Is ‘solar on steroids’ a solution?


There is lots of stuff in the carbon tax policy – and I am not going to try and cover it. But one part of the ABC’s story worried me. There will be an extra $10b for ‘large scale renewable projects’ and the assumption appears to be that much of this will go to solar. I hope this assumption is wrong for the following reason.

In the absence of mass power storage, solar electricity may reduce carbon emissions but it does so in a way that pushes up electricity prices. The problem is that, without storage, solar power cannot provide peak-load security. This means that cheap solar simply ‘crowds’ the non-solar capacity that is needed for peak security so that it must recover its fixed costs over a shorter period of time with less dispatch. This means lower carbon emissions but higher electricity prices. 

As an example, suppose peak demand is 10,000 MW. Further suppose that this peak may occur on a hot summer’s day (due to airconditioning) or a cold winter’s night (due to electric heating). Solar helps meet capacity on a hot summer’s day, but doesn’t help on a winter’s night. So on the winter’s night the price will rise until the non-solar generation is producing 10,000 MW and covering the peak load.

But the non-solar power stations have to cover their fixed costs. Let’s suppose that they previously covered these costs by the price rising to $5000 per MWh in both the summer and winter peak periods. Now more solar generation means that the price in the summer peak is lower and/or less non-solar is dispatched in the summer peak. Either way, the non-solar generators must be earning more money in the winter peak if they are to cover their fixed costs (given that they earn less in summer). So increased solar will push up the price of electricity in the winter peak.

A bit more economics (assume free entry in non-solar generation) shows that, if there is to be enough generation capacity to cover the winter peak (or, more generally, give system security even when solar is not operating) then non-solar plants must still be accruing the same annual revenue (to cover their fixed costs) with or without solar. As solar reduces the amount of electricity sold by the non-solar generators, this means that consumers must be paying higher electricity prices on average.

Note that this argument does not depend on the price or efficiency of solar power. Solar power could be dirt cheap and extraordinarily efficient. Rather, the argument depends on the inability of solar power to provide system security for peaks in demand that arise at night.

Another way to think of this is the following: solar adds to generation capacity at some times but can’t be used for system security at all times. So if there is an increase in solar generation, customers must pay to cover all the non-solar generation needed for system security and also pay for any new solar capacity. This means they will be paying more in total for electricity.

Now there are a range of assumptions here. First, solar can help system security to the extent that the electricity network has summer peaks (e.g. South Australia and Victoria) rather than winter peaks (e.g. New South Wales). So solar may enable some decommissioning of non-solar plant in some states.

Second, the analysis is long-run. In the short term, the existing generators have sunk their costs and if they do not recover them then they will keep operating, so long as they cover average variable costs. So solar power may drive down electricity prices in the short term. Some existing non-solar generation owners may go bankrupt but someone else will buy the plants at a low price and keep operating them. But as older non-solar plants need maintenance and refurbishment, they will be taken out of the system if the electricity price is not high enough to pay for their continued operation.

Third, as the mix of generation needed changes, some coal-fired base-load generation may be pushed out and replaced by gas-fired peaking plant. Solar may take over some of the ‘base load’ responsibilities. But the gas-fired peaking plant will still be needed for system security and it will need to earn an economic return.

Note, however, that the above argument disappears with renewables that do not depend on the the vagaries of nature (the sun shining or the wind blowing). So, for example, tidal power stations may be an appropriate solution (I would have to rely on the technical people for this). It also, of course, disappears if electricity becomes economically storable in large amounts.

All of which means that, if there is $10b on the table for renewable energy, large scale solar and wind generation projects may be the exact wrong place to start.



14 Responses to "Is ‘solar on steroids’ a solution?"
  1. First of all, note that the summer peak is MUCH higher than the winter peak. In summer, air conditioners run during the day, at the same time that all the factories and offices are using power. Winter heating mostly occurs outside of office hours, tends to be gas powered, and is only needed in the southern states.
    The main downside I can see with solar power is that it matches peak demand much better than coal, so as the most polluting coal power station are taken out of service, we can say goodbye to really cheap off-peak power. And that’s going to hurt the business case for electric cars.
    As for other types of renewables, when you’re talking gigawatts there’s not much choice. We’ve tapped the available hydro sources, and there are limited tidal opportunities (even if it was cost-effective). Hot-rock geothermal has potential, but it’s a long way off. But solar works, and it can scale up indefinitely.

  2. There are potential solutions for the storage problem: (1) pumped hydro storage (I believe that the Snowy River scheme used cheap baseload coal power to pump water back into the system and then release the pumped capacity for sale of electricity at peak load prices – a good arbitrage opportunity) and (2) molten salts that store solar heat for a given time period.

    Solar thermal ‘farms’ have been written of as a large-scale solar solution but the prototypes operating in Spain are quite expensive in terms of tariff rates.

  3. This analysis assumes that electricity will continue to be charged at a single average cost day or night, year round.  When consumers are on tariffs that can charge more at certain times of the day, then it is only those using the expensive night-time power that need to pay for it.  It’s already possible to select such tariffs as a retail electricity customer – I suspect that they will simply become more common.
    It’s also worth noting that this disparity in the cost of electricity between daytime and nighttime performs two important functions: it pushes consumers towards using the less scarce resource (daytime-generated electricity), and it encourages the development of energy storage technologies.

  4. A new solar thermal station in Spain has just clocked up its first 24hour period of uninterrupted electrical supply to the grid. While it is described as commercial scale, its 19.9 MW capacity is still small compared with typical coal-fired base load generators.

    The report gives no information on the economics of the enterprise. One of the partners in the development is an Abu Dhabi company, from a nation where the economics of electricity productions are rather different from Australia. 

  5. How variable is coal power/gas power to consumption demand? If the demand side decreases can they turn off production “in real time”. I’m imagining a senario similar to turning off your car at the lights. If solar is generating sufficient capacity are you able to turn off coal/gas and bring it back on for colder nights

  6. This is why the opposition leader is concerned about the quality of econonomists. “Solar” and “crowding out” is an oxymoron. $10b will buy about 1000MW of solar generation (peak). Thats 200MW continuous if you could store it. About 0.7% of Australia’s demand. All you get is a massive white elephant. If you want to really put a dent in fossil fuels with solar it will cost about $1.5T.

  7. To answer “Debt Consolidation Nation”‘s question about supply variability – sure you can turn off coal power stations, but it takes over 24 hours to start them up again, so they can’t meet the daily peak/off-peak cycle. They can match seasonal fluctuations by going down for maintenance during winter.
    Gas power plants can be switched on/off to meet demand, but they’re very expensive assets to leave sitting idle.

  8. Matt
    First – thanks for answering the question on supply variability and coal-fired base-load plant. Exactly correct. They are inflexible – unlike gas plants which have a significantly higher marginal cost (approximately 50% to 100% higher than that of coal fired).
    But, your comment on the peak is not quite right. If you look at the AEMO market reports you will find that for NSW the summer peak on Feb 1 2011 (highest in Jan and Feb and a record) was 14,744 MW. The winter peak on 29 July 2010 (highest in July and June) was 12,400 MW. So – yes the summer peak was higher – but I am not sure ‘MUCH’ is correct (of course, ‘much’ tends to be in the eye of the beholder). The winter peak was at approximately 5.30pm (around sunset). 
    Indeed, reading the reports, it looks like Tasmania only has winter peaks – but I have only reviewed four months of reports.
    Finally, the summer peaks are much more common in NSW. But system security depends on the size of peak demand. 

  9. rick gann:

    $1.5T – what a Big Scary Number. The per GW cost that you quote is around that of the Spanish plant, but that cost should drop substantially for a couple of reasons. Firstly, as more such plants are built, knowhow accumulation, manufacturing improvements and economies of scale will come into effect. Secondly, the Spanish cost includes backup gas generation. With a network of solar plants able to stagger their off-grid periods, amount of backup gas generation requirement will be much reduced.

    So say by 2020 the $1.5T has dropped to $750B. By then Australia’s GDP should be well over a trillion dollars/year. If construction is spread from 2030-2050, that requires an investment of $25B a year. 

    This still looks unaffordable but the gap between that and affordability looks more like a technology challenge than simply an impossibility. 

  10. MikeM

    I repeat, our opposition leader is very concerned about the quality of economists.

    With the mere stroke of a mouse you cut my conservative $1.5T to a “mere” $750B.

    Solar power technology is not exactly cutting edge – most of the components are mature technologies so learning curves don’t apply – even if you get it for free it covers massive areas, requires massive construction effort and huge transmission costs (not included in my conservative $1.5T). Transmission costs have to be allocated to 20% availability factor instead of gas/coal 95%+.

    Then you say it is a meresnip at $25B per annum for 20 years- another sweep of the mouse.

    On a per household basis this level of investment is frankly astonishing.
    I know economists don’t like to deal with tacky tech stuff but solar energy has low energy concentration – even if you get 100% efficiency (100%+ is not allowed in our universe!) – It will still be marginally better than now.

    Solar will never cut it – a bit like the Great Wall of China – it will cause great cost to generations, ruin an economy many times over and in two or three centuries become a marginally cash flow positive (if you cook the books) tourist attraction.



  11. I don’t think many people especially politicians understand the difference between fixed and variable costs, let alone the way fixed costs are allocated.  

%d bloggers like this: