Last week’s post on The Energy return of Solar PV caused quite a stir. Yesterday I received a response to some of the comments from Ferroccio Ferroni and Robert Hopkirk addressing some of the queries raised by readers. Their response is given below the fold. But first I have a few comments to add.
Let’s kick off with the unshakeable enthusiasm for renewables of every flavour from the Scottish National Party. Member of Parliament Callum McCaig (inset pic):
I think Scotland is very much leading the way….
….in how we embrace the renewables agenda… whilst we have a government who seem to be investing in the white elephants of Hinkley and, at best, 1950s technology. It is not a lost cause but we do need to keep beating the drum for solar, for wind, for storage and we get these things right we can say cheerio to things like Hinkley Point C.
Solar in Scotland produces virtually no electricity in the winter months, it produces absolutely zero electricity at 18:00, the time of peak demand in winter (Figure 2). Bridging the annual solar cycle with storage is impossible (at least totally impractical). Hinkley C, despite its many flaws, would provide 3.2 GW of power 24/7 and it is cast as a white elephant! The political commentary on energy in Scotland has become dangerously deluded.
Figures 1 and 2 UK solar output according to National Grid for June 2015 (top) and December 2015 (bottom). Both charts at same Y-axis scale are taken from UK Grid Graphed. UK wintertime solar is totally useless as a power generating technology producing a pathetic midday blip. According to DECC the UK had 8.14 GW installed PV capacity at the mid-point of 2015 and 8.92 GW at the end of 2015. In December, those 9 GW of installed capacity pushed out 0 GW of electricity for most of the time.
But do we know how much electricity solar PV is generating in the UK? We can download data from National Grid or DECC but what does it mean? Most domestic PV systems do have meters that evidently are not read. I checked with the Renewable Energy Foundation who were of the opinion that UK PV generation was model based. In other words synthetic data based on model inputs. And this is DECC’s description of UK solar PV production statistics:
Note 6. Actual generation figures are given where available, but otherwise are estimated using a typical load factor or the design load factor, where known. All solar photovoltaic generation is estimated this way.
Despite what many may think I do not have an axe to grind on solar PV. If it can be shown to work unsubsidised with ERoEI >>5 then fine. But if it is true that high latitude PV has ERoEI approaching 1 then we are digging ourselves an energy grave. The book by Prieto and Hall on the ERoEI of Spanish solar PV, and this excellent pdf summary, settled on a figure of 2.4 for a capacity factor of 17%. Normalise that to a capacity factor of 9% for high latitudes like the UK and the ERoEI comes out at 2.48*(9/16) = 1.4 and Prieto and Hall do not account for the energy cost of intermittency. Take that into account and their number is not a million miles away from the controversial figure of 0.83 reported by Ferroni and Hopkirk. And please note that ERoEI for solar PV is not a constant but varies with latitude, sunshine hours and orientation of installation. It seems in the UK at least that we do not have reliable figures for the energy return part of the ERoEI equation.
And so to the response from Ferroni and Hopkirk:
To the blog: http://euanmearns.com/the-energy-return-of-solar-pv
We are clarifying here our position regarding the objections raised in this discussion-panel:
1) The value of the PV-electricity production in Switzerland during the lifetime is too low and PV-systems are producing per year much more electricity (as much as 153 and up to a value of 185 kWh el/m2 have been cited in other comments). We give a value of 107 kWh el/m2/year and with degradation, a lifetime value of 2203 kWh for 25 years. One should note that also PV modules integrated in building façades are included in this average.
- We base our value on the Official Swiss Statistics and we explain how to use it. The weak point of the all PV- statistics is that they are not based on the module surface but on the fictive value of kW peak. This was enforced by the solar industry in order to confuse the results. On the other hand, the value of the energy absorbed in thermal collectors is given per unit surface area. Going to the statistics one obtains the total energy produced up to the end of the corresponding year (normally in GWh) as well as the total installed capacity or power in MW peak as direct current at the end of the corresponding year. Dividing MWh by the average capacity between the end of year and end of the previous year we obtain MWh / MWp. We assume here that during the year a linear capacity increase due to the construction of new PV-plant is to be expected.
- Here we have to distinguish between the value indicated by a module supplier and what is used in the real planning of PV-Systems. Modules are sold on the basis of money per peak Watt, which is understood to come from a reference sunlight intensity of 1 kW/m2. If the conversion efficiency were perfect, the area of the module would be 1 m2. For a conversion efficiency of 20 % the required area would be 5 m2. But efficiency is measured at a standard temperature of 25° C and a vertical radiation incidence of a lamp, that cannot simulate 1 to 1 a solar radiation. We suggest measuring efficiency at an incidence of 45 degrees inclination and at a temperature of 40 °C and this would correspond more closely to reality, the resulting efficiency then being much lower. In real planning other values are used: In Germany a value of 10 m2 per kWp is used (information from insiders – i.e. from people working in the field) and in Switzerland, Swissolar is recommending also to plan PV-plants with the same value. This means that, when in Germany one will obtain 1000 MWh/MWp from the statistics and then, dividing by 10 m2/kWp, 100 kWh/m2 of electricity production. In our case we have made an average over the last 10 years, but we have used a conservative value of 8,2 m2 pro kWp to determine the electricity production. The value is based on the evaluation of projects realized in Switzerland where previous projects were planned with 8 to 9 m2 per kWp.
- I have been in front of a court to deal with the question of declaration of the payback period of PV-Systems. I had stated that the average specific performance value is 100 kWh/m2 without degradation. The utility did not claim that my assumption was wrong. Therefore, our value of 106 is court-proof.
- One utility has forbidden us to publish results and I suppose that they were below 100 kWh/m2.
- When I was asking some utilities to give me measured results for publication, I received only the project data, which usually indicates that these values are higher than any measurements. All computer programs for the calculation of the production give results, which are too optimistic.
- We have assumed an average lifetime of 25 years. We have however, insider information claiming that a lifetime of only 20 years is more realistic.
In conclusion, we are aware that some PV-plants produce much more than our average. But we are confident that at the present stage of development our average value is conservative and of course we are not appreciating the non-transparency of the utilities.
2) Energy involved in labour and capital: In our paper we have demonstrated that PV generation is both labour and capital-intensive – for labour by a factor of 7 and for capital by a factor of 10 in particular with respect to nuclear energy. We know that the current life cycle analysis does not consider these two factors. As a result, the activities related to the PV are producing a destruction of resources without being able to satisfy any of the energetic needs as given in form of a pyramid going from extract energy to arts in the papers of Hall/Prieto. It is similar to digging out a hole in the earth and then, when finished, filling it again, but even worse, not completely. Could one explain me the benefits to our society of such an operation? This is in our opinion unethical, since you are not only allocating but worse you are diverting resources to the real need of the society which are for instance the elimination of the poverty. PV in region of moderate insolation are increasing the poverty or destructing wealth. We are of course welcoming similar calculation also for the nuclear or fossil field.
3) Price data: As commented, Switzerland is an expensive country. Our study is uses Swiss data. The value of 6000 CHF/ kWp – the average of the 5000 to 7000 is from Swissolar, the lobby association of the solar industry. It is always difficult to compare prices: Looking at at the Sunroof Project of Google for California, you will find at the moment a value 4100 USD/kWp. Knowing that our hourly rates are higher than in US, we judge the assumed value to be reasonable. But the capital cost of 1000 CHF/ m2 indicated in our paper for a mixture of 2/3 roof-mounted and 1/3 field-mounted installations is – in our opinion – too low. Note that land in Switzerland is very expensive. Therefore, we have calculated a low capital requirement. In reality more capital will be needed. Our results should be considered as conservative.
4) References: one comment was related to old references (2009) regarding the weight of the supporting material. To size a support one needs to perform a stress analysis considering loads such as weight, high-winds and earthquakes. Stresses remain stresses and also allowable stresses have not changed in last 100 years. Of course it would be possible to provide newer references.
Many thanks to all contributors for their comments and for new references received. Some were new to us. After this interesting discussion, we remain of the opinion that photovoltaic technology, at least for Switzerland and Germany, is leading our society towards an enormous destruction of resources. In this connection we see elements of questionable conduct from some politicians, against which the civil society should counter-attack. Note that in our paper we have limited our study to the energy, without considering the emission of CO2. We have published in German an article showing that a modern coal plant emits fewer greenhouse gases per unit of electricity than does a photovoltaic plant.
Ferruccio Ferroni/ Robert J.Hopkirk