Forecasting emerging technology like Solar Energy

The US Energy Information Administration (EIA) and the International Energy Agency (IEA) have had terrible projections of the energy market especially in regards to solar and wind.

Quartz and Steinbuch and Ramez Naam among others have covered this massive year over year failing.

In a tweet in May, 2017, the IEA said that their WEO scenarios do not forecast the future but provide a yearly picture of what will happen if nothing changes. They are basically saying that your car will be 400 miles away in 20 hours if it is currently moving at 20 mph.

This was ok when the energy world was mostly static from 1980 to 2000. They had time to adjust their yearly report when price changes to oil, gas and coal slowly adjusted the situation or when China expanded its economy over many years. They just had to make moderate GDP projections over decades to get to new overall world consumption.

They are getting grief for doing no homework in terms of solar projects that were funded and in the process of being built in 2016 for their 2017-2026 projection.

Basically IEA and EIA are not even trying to get their projections right for a changing world of energy.

Ramez Naam has been more accurate in projecting the growth and price reduction in solar power.

Ramez uses a learning curve for solar energy costs based on the doubling factor.

Possibly accurate projections and scenarios for a changing world of technology need to consider physics, projected engineering, supply chains and building up economic models from first principles.

28 thoughts on “Forecasting emerging technology like Solar Energy”

  1. The German experience (Hirth et al) is that each percentage point penetration increase yields a 4 percentage point reduction in value in high-resolution spot-markets. So the straight-line red comparison line (the natural gas price) should be replaced with a line that shows the falling average value of solar in the spot markets. Then you’ll find that value reduction catches up with cost reduction somewhere in the 8-16% range, making solar unappealing again.

  2. To Goat Guy: Caution needs to be exercised in estimating solar system payback periods. If these calculations are based on a single time period, or worse, on the time of peak insolation, the payback period would be significantly underestimated. For example, at the NOAA solar measuring station in Desert Rock, Nevada the measured insolation average for June, 2014 was about 290 watts/square meter, but the isolation average at this location in January was only about 105 watts/square meter. If one designs a solar panel system for an approximately fixed load, like an electric vehicle used for local trips or hot water system, the solar system would be too small in January, if it was based on June sizing measurements. Conversely, if the size of the solar system was based on January measurements you would have excess generation which could lead to “dumping” or “Duck Curves”, which degrades the overall system’s economics. It would appear that to avoid these extremes one would have to add energy storage, but batteries remain expensive. In any case, the cost for storage would have to be factored in to the payback period analysis.


    Apply For an urgent Loan Today… apply for Private loan, commercial and Personal Loans with very Minimal Interest Rate as Low as 3% within 1 to 33 years payback period, to any part of the world.

    We give out loans within the range of $10,000 to $10,000,000(EUROS, POUNDS, RUBLES, PESOS) or any other currency you need the loan. Our loans are well insured for maximum security .

    Contact us via email:

    Or Contact phone number: +12109267927

  4. Correction:
    Latest offers received in Saudi Arabia for October-2018 300 MW PV bid: Low Bid… $0.0178/Kw-hr – Runner-Up… $0.0234/Kw-hr. This is the real cost of PV today. The models above are way outdated.

  5. Latest offers received in Saudi Arabia for October-2018 300 MW PV bid: Low Bid… $0.0178/Kw-hr – Runner-Up… $2.3413/Kw-hr. This is the real cost of PV today. The models above are way outdated.

  6. They just need to understand expotenial growth. Of course one of the biggest problem with exponential growth is that a slight change in the growth rate can give big difference in the result.

  7. I think what would be much more interesting would be forecasts about battery technology, its impact and pricing. I realise that “Greentards” all go to bed when the sun goes down, but I actually do things at night.

  8. Caveats:

    1. As solar is adoption passes thousands of GW of installed capacity the cost of solar power moves from the bottom yellow line to the top green line. The reason is that all the best (aka “sunniest”) places have solar placed on them and solar moves to locations that are less than ideal.

    2. This chart tells me that if the cost of hundreds of GWhr of batteries was zero that solar would never displace methane in the US. The reason is that you would need 5x the solar installation capacity to produce the power you need for the entire day (due to the common 0.20 capacity factor of PV solar) so 100% solar would wind up producing power in 2035 at 0.125 $.

    3. The assumption in the chart is that the cost of natural gas power is constant from now until about 2035. Natural gas prices have been falling over the last decade.

    4. Obligatory comment about the sucky comment system.

    • What Combinatorics said makes no sense
      It has already been discussed ad nauseam that the need for backup storage has been vastly overestimated but he, along with all the pro-nuclear condrum, does not get it.
      He really does not get it
      As a side note, why is Brian so against Disqus and keeps using such s****y commenting systems

    • Solar and wind power gets dispatched first because their incremental cost of the next MW is zero. They can always undersell fossil fuel power. Remember most of their cost is interest so if they go bankrupt somebody else will buy them and sell the power for even less money.

    • “100% solar would wind up producing power in 2035 at 0.125 $.” The price already accounts for that. If batteries cost zero, 100% solar would cost ~0.025$. The chart is already in price per KWh, not capacity.

    • I back calculate about 500 megajoules for a 100 watt solar panel per year, which is like 3.6 gallons of kerosene. I agree that they will pay for their energy that in a matter of years.

      • A little pessimistic. I get to about 750 MJ/a in good locations. But maybe overall, not a bad estimate (your 500).

        BTW, 500 MJ = 500 miles of mid-sized vehicular travel too. 500 miles from a 100 W panel is 5 miles/watt-year. Cool. I drive about 20,000 mi/year. A little hard on my car, but California is Big and I go on frequent longer weekend and vacation trips. Love my state, hate its taxes, love its diversity, hate its socialism. Oh well. Things of interest can be far away.

        20,000 mi = 20,000 MJ = 5,500 kWh/year. ÷ 365 = 15 kWh/day.

        My electric bill is about $300/mo at 18¢/kWh. 1,660 kWh/mo ÷ 30 = 55 kWh/day.

        So I need 70 kWh or so per day. MORE to sell back to the electric company at some stupid discounted rate, then buy back in the cold dark months of winter, I guess. Call it 80 kWh with sell-back margin.

        Each kW of panel generates about 6 kWh/day here, on the average. 80 kWh ÷ 6 kWh/kW-day = 13 kW of generating panels. At 0.18 kW/m² = 75 m² of panels installed. 75 m² ÷ 0.3048² = 800 ft². For us Imperial estimators.

        Do I have 800 ft² of rooftop unobstructed by trees ‘n’ things? Hmmm… have to go see. I think so. But I have a peaked roof with only one side more-or-less facing the right direction. I guess I’ll have to do an “overhead wing” design, that rises up above the lee side.

        The upside is that rain won’t collect.
        The downside is that wind storms will require STOUT trussing.
        The downside is also that my roof will necessarily need puncturing.
        And thus leaking potential.

        The other downside is that it is expensive. At $0.85//W for panels, and about $0.50 for mounting bullshot, and about $2,500 to get an installer to really do the job right, and another $1,500 in inverters and electric-connections, I’m looking at $21,000 in acquisition and installation costs.

        I wonder what the other costs are.

        This is displacing about $400/mo = $5,000 a year in direct electricity costs. They change to about $500/year in electric-company share-the-grid costs. So about $4,500 yield off the whole $21,000 thing. Payback? Oh… about 5 years, right?

        Providing something doesn’t break or need expensive repair.


  9. Do solar panels put out as many joules in electricity over 20 years as it took to make them from rocks? At which point do they become “positive” in this respect with their typical capacity factor < 15%? Solar panels are not green until they pay back the fossil energy at it took to make them.

    The linked article is pretty interesting; it discusses how quickly solar power is taking off and how predictions never get it right.

    For the record I don't think it's fair to consider solar panels to be energy producers (green or otherwise) until they pay back the energy debt from their manufacture.

    • @scaryjello re: “solar panels generate more joules than invested to make ’em”

      Yes. At 18% single-crystal present-day efficiency average, 1 m² of panel generates close to 200 W at high noon in an optimal location at 35° latitude. Averaged over a day, the same panel produces something a bit over 1.2 kWh per day. The investment cost of making the solar cell panel is today (barring influence from national subsidies) very closely related to the energy-to-make as well as the whole-plant-to-manufacture the things. We presently can buy pallet-loads of fully assembled 200 watt solar panels for about $0.70 a watt. $140 a panel. If present competitive price for electricity is 10¢/kWh, then $140 ÷ 0.10 = 1,400 kWh of energy output needed to “pay for” the panel. If the panel averages 1.2 kWh/day, then 1,400 ÷ 1.2 = 1,200 days. Divide by 365 and you get 3¼ years. Payback in full.

      After that you have to figure out how to pay for the mounting systems and maintenance, the cost-of-land and the regulation/injection of produced energy into the grid … in order to determine the payback point. 10 years? That’s about right.

      However, as INVESTMENTS go, this isn’t a good payback term. There is a lot of things that can go wrong with solar panels, not least of which on St. John Island … hurricanes. Which basically destroyed an entire 5 MW hillside facility.

      Just saying. Doesn’t have the payback-to-cost ratios of hydroelectric or just plopping a nice new methane-gas plant on an old decommissioned coal power plant pad. (which already has all the power lines and transformers, breakers and so on in place!!!)


      • Actually, good ol` Goat is little bit too optimistic on this.
        At first, cost of installed solar at utility scale is still $3/W on average
        $2/W in best cases at least until now
        First Solar CEO Says that cost will eventually go down to $1/W
        This means that at the moment, the cost of a 6KW installation will be around $12K, possibly going down to, say $8K sooner or later
        Assuming conservatively 1200 hours of sun full power per year, this means 7200KWh saved per year, at a $0.12c/W
        Which means around $1000 saved per year at best.
        Probably more around $800
        This means that even if we will reach the price of $1 pe r installed W in the future, it will still need 8 years to break even. Now it is more like 12-14 years
        However, I would still recommend to go solar because of
        1) the costs of coal for the environment
        2) the costs of coal for health
        3) when cars will go electric, the Saudis will not get any money
        4) this will bring more democracy there

        Nuff said

      • Goat, you do know there is a lot of bad thinks that can happen to other types of power plants especially nuclear. Ask the Japanese. Also power plants require a lot of maintenance and monitoring.

      • As a side note, all such discussions are irrelevant since most solar is now utility scale where prices are around 5c/W

    • For the record I don’t think it’s fair to consider solar panels to be energy producers (green or otherwise) until they pay back the energy debt from their manufacture.

      Yup. Not to mention the toxins involved in manufacturing as well as EOLing them.

      You never hear the Solartards mention that. And when you challenge them to do so, they respond with weird non-contextual ‘logic’ or scream ‘FOX News!’ when you corner them on that.

      Solar cell costs and efficiencies are improving exponentially. This is true. And it seems that less toxic materials will be involved in future cells as well. But the rest of the panel and installation costs are pretty fixed and in some cases increasung in pace with the overall productivity fall in US construction that has gone on for two decades now thanks to zoning, regulatory protected construction practices and lack of adoption in automation.

      At these the log-jams concerning the last and part of the second are finally starting to fall.

      • > You never hear the Solartards mention well as EOLing them.

        That’s because silicon solar panels don’t contain toxic materials. They contain silicon (duh), aluminum (the frame), glass (front cover), plastic (back cover), copper (interconnect wiring), and a small amount of silver (cell contacts). All of these materials are recyclable, although so few solar panels are old enough to need it, the recycling business is basically non-existent so far. Panels are typically mounted on steel support structures, which isn’t toxic either.

        > But the rest of the panel and installation costs are pretty fixed

        Also not true. Both panel and other systems costs have fallen:

        Since you resort to calling the people around the world who are installing and favor solar names, and are completely wrong on several points, you must be the “Fake News!” that Trump talks so much about.

        • What say you of the materials such as silicon tetrachloride, hexafluoroethane, nitrogen trifluoride, polyvinyl fluoride used in the manufacture of solar products?

        • I believe some thin film PV – generally considered the future of PV – does have toxic material in it. Cadmium and Indium for example.

          Probably not enough to worry about from having them on your roof, but enough to worry aboutproper handling in the production/recycle stages. Maybe enough for firemen to worry about if your solar PV powered house burns down??

        • danielravennest

          The toxicity of materials (cadmium & tellurium anyone?) used in borh manufacture and content has been well documented. As has studies showing how the rest of the panel parts as well as installation costs are not falling as much or even at all (particularly labor). I did mention that some of those costs can come down and why that would be so.

          NBF has even posted articles about these issues.

          I note that you do not present anything that disproves that in your claims otherwise.

          So yes…you gave the perfect example of how I described the typical non-response Solartards give. Just subsitute ‘Trump’ for ‘Fox News’. What does ether have to do with this subject matter?

          What shocks me is that I expected better from you. Not necessarily agreement, but not the intellectually dishonest ravings of a raging Solartard.

        • Also for danielravennest:

          What say you of the materials such as silicon tetrachloride, hexafluoroethane, nitrogen trifluoride, polyvinyl fluoride used in the manufacture of solar products?

Comments are closed.