, except it is a tenth the mass of a battery and offers twice to three times the range. Which is why the EV’s have such miserable range – the battery is already too big for the frame.

    Sure, liquid fuel vehicles have longer range, so what?

    I’ve never once pushed by car to the limit of its range on one tank. Not remotely close.

    Even on long country drives I still go past many petrol stations before the tank is low enough to make it worth stopping… and by far the majority of vehicle journeys are not long country drives.

Comments are closed.