A Potential Solution to the Problem of Storing Solar Energy – Don’t Store It.

In How Much Battery Storage Does a Solar PV System Need? I assumed that the rooftop PV system would generate just enough power to fill annual domestic demand and that the surplus power generated in summer would be stored for re-use in the winter in Tesla batteries. The result was an across-the board generation cost of around $35/kWh. Clearly the Tesla battery storage option isn’t economically viable, or at least not under the scenario I chose.

As Phil Chapman and others pointed out in comments, however, this is not the only way a domestic solar PV system can generate enough year-round power to allow a household to go off-grid. Another is to overdesign the system so that it’s large enough to fill demand in winter when solar output is at a minimum and simply curtail the excess power generated in summer. How does this “no storage” option pan out?
To evaluate the no-storage case I use the same four scenarios as for the battery storage case (Equator, 20, 40 and 60 degrees north latitude) and add another (50 degrees north latitude) to provide more detail. The following assumptions are the same as before:

Household consumption is 5,000 kWh/year.
Household demand is constant through the year at 13.7 kWh/day, or 0.57 kWh/hour.

One assumption is clarified:

The impacts of short-term changes in cloud cover are ignored.

What this means is that I haven’t allowed for an unusually cloudy or unusually sunny week or month. Latitudinal variations in cloudiness are, however, allowed for in the load factors used to estimate generation (15% at the Equator, 17% at 20 degrees N, 16% at 40 degrees N, 12% at 50 degrees N and 10% at 60 degrees N) which are derived from metered panel output data.

One assumption is changed. PV panels are now pointed in the direction that provides maximum winter generation, not maximum year-round generation. To do this panels at 60N must be angled at 83 degrees south relative to the horizontal instead of 49 degrees, panels at 40N at 63 degrees instead of 35 degrees, panels at 20N at 43 degrees instead of 18 degrees and panels on the Equator at 23 degrees instead of zero degrees. The impact is to increase winter solar generation (by 21% at latitude 60N) and decrease annual solar generation (by 14% at 60N) everywhere except on the Equator, where the quirks of orbital geometry generate the opposite effect.

And while I refer to it as the no-storage case there will inevitably be times when the rooftop solar system won’t generate enough energy to meet household demand, meaning that some backup storage will be needed. To supply it I add one 10kWh Tesla battery storage unit, which would be capable of filling demand for at least one powerless winter night, and just in case I also add a 3kW backup gasoline generator. (Fossil fuel backup for a “100% renewables” installation is acceptable. The island of Eigg in Scotland backs up its hydro, wind and solar with diesel generators and the island of El Hierro in the Canaries backs up its wind/pumped hydro system with a 10MW oil-fired plant, and nobody complains.)

Now to the results. We will again discuss the scenarios in sequence. The solar module output  data used to construct them are from PVeducation.

Rooftop solar system on the Equator:

Figure 1 shows annual generation from a rooftop solar system on the Equator that is large enough to meet demand at the time of minimum solar generation, which with the panels angled 23 degrees south (or north) relative to the horizontal occurs at mid-year. System specifics are:

Installed capacity: 5.4 kW
Load factor based on consumption: 10.6%
Annual generation: 6,518 kWh
Annual consumption: 5,000 kWh
Curtailed: 1,518 kWh (23%)

Figure 1:  Demand, consumption and curtailment, rooftop solar system on the Equator with panels at 23 degrees to the horizontal 

As noted earlier, however, this option is actually less efficient than adjusting system capacity to meet minimum demand with the panels pointing vertically upwards (Figure 2). Specifics of this case are:

Installed capacity: 4.0 kW
Load factor based on consumption: 14.1%
Annual generation: 5,263 kWh
Annual consumption: 5,000 kWh
Curtailed: 263 kWh (5%)

This option meets demand with a smaller PV system (4.0 versus 5.4kW) and cuts the amount of surplus generation that has to be curtailed from 23% to 5%. This is therefore the option we will go with.

Figure 2:  Demand, consumption and curtailment, rooftop solar system on the Equator with panels at zero degrees to the horizontal

Rooftop solar system at latitude 20 north:

Figure 3 shows the data for this case. The 43 degrees south panel inclination generates a double-peaked generation curve with minima at the end of the year and at mid-year. System capacity increases by only 0.1kW relative to the Equator and curtailment is still low at 10%. System specifics are:

Installed capacity: 4.1 kW
Load factor based on consumption: 13.9%
Annual generation: 5,556 kWh
Annual consumption: 5,000 kWh
Curtailed: 556 kWh (10%)

Figure 3:  Demand, consumption and curtailment, rooftop solar system at latitude 20 degrees north

Rooftop solar system at latitude 40 north:

Figure 4 shows the data for this case. At this latitude system capacity has to be increased to 6 kW to meet winter demand and curtailment becomes significant at 33% of total annual generation. System specifics are:

Installed capacity: 6.0 kW
Load factor based on consumption: 9.6%
Annual generation: 7,429 kWh
Annual consumption: 5,000 kWh
Curtailed: 2,429 kWh (33%)

Figure 4:  Demand, consumption and curtailment, rooftop solar system at latitude 40 degrees north

Rooftop solar system at latitude 50 north:

Figure 5 shows the data for this case. At this latitude system capacity more than triples relative to the Equator, more than half of the power generated has to be curtailed and the load factor decreases to less than 5%. System specifics are:

Installed capacity: 12.6 kW
Load factor based on consumption: 4.5%
Annual generation: 11,516 kWh
Annual consumption: 5,000 kWh
Curtailed: 6,516 kWh (57%)

Figure 5:  Demand, consumption and curtailment, rooftop solar system at latitude 50 degrees north

Rooftop solar system at latitude 60 north:

Figure 6 shows the data for this case. At latitude 60N the winter sun is so weak that 87kW of installed capacity is needed to meet demand, the load factor falls to less than 1% and over 90% of annual generation is curtailed. The fact that 87kW of PV panels occupies an area of over 400 square meters and won’t fit on most rooftops doesn’t help either. System specifics are:

Installed capacity: 87.0 kW
Load factor based on consumption: 0.7%
Annual generation: 65,570 kWh
Annual consumption: 5,000 kWh
Curtailed: 60,570 kWh (92%)

Figure 6:  Demand, consumption and curtailment, rooftop solar system at latitude 60 degrees north


I estimated capital costs and cash production costs ($US) for the above cases based on the following assumed installation costs, which although subject to uncertainty should at least be in the ball park:

PV panels: $4,000/kW installed.
10kW Tesla battery storage unit: $5,000 installed
3kW Honda generator: $3,000 installed

I estimated generation costs simply by dividing the capital cost by 100,000 kWh, which is the usable power the system will generate assuming a 20-year life. This is of course a very crude way of doing it but it’s the way homeowners with rooftop systems usually look at economics. (My neighbor, whose $9,000 system cuts his electricity bill by $1,000/year, claims an 11% return on investment.) The results are summarized in the Table below:

And Figure 7 plots cash costs against latitude:

Figure 7:  Cash generation cost versus latitude, rooftop solar systems

I draw the following conclusions from these results:

1. The no-storage rooftop solar option is vastly more economic than the battery storage option discussed in the previous post.

2. Households in some parts of the tropics can already install rooftop solar systems that will allow them to go off-grid without suffering an economic penalty. The $0.24/kWh cash costs for rooftop solar in tropical latitudes are comparable to what residential users now pay for grid electricity in a number of countries.

3. Latitude places limits on where a rooftop solar system will allow a household to go off grid. The farther north (or south) of the Equator we go the more inefficient the system becomes, until at latitudes of much over 40 degrees the economics become marginal at best and at latitudes much over 50 degrees they become prohibitive (sorry, Scotland).

4. Latitude will constrain the growth of off-grid residential rooftop solar because most of the residential rooftop market is at higher latitudes in the Northern Hemisphere.

But going off-grid with rooftop solar might still be economically-viable if you happen to live in the tropics, right?

Well, not exactly. There’s a hitch.

My solar system:

I do happen to live in the tropics – at latitude 20N in Mexico to be exact – and two years ago I installed 2.25kW of PV panels on my roof at a cost of about $7,000. Since then the panels have operated at a load factor of slightly better than 20%, generating ~4,000 kWh/year. They have cut my consumption of grid electricity from about 4,800kWh/year to about 800kWh/year and reduced my monthly electricity bill from over $100/month to about $5/month (electricity in Mexico is heavily subsidized for low-consumption users). So on a cash basis my panels will pay for themselves in six or seven years.

And by installing a battery and buying a backup generator I could go completely off-grid. Why don’t I do it?

Because it would cost me thousands of dollars more and save me less than $100 a year. It’s far cheaper for me to buy a few kilowatts from the Comisión Federal de Electricidad when I need them than to install storage of my own.

And there’s the hitch. A rooftop solar system may make overall sense if the cost of grid electricity is high enough, but batteries and backup generators still aren’t remotely competitive with grid electricity when it comes to load-following.

This entry was posted in Energy and tagged , , , . Bookmark the permalink.

69 Responses to A Potential Solution to the Problem of Storing Solar Energy – Don’t Store It.

  1. Willem Post says:


    This sounds pretty good with only 2.25 kW of panels. Soon there may be a mass migration of retirees to these latitudes. Real estate and land values will increase

  2. Modelling a system comprised of wind turbine backed-up with pumped-hydro storage I found the most cost effective solution with a wind turbine maximum capacity of 5.5 times peak demand and a pumped-storage hydro energy capacity of 1.11 peak-demand days.

    See this post on my blog – “Modelling of wind and pumped-storage power”
    also this image

    • Willem Post says:


      It would be better to have separate time-line graphs.

      That way you can place a vertical line and see how the various actors are performing at that time, such as the rates at which:

      – wind energy, MWh, is fed to pumps
      – storage, MWh, is being added
      – demand is draining energy, MWh, from reservoir
      – energy, MWh, is being stored

      Using historical the daily energy demand and the daily wind energy generation, one should be able to determine bottlenecks, i.e., periods of little or no wind energy.

      • Willem Post says:


        The variable energy of the wind turbines would first need to be balanced by other generators before it can be used by the pumps to pump water into the reservoirs.

    • Euan Mearns says:

      The trouble with this is that it will produce a vast surplus of power that has nowhere to go. And while it may be”the most cost effective” solution for a wind based system, I’d be interested to know how the costs compared with other solutions. And its very difficult to make a market based economic case for the pumped hydro that will only get used occasionally.

      I presume you are aware that there is public concern in some quarters about the impact on landscape and also about power bills.

      And the problem remains of long lulls where in the case of Scotland we need 100% back up.

    • I would be interested to know how much pumped hydro storage the rest of the UK would need to handle the massive wind spikes that Scotland proposes to export to it.

      • Leo Smth says:

        There is only one place really that pumped storage in the UK could be increased and that is Scotland: It would indeed be in a semi-autonomous Scotland’s interest to do it too, as it would enable electricity to be sold into the highest priced market.

        Which is why it will never happen under the SNP ;=)

      • Jamie says:

        Not much, you just need EVs that the grid operator can dispatch when needed. For every 1 million EVs you have plugged in overnight you have about 3GW of demand that you can draw on. Obviously we need some smarts built into that but that will come as wind capacity, the EV market and the need for balancing services grow. More important in the short term is to beef up the North-South capacity which is more than doubling next year with the opening of the Western Link.

        • roberto says:

          “Not much, you just need EVs that the grid operator can dispatch when needed. For every 1 million EVs you have plugged in overnight you have about 3 GW of demand that you can draw on. ”

          Not even close!… do the math, man!… 🙂

          3 GWp (that little “p”) makes all the difference!… in Scotland from October to March would produce how much????

          Went here…


          … and clicked on the map somewhere a bit east of Edinburgh… optimized slope and azimuth… this is what 1 kWp would produce every year (copy and paste):

          Jan 1.00 31.0 1.20 37.3
          Feb 1.68 46.9 2.05 57.5
          Mar 2.69 83.5 3.37 104
          Apr 3.40 102 4.38 131
          May 3.87 120 5.07 157
          Jun 3.38 101 4.53 136
          Jul 3.39 105 4.55 141
          Aug 2.99 92.6 3.96 123
          Sep 2.62 78.6 3.40 102
          Oct 1.83 56.7 2.29 71.0
          Nov 1.39 41.7 1.69 50.7
          Dec 0.83 25.6 0.99 30.6
          Year 2.42 73.7 3.13 95.1

          First column after month name is average daily electricity production in kWh, second is average monthly, thirdand forth are average daily and monthly irradiation (in kWh/m2).

          Last row is the sum: from Nov to February, included, 4 loooong months… this ridicolous technology would produce 145 kWh… this means that 3 GWp would produce 3million times more, which is only 0.44 TWh… i.e. as close to nothing as one could go.
          How much electricity does Scotlan consumes on average during those 4 months?


          • Jonathan Maddox says:

            In Scotland, wind power is the intermittent renewable of choice and the posts you are replying to were clearly discussing wind — which, usefully, peaks in winter when energy demand is greatest. PV is something a niche technology so far north (but not useless, because it’s less windy in the sunnier months).

  3. matthew_ says:

    Nice Roger!
    Figure 7 is pretty clear. Beyond the 40th parallel off-gridders need to think of something other than PV and batteries to supply winter power.

    • roberto says:

      … and let’s notice that Roger has been magnanimous with PV… he has assumed a 16% load factor at 40 deg and 12% at 50 deg latitude… but Germany, overall, barely gets to 10% as an average… even sunny Italy got only 14.2% for all of 2014…


      • Roger Andrews says:

        Roberto: I didn’t assume the load factors. I estimated them from actual generation data from hundreds of individual PV systems. I did it this way because load factors estimated from national generation statistics aren’t reliable. It’s all discussed here:


        • Roberto says:

          Thanks for the link, but I disagree… In my opinion it is the opposite, it is sparse data from few hundred pv systems which is not indicative of the real numbers… there are consistent data for entire countries were are paid and installed power and production are carefully monitored… Because it’s big bucks… Italy and Germany are the v two prime examples of this.
          Germany rarely exceeds 11%, Italy 15%… the latter in spite of having a large chunk of its total power installed in the sunny southern part of the country.
          I can provide the data… in Italian…

  4. A C Osborn says:

    The question not addressed here is, will the Panels actually carry on producing at the same rate for 20 years and will the Inverters also last 20 years?
    I have my doubts, especially about the inverters.

    Does Solar Energy impact household Insurance?

  5. Dave Ward says:

    “It’s far cheaper for me to buy a few kilowatts from the Comisión Federal de Electricidad when I need them than to install storage of my own”

    And if enough people do the same, as Tesla advocates, you’ll eventually find the grid suppliers will go bust, as they can no longer operate such a wildly variable system. Then you really WILL depend on that small generator….

    • The Comisión Federal de Electricidad is government-owned and isn’t going to go bust. Demand for rooftop solar is limited here anyway. And if the ultimate goal in other countries is to bankrupt the grid suppliers then all I can say is – see you in Mexico.

      • JerryC says:

        I think the point is that you are essentially free-riding on the conventional grid, which limits the scalability of your approach. We can’t all free-ride off each other.

        • I get sick and tired of people who know absolutely nothing about Mexico telling me that I’m free-riding on the grid, or “exploiting the poor” as I’ve been accused of in the past. My rooftop PV system is completely unsubsidized, has no impact on the rates charged to poor Mexican consumers and has significantly cut my carbon footprint (how much grid electricity do you consume in a year?) Do you really think the world would be better off if I got rid of my solar panels and went back to 100% grid electricity, much of which in Mexico is still generated by burning oil?

          • JerryC says:

            No need to get all defensive. I’m simply pointing out that your system is not somehing that would scale to general usage because widespread adoption would completely wreck the economics of grid production.

          • The big free-riders are the owners of utility-scale wind and solar plants. They dump their subsidized power on the grid whenever the sun shines or the wind blows whether the grid needs it or not and leave it up to conventional generation to sort out the mess while paying absolutely nothing for the privilege. They are the people who are wrecking the economics of grid production, not the residential users.

  6. Euan Mearns says:

    Roger, I like the post because it puts numbers on the folly of deploying solar N of 50˚. Aberdeen is 57˚N. You construct a system that has over-supply in summer when it is not needed and effectively zero supply in winter when it is. Its interesting to see the exponential change N of 50.

    The pic is from an earlier post where I was complaining about solar panels in Aberdeen being deployed on roofs pointing all ways. I think the argument goes that there is so little direct Sun that it doesn’t really matter if you leave the things in their box.


    • Euan: You might care to add a “bonkers” tag to the post 🙂

      • Willem Post says:


        Because inverters have lower efficiencies at PV solar outputs of less than 20% of inverter capacity (occurring mostly during winter, and dawn and dusk throughout the year), the monthly energy feed-in ratio is about 4/1 in New England. In Southern Germany, further away from the equator, it is about 6/1. See monthly output from 2 monitored solar systems in Munich. In Vermont, the hours of sunshine ratio is about 2.54 and production ratio is about 3.8 for fixed-axis systems. Here are two field-mounted examples, one fixed-axis, one 2-axis tracking.


        The Ferrisburgh Vermont solar farm, 1,000 kW, south-facing, correctly angled, field-mounted, has monthly averages of 4 years of production that show the monthly energy feed-in ratio of July/December = 1.000/0.263 = 3.80, and a 4-yr average CF = 1,323,879 kWh/yr/(8,760 hr/yr x 1,000 kW) = 0.151.

        The South Burlington Vermont solar farm, 2,200 kW, 2-axis tracking units, field-mounted, has monthly averages of 4 years of production that show the monthly energy feed-in ratio of July/December = 4.936, worse than fixed-angle, and a 4-year average CF = 0.167, which is .167/.151 = 10.6% better than fixed-angle, even though such trackers are claimed to be up to 45% better! In Vermont, the better performance of 2-axis, up to 21%, occurs mostly during May, June, July and August. Snow would readily slide off the panels at the steep winter angles. Such systems would be about 25% to 30% more costly and require greater O&M expenses, which will reduce any economic advantage.

  7. Ed says:

    Horses for courses. If you live in Scotland you are better off clubbing together as a community and building a wind turbine to serve that community because one large turbine is more efficient (I think I am correct in saying) than loads of small individual ones. On the other hand if you live in Spain you can add individual roof top solar into the mix. The important thing for me is the sense of empowerment, not waiting for the government to get their act together, or letting big corporations have control. This cuts through the faux divisions of being Conservative or Green or somewhere in between. Am I being unrealistic or romantic ? Maybe.

    Two of my concerns would be firstly what happens in cities and secondly, how long term is wind or solar; can they produce enough surplus energy so they can renew themselves without fossil energy?

  8. Willem Post says:


    The ratio of solar energy production of the most productive month divided by the least productive month is about 3.8 in Vermont.

    However the solar irradiation ratio in Burlington, VT, is 2.54, per NREL.

    The culprits are dec, jan, and feb, which underperform by 30, 52 and 34%, respectively.

    It may have to do with snow/ice cover, and the inverter being very inefficient at low percentages of rated output.

    An added low-capacity inverter, at about $2,500 installed, may be needed for ALL hours with low solar DC production.

      • Willem Post says:


        The cloudiness part may already be covered in the irradiation values.

        I called them irradiation values, but the website stated “average hours of sunshine” for the 12 months.

        At your 20 deg latitude, low DC solar outputs during winter are much less of an issue. The inverter operates mostly above 20% – 25% of its rated output.

        The inverter efficiency rapidly decreases to 0 below those percentages.

          • Willem Post says:


            I went to the site. Put in 44 deg North for Burlington, VT and a tilt angle of 54 deg for maximum winter generation, at a sacrifice of some energy in the summer. This will not maximize energy production, but flatten it.

            I assume the 54 is the sharp angle between the panel and the horizontal.

            These numbers do not take into account climate effects and any system factors relating to actual tilt, actual facing south, being not dusty, being not shaded, being not aged. Those real-life factors reduce Germany’s annual energy production by about 15%

            Germany’s installed solar capacity looses about 1/2 percent/y of production due to aging; 38,000 MW installed less 190 MW a year later!!

          • Willem: These graphs from my “Estimating Solar Load Factors” post take climate effects, tilt etc. into account:

          • Willem Post says:


            Interesting graphs.

            I notice Germany with a CF of about 0.12 which is higher than the actual of about 0.103.

            Germany keeps good records of production and installed capacity.

            Some of the outliers points on the high side may be 1 or 2 axis tracking systems.

            I am always amazed how you manage to come up with the data.

            What you call load factor I call capacity factor.

            In general, a load is imposed on a system which responds by operating at a capacity factor.

            In case of solar, almost all systems respond to the sun.

            That is what makes their energy so fickle, that it requires coddling and baby sitting.

        • Willem Post says:


          PV systems are set up to maximize production for maximum subsidies.

          To get maximum production from the sun in winter, subsidies would need to be adjusted on a monthly basis for people to bother to adjust the angles.

          Also, with the grid available as a battery at a constant cost there is less need to bother.

          There is perversity all around that messes up idealized approaches regarding storage.

          Other factors are: not facing true south, not having properly angled roofs, not being clean, being partially shaded by trees or clouds, being covered with snow and ice.

    • Willem Post says:


      Here are two German monitored PV solar systems showing a monthly production ratio of about 6.0, which exceeds the solar “hours of sunshine” ratio, as is the case in Vermont.

      See monthly output from 2 monitored solar systems in Munich.

      An analysis regarding storage should be based on actual production ratios.

      As I mentioned, at 20 deg, your ratio may be much closer to 1.5. Do you have monthly data to determine it?

      • Willem:

        a monthly production ratio of about 6.0, which exceeds the solar “hours of sunshine” ratio.


        • Willem Post says:


          The corresponding Vermont ratios are 2.54 for sun and about 3.8 for production.

          The 3.8 ratio is an average of two large, field-mounted systems over 3 years.

          The 3.8 ratio is about the same for many Vermont projects.

          Monthly production is listed for each project on the Department of Public Service website.

          The German ratios are…..for sun and about 6.0 for production.

          I do not know the German ratio for sun.

          • Willem: The ratios of what to what?

          • Willem Post says:

            The ratio of the most sunshine month divided by the least sunshine month and the ratio of the energy production of these months.

            For Vermont the ratios are 2.54 and 3.8, for Germany they are ? and 6.

          • Willem Post says:

            The production ratios being greater than the solar ratios means the winter production is really poor for various reasons, including the inverter inefficiently operating at the low end of its rated output for many more hours than in summer.

          • Another possibility is misaligned panels, which work much more efficiently in the summer than in the winter. Here are some numbers for horizontal panels:

            New Hampshire (43N)

            Summer Maximum:
            Incident power 11.86 kWh/m2/day
            Module power: 11.18 kWh/m2/day
            Efficiency 94%

            Winter Minimum:
            Incident power 4.38 kWh/m2/day
            Module power 1.74 kWh/m2/day
            Efficiency 40%

            Germany (50N)

            Summer Maximum:
            Incident power 12.21 kWh/m2/day
            Module power 10.92 kWh/m2/day
            Efficiency 89%

            Winter Minimum:
            Incident power 2.81 kWh/m2/day
            Module power 0.80 kWh/m2/day
            Efficiency 28%

            Lots of misaligned panels around.

  9. So, double the costs for the average American at 10,000 kWh per year?. At 40 degrees = $64k + $20K interest on loan = $84K?

    It’s not uncommon in Seattle to not see the sun for weeks on end.

  10. Hugh Sharman says:

    Umm…thanks Roger!

    I completely agree that Elon Musk and his over-hyped “Powerwall” are likely to come to grief.

    I AM impressed that your two-years old roof mounted PV can achieve 20% capacity factor. There are many billions who live between 20° north and 20° south of the equator who can benefit from PV as the costs of PV installations continue to fall

    Your interesting article is based on a very specific example from which one cannot draw any general conclusions at all. Mexico, dometic application, sounds like a “best case” example for the argument you are making for PV without storage, especially the PowerWall!!

    As a different example of which I am well aware, the whole of mid to north Nigeria has only a very limited chance of ever seeing reliably delivered electricity within the lifetime of more than 100 million citizens of Nigeria today. Unless massive amounts of PV (many, many GW) are delivered and co-generated with its pitifully under-performing grid, existing and new hydro (including small scale hydro), electricity storage, diesels, landfill gas power and synchronous condensers, these populations are doomed to continued, wretched poverty.

    You notice electricity storage is in the list. Putting aside the PowerWall, system costs of storage are also falling. New chemistries for large scale storage, using zinc and lead (for example) will drive expensive lithium chemistries (requiring lithium, manganese and colbalt) out of large-scale storage applications. Storage applicatons in off-grid, distributed and grid systems, at power stations, within distributed grids and “behind the meter” are coming along quite nicely, admittedly at the scale of only hundreds of MW globally at present.

    My main concern about the future viability of storage is the likely, absolute scarcity of the necessary metals as exponential growth in global storage capacity takes hold during the next decade while the slow-down in population growth happens more slowly than the expectation of that increased population. Witness the mess with seaborn refugees in Africa and East Asia, today.

    On a related subject, are we on peak (affordable) oil? Are we heading towards peak affordable gas?

    • Ed says:

      Organic flow batteries that are in the early stages of development don’t need expensive metals. http://www.sciencedaily.com/releases/2014/01/140108154238.htm

    • Olav says:

      Finally got my “non battery” storage as commented in http://euanmearns.com/the-cost-of-energy-storage/ up and running. All of April and until now in May I have made dinner every day at 18;00 from the last day or yesterday(s) sun. Have not used the electric cook top at all since April 1. Weather has been unusually lousy and I live on Latitude 58 on West coast in Norway which I believe has less sunshine than Aberdeen. I northern Nigeria this will cover a family need for cooking heat easily but ecomically it will not fly there even at a tenth of Tesla batteries cost
      High temperature storage for cooking is a good replacement for electricity and it would help the grid at the most valuable time of the day. Charging the storage as the sun shines is also beneficial. Ref the “California duck”

      • Roberto says:

        ‘. I northern Nigeria this will cover a family need for cooking heat easily but ecomically it will not fly there even at a tenth of Tesla batteries cost’

        In third world countries, were a tiny fraction of the population gets part 5th grade, any pv system will stop right after its first technical glitch, there is no chance whatsoever to set up a reliable maintenance program unfree such circumstances.

        • Olav says:

          We should not underestimate the Africans possibility to adapt. In rural Africa is landline phone a technology they did not enter, they go directly to Mobile phone. I also believe they will skip centralized power. Even a few poles to reach a 1 KW a day user is far to expansive unless heavily subsidized. Small PV systems is available at a lower cost and is taken in use at a very high rate. Those small PV systems are far more complicated than the system I tested for cocking. I do not need MPPT Battery charger, batteries and Inverter which have potential for some technical glitches.

    • Willem Post says:


      Here are some calculations regarding using a TESLA unit: charge with grid energy when rates are low, and use the energy when rates are high.

      At 90% AC to DC inverter efficiency, and allocating half of the 8% DC-to-DC loss to the charging side (the TESLA unit has a round-trip DC-to-DC efficiency of 92%, per spec sheet), it would take 7/(0.9 x 0.96) = 8.10 AC kWh of off-peak grid energy to charge 7 DC kWh into the unit.

      During on-peak hours, one would get back 7 x 0.96 x 0.90 = 6.05 AC kWh to use in the house. A big percent loss of energy!!

      The INSTALLED cost of the 7 kWh TESLA unit = $3,000 + S & H + Contractor markup of about 10 percent + $2,000 for an AC to DC inverter + Installation by 2 electricians, say 16 hours @ $60/hr = $6,500.

      In Southern California, base rates are $0.11, off-peak, and $0.46, on-peak; which likely is THE best-case scenario in the US. But this rate ratio is only for 6 months.

      The 8.10 AC kWh, off-peak, would cost $0.89. The avoided cost of 6.05 AC kWh, on-peak, would be $2.78, for a profit of $1.89/day, or $691/yr.

      Adjusting for 1/2 year, the SIMPLE payback would be about 10 x 2 = 20 years, not counting the cost of financing, PLUS any costs for O&M, PLUS any capacity degradation due to deep cycling.

      The TESLA warrantee is for only 10 years (for manufacturing defects, not performance), and during these 10 years, there would be 3,650 deep discharge cycles, which far exceeds what such batteries are designed for, meaning there would be capacity degradation that would further lengthen the SIMPLE payback period.

      • Willem Post says:


        As more info of the TESLA units has become available during the past weeks, some updating is needed.

        According to a recent article in the WSJ (paper version), the TESLA 7 kWh unit is:

        – Designed to have multiple charge/discharge cycles PER DAY; the article does not say how deep.

        – Meant to be paired with a PV system

        – Has an installed cost, with other equipment, of about $7,000.

        According to the article, the TESLA 10 kWh unit:

        – Is designed to provide power up to 2 kW during a power outage.

        – Can be cycled up to 50 times per YEAR for buying energy during low cost periods, storing, and using the energy during high cost periods*.

        – Has an installed cost, with other equipment, of about $5,000 under lease, or $7,140, if bought.

        * This involves the following energy loss:

        Assuming a 50% charge/discharge, and a 90% AC to DC inverter efficiency, and allocating half of the 8% DC-to-DC loss to the charging side (the TESLA unit has a round-trip DC-to-DC efficiency of 92%, per spec sheet), it would take 0.5 x 10/(0.9 x 0.96) = 5.787 AC kWh of off-peak grid energy to charge 5 DC kWh into the unit.

        During on-peak hours, one would get back 0.5 x 10 x 0.96 x 0.90 = 4.32 AC kWh to use in the house, for a (1 – 4.32/5.787) x 100% = 25.3% loss of energy!!

        In Southern California, base rates are $0.11, off-peak, and $0.46, on-peak; which likely is THE best-case scenario in the US. But this rate ratio is only for 6 months.

        The off-peak cost would be 5.787 x 0.11. The on-peak avoided cost would be 4.32 x 0.46, for a profit of $1.35/day!!! The monetary gain of this arbitrage is miniscule.

        The cost of financing, PLUS any costs for O&M, PLUS any capacity degradation due to cycling are ignored.

        The 10 kWh unit providing up to 2 kW of “quiet, clean” energy for a few hours of outage is quite a stretch, based on need, and is making no economic sense, unless grossly subsidized; 2 kW for an upscale mansion is insufficient and multiple wall-hung units would be required.

        Even a PV solar system owner doing some energy “shifting” with the 7 kWh unit is quite a stretch, based on need, and makes no economic sense, unless grossly subsidized.

        Utilities could be using multiples of the 100 kWh units for stabilizing their distribution grids with large numbers of PV systems. The US DOE, getting on that bandwagon to look progressive, will be subsidizing those applications for “demonstration purposes”.

        The more we shift to expensive RE, the more we shift the US wholesale price of the mix of energy from the current 5 c/kWh to about 10 – 15 c/kWh.

        That trend of increasing wholesale prices would be more visible, if many of the RE changeover costs were actually charged to the US energy system.

        Instead, they are “socialized” by means of taxes, fees, surcharges, bond issues, grants, etc., by POLITICIANS, because they do not want to be blamed for raising the cost of electricity.

        A perfect example of such follies is the wood chip-fired, Montpelier District Heating Plant; taking from politically deluded Peter, to do favors for politically well-connected Paul.

        Those various cost, due to increasing RE in the US, will have a MAJOR impact on making ALL goods and services, not just energy, much more expensive, as is already happening in Germany, although many RE proponents and politicians blame it on other factors; somewhat like Miss Piggy: MOI?

        In fact, Germany, being THE economic engine of the EU, has experienced slowing economic growth, due to its growing ENERGIEWENDE, in the past few years.

        Weaker EU countries are significantly affected by the German economic slowdown.

        Germany and other EU countries losing part of the very profitable Russian market and throwing billions/year into a black hole, called Ukraine, is an additional headwind.



  11. Jamie says:

    Again I reiterate that if you were going completely off grid you wouldn’t just carry on consuming what you’ve been consuming before. Before specifying your PV + storage system you would obviously replace all of your lighting with LEDs, buy efficient appliances and generally hammer down your demand as far as possible. You might even adopt strategies such as operating a freezer in summer but not winter and you would definitely want to make sure you’re dumping your excess generation into hot water so the assumption that the excess would be curtailed doesn’t hold.

    The reason why I can be so confident that people would take these steps is that the economics of cutting electricity consumption far out strip the economics of supplying it so it shouldn’t be ignored and makes a big difference to the results.

    • The option of going completely off grid with a rooftop PV system is limited to those who have a) enough money to pay for the solar panels, batteries, inverters, energy-efficient appliances etc. b) enough roof space to accommodate the panels and c) a willingness to alter their lifestyles to cut their energy consumption. The vast majority of people even in the developed countries lack at least one of these essential ingredients.

      • Jamie says:

        Indeed although I’m not sure what your point is. My point was that if you make unreasonable assumptions you get unreasonable answers.

        While you have access to a robust grid connection, complete off grid living won’t make economic sense.

        Short term battery storage will make sense in the long term though because of the need to cut peak demands on the grid. Having millions of customers drop off the grid for the evening will be of great help to grid managers and being able to have their batteries soak up overnight wind generation will be another huge benefit.

  12. David MacKay says:

    This is a nice clear post, well done! One minor issue: you’ve assumed the one battery will last the full 20 years (delivering a daily day->night storage service). I think this may be unlikely – the US target for automotive batteries is that they can deliver about 1600 cycles. So perhaps in your scenario the owner would have to replace the battery a few times over 20 years. This will bump up the cost per kWh somewhat, by a roughly similar amount at all latitudes.

    • louis says:

      I agree, though also think that leasing the storage is a better way forward. It’s noticeable that most of the leased PV systems that I know of in the UK don’t come with any local storage option. Though the Tesla Powerwall ? solution is the first t o capture the imagination of the press I would hope in future that allowing a leased storage solution to pull the thousands of households with systems already in place would at least be a start … if nothing else there are the generating points to at least experimeriment with developing a turnkey solution to orfer the option/plan for the contingency.

      • Willem Post says:


        “Though the Tesla Powerwall solution is the first to capture the imagination of the press….”

        What catches the imagination of the hard-pressed press is a chance to write about hot topics, such as energy storage, to sell papers, etc.

        See my above comment, which indicates the economics of the TESLA unit is completely dismal.

        The press probably could never come up with this type of calculation, because it would kill their story.

        To tie this cripple into EVs makes matters even worse.

        • louis says:

          I doubt that the press were ever actually remotely interested in the Tesla battery viable or not. However Elon Musk et al. have never had a problem generating column space …. but Turnkey solutions are necessary if a concept is to be at least explored within the public’s energy debate even if only to be discounted.
          Though I’m sure the press will be all over it once the first house burns down as a result of a PowerWall failure …

  13. Jonathan Maddox says:

    “maximum winter generation … panels on the Equator at 23 degrees instead of zero degrees.”

    In which months does winter occur at the equator?

    • Roberto says:

      Well, never thought about that… I’d say that since the sun’s orbit oscillates +/- 23.5 degrees north and south at the equator, there are 2 days of minimal daylight, corresponding to the two solstices, 3/22 and 12/22 (or is it 21st?)…


  14. jengel says:

    One aspect not included in these comparisons of solar performance at various latitudes is that the energy needs change dramatically at the various latitudes. For example, 5MWh/yr at the equator or 20N could provide a modern lifestyle with all the techno-gadgets and entertainments and even some left over for the occasional AC binge.

    However, as someone who lives at 45N, has 5kW of rooftop solar (self-installed in 2010 at a total cost of ~$17k) and tracks his family of four’s energy use can tell you, the 5MWh will disappear quickly in the winter. My 1800 square foot home is older, but updated with an eye towards efficiency and run fairly frugally energy-wise, and yet our annual energy usage is about 25MWh/yr from all sources with about 4MWh/yr for baseline modern lifestyle electrical energy use (lighting, appliances, gadgets, etc). Approximately 19MWh/yr of this is sourced via natural gas for heating living space. We have an air-source heat pump that does AC duty in summer and shoulder season heating which cut our annual gas usage from 800 therms to 500 therms but below 25F the heat pump gives way to gas. Even assuming we had a heat pump that could maintain a COP of 3 at all temps (COP of 3 for geothermal vs COP of 1 for gas) our annual energy usage would still be 13MWh/yr of which the lions share would be used for conditioning living space. Our location has about 8000 heating/cooling degree days per year for comparison.

    Assuming a structure that averages 3kWh of energy needed for all purposes per degree day as ours does, we need 5MWh in a single winter month with 1700 heating degree days. Move North to 60N and it only gets worse. My guess is that 3kWh per degree day is low for the average US consumer. We’d need 850kWh in a 5 day span in winter. Which using your math for 40N would require ~73kW of installed rooftop solar. Not sure I could fit that on my roof.

    My point is that the solar energy footprint needed for a reasonably comfortable life isn’t just a function of insolation, but rather energy needed. While you could live comfortably at 0N or 20N with 5kW installed solar, you would live a primitive life by comparison at 40N or 60N. Which makes me wonder about the sustainability of living at the higher latitudes long term. How “passive” a home could you build in a temperate climate with 8000 degree days per year?

    • energy needs change dramatically at the various latitudes

      It’s more a function of wealth than latitude. Here are some per-capita electricity consumption numbers from the World Bank. Note the contrast between Singapore (rich) and Indonesia (poor), both close to the Equator. Also the contrast between France (rich) and Ukraine (poor), also both at the same latitude.

      UK 5,472
      Germany 7,081
      France 7,292
      Ukraine 3,662
      US 13,246
      Japan 7,848
      Australia 10,740
      Singapore 8,404
      Indonesia 680


  15. jengel says:

    Gah! I think my lengthy, detailed, and insightful comment was eaten by the dog…

    Let me summarize: Our semi-efficient 1800sf house at 45N uses 25MWh/yr of which 19MWh/yr is for heating using gas in support of an air-source heat pump. Our 5kW of rooftop solar covers our baseline electrical energy usage (4MWh/yr) but not energy for space conditioning. Our climate is 8000 degree days per year and we average 3kWh per degree day. In the dead of winter we need 850kWh per 5 day period which by your math for 40N would require about 73kW of rooftop solar. Not going to happen. Even at a vastly more efficient 1kWh per degree day we’d need 24kW of rooftop. Can passive techniques achieve 1kWh per degree day?

    My point is that while 5MWh/yr is good for reasonable lifestyle in some locations, northern climates will not be livable at that energy level with current building stock.

    • Willem Post says:

      Per Passivhaus criteria, a 2000 sq ft house would consume for space heating about 15 kWh/m2/yr x 186 m2 = 2,790 kWh/yr, or 23.3 kWh/day, if averaged over 4 months, and its PEAK demand would be 10 W/m2 x 186 m2 = 1.86 kW, or 6,348 Btu/hr, or 3.2 Btu/sq ft/hr, i.e., a 2 kW electric heater in the air supply duct COULD be the heating system!!

      For comparison, here is the URL of a 1,232 sq ft house with a heating design demand of 10,500 Btu/hr, or 8.5 Btu/sq ft/hr. It uses for:

      – Heating two Mitsubishi, Mr. Slim, ductless, minisplit, heat pumps (one downstairs @ 12,000 Btu/hr, and one upstairs @ 9,000 Btu/hr), installed cost about $5,250.
      – Ventilation a Lifebreath 155 ECM energy-recovery ventilator.
      – Electricity a grid-connected, PV system, 5.7 kW, roof-mounted with Fronius IG 5100 inverter, installed cost about $22,000 less subsidies.

      • jengel says:

        The URL shows a nice solution for building a new house to be low/zero net energy. Awesome. For those of us “in place” getting 12″ thick framed exterior walls is not easily done, but one can dream.

        To meet passive house the HVAC budget is 2790 kWh/yr for 2000sqft, which is pretty close to 1/3kWh per degree day if you have 8000 degree days. Not too bad.

        I find the passive house primary energy numbers quite surprising at 120kWh/m2/yr, or 22.3MWh/yr at 2000sqft (186m2). That is a huge amount of energy, considering only 2.8MWh/yr (15kWh/m2/yr) are allowed for HVAC. It would take a lot of TV’s, hair dryers, halogen lights, and iPhones to use the remaining 20MWh/yr. Why work so hard on conserving HVAC energy and then leave all the lights on? My primary energy is only 25MWh/yr right now and if my HVAC met passive house, my annual primary energy would be 7MWh/yr, or 40kWh/m2/yr and 3x smaller than passive house standard. Am I missing something?

        Thanks for the discussion.

  16. Luís says:

    The assumption of constant consumption throughout the year, throughout each month and throughout each day has no adherence to reality. Neither does have the assumption of constant output throughout the day. This exercise is completely meaningless.

    There are various reasons to consider the Tesla battery a money loser, these out-of-touch calculations are totally unnecessary.

Comments are closed.