300 GW/a: PV as a global energy pillar09 / 2012, Markets & Trends | By: Hans-Christoph Neidlein, Jonathan Gifford
Interview: While 300 GW per year (GW/a) may give hope to a PV industry facing many challenges, Christian Breyer believes that achieving the number is vital in realizing a worldwide energy transition. With colleagues at the Reiner Lemoine Institut, Breyer is creating a multifaceted model for a 100% renewable future and pv magazine found out what it would mean for PV.
If we are to consider the global PV industry growing to 300 GW/a by 2025, it would be a part of a major energy transformation. Why do you think such a transformation is necessary?
The key is energy, always energy. It’s the driver of the world and of course we hear that about oil but at the end it’s energy. We need to reach a minimum level of wealth to get the population growth down, and then we have a chance to manage all the further problems in the world. To stabilize the population at 10 billion people, we need to get people much wealthier than they are today in the developing countries. There is a minimum per capita income to get the birth rate per woman down to a level of two, and for that minimum income you need a minimum energy level. Without managing global population growth, we will for sure lose everything. If we get that fixed then we have a good chance to find a path to sustainability.
How then can this energy be provided in a sustainable way and how significant will the role of PV be?
If we start with the assumption that the world needs 100% renewable electricity, the core pillars in this century will be PV and wind power. If we envisage the world having a population of 10 billion people, they consume 7,000 kilowatt hours/a per capita, this might be the final electricity demand. Then 10,000 GW to 12,000 GW of PV installed capacity is needed [additional to wind and other renewables]. Then if you have to replace 12,000 GW when it comes to the end of a 30 year lifetime, you need an annual installed capacity of 400 GW – just to stay in equilibrium. For at least the long term this should be the order of magnitude we should think. The question is then how stable is this 12,000 GW and when it can be realized?
Fuel replacement applications for PV, in which PV can be added to existing diesel or oil generation facilities, seem to be an area where an economic case can be made for large-scale PV deployment today. How much diesel capacity do you think could be enhanced in this way?
Of course there is still the base capacity, and this the diesel can provide. But with these power plants, the cost of generating the electricity is 80 to 90 percent dominated by the fuel. So if you do not burn the fuel it costs you nearly nothing. So, of course, the capacity is available and it balances these hours when the capacity is needed. For oil-fired power plants, the number would be roughly 600 GW and most of them are used only for several hundred hours a year – for really dramatic situations on the grid, like, for example, in February this year during the northern hemisphere winter, or if you are close to the equator in the summer period when it’s very, very hot and electricity is needed for air conditioning.
For only the diesel, which is a fraction of all the oil-fired power plants, there are several numbers out there. I claimed in one piece of research that 30 GW might be possible to be upgraded by PV, I have now seen in the latest quarterly report from SMA, the number is put at 20 GW and I’m quite sure it’s somewhere in between.
But then the question is: What is the order of the diesel generators you can really upgrade with PV? Typically in many countries you have it really as a backup capacity, or for when you have shortcuts in the grid. But for those cases it’s very difficult to combine that with a PV system because you never know when you need it. We have done some calculations with the combination of PV, battery and diesel and we have seen that in the off-grid region it is really highly economic. The amortization period in many parts of the world is only four to six years, so that is very profitable.
How would you break down the 300 GW/a by market segment: residential, commercial, rooftop or utility-scale?
It’s really a question of whether the PV market will evolve in a more decentralized or a centralized way. And I’m quite sure in the very end it depends on the social and political system. For example, if it’s a very liberal country then I’m quite sure it’s more likely to have the market develop more towards the homeowners and if it is a more centralized country then it might tend towards more centralized power plants.
So if we compare Europe and the MENA region, you have hardly any discussion about PV systems on the rooftops in the MENA region, but you have a discussion about large-scale power plants. And in Europe it’s always a discussion that we have it on the rooftops but it also includes large-scale PV power plants. It really depends on the system.
So globally it’ll always be a mixture of both, the decentralized component is very powerful for PV, so it might be two thirds rooftop, from the very small up to several megawatts, but also with PV power plants. But from country to country it really can change, so both are possible.
Continued cost reductions for PV will be important in all market segments, do you see the potential for the PV learning curve to continue to deliver that?
If you look way back to the mid 1950s, we can observe a really stable learning curve on that line (see diagram above right, which starts in the 1970s). With every doubling of production capacity we see a cost reduction of roughly 20 percent [for PV modules]. And this trend is very stable over five decades now. So you should never bet against the trend, and the trend is very stable. We had exactly this in the semiconductor industry where we are familiar with Moore’s Law. It’s very stable, that for every doubling you can reduce the cost. If you have a closer look at the mid 1970s, where we had the beginnings of terrestrial PV, we see maybe a few different trends, but it’s more a question of statistics – where we had a higher learning rate from the mid 1970s to the late 1980s and then a somewhat lower learning rate in the 20 years afterwards. But if you see the price of a crystalline silicon (c-Si) module, in the years 2010 and 2011 they are significantly lower than the learning curve. So now were are back to a very steep line, so up to now for a long-term period we have been very stable on that line.
And how stable do you believe that learning curve will be in the future?
We performed some research two years ago looking into research and development investments into PV to investigate whether there was a limiting factor in the cost degression: Are there enormous investments to bring the cost down to a stable level?
What we saw in our research was that there was an enormous explosion in R&D investments, starting roughly in the year 2000, when the feed-in tariff (FIT) markets started to become relevant for the PV industry. And if we know how long it takes to have a new technology established in the market, then we see in the last five years at least, they are still before market introduction. So now we will see all that new machinery in the market; we have high efficiency cells in research for ten to fifteen years and now we see that they become part of the products. We now see low cost c-Si modules with 18 percent efficiency, or something like that. This is really new and this is a consequence of technological advancements.
These advancements are not only on the cell level, they are on a material level: in ingoting, in wafering, in the modules, and all the devices as well, like the inverters and the machinery. It is right across the full value chain.
We do not know – even some thin film technologies will survive. But all these enormous R&D investments are starting to become effective on the product side, and of course they reduce the cost. The cumulative R&D investments by the end of the 2000s was in the order of €50 billion, but it’s less than two percent of the investment in nuclear, which gives you an idea of just how low this number is compared to the powerful outcome of the technology that we got for that relatively small amount of money. We have also got a lot of progress to come in the coming years. To conclude, I am quite sure that the learning curve is at least stable in the coming years and by the end of the decade we will end up with systems delivering between €0.70/Wp up to €1/Wp depending on market segment.
What about the learning curve on the system level?
We can observe that if we have a 20 percent reduction with a doubling of capacity on the module level, then it’s roughly 15 percent on the system level, which is still a very high learning rate. It’s a little bit lower but not significantly lower and this is very, very relevant.
What about the costs between different markets, say the U.S. and Germany, on a system level?
Well we are talking about 300 GW/a global market, so if you look at all the markets of the world, many are several hundred MW/a and if you look at those then they always tend to be not far away from the most competitive markets in the world, which currently have three to seven GW/a.
Is that significantly affected by the political environment, for example the permitting processes, because we see big variations between say the German market and then, by contrast, parts of the Californian market?
If you don’t have the political will, you won’t see PV. A factor is the willingness to have PV. But I believe that the pressure will sharply rise to introduce PV, because governments will have to explain to citizens and consumers in certain countries that energy prices are really high. People will know that in a neighboring country they are much lower because they have renewables, then governments will have to explain that. In a lot of countries there are riots due to high energy prices, and typically they are paid for by subsidies, but also these subsidies explode. In some countries like South Africa or Egypt, there are enormous subsidies in their energy system, but they have to finance it on the national budget level and of course the higher the fossil fuel prices are and the lower the renewable costs are, then it becomes a question of why should citizens pay so much in subsidies, because in the end a lot of people earn from these subsidies through corruption.
At what point do you see storage entering the equation?
Storage is really one of the last system components we would use, nevertheless the market will start much earlier, because from the end-user or grid-parity perspective it makes economic sense. If consumers have to pay high retail prices for electricity, and you can store the electricity you can produce cheaply – the PV power from your rooftop in a battery system – then it makes sense from the end user perspective. That is why I expect that four to five years after retail grid parity we will see PV plus storage. But this is then a question of legislation: Is storage allowed? Must you pay a higher base price for the power excess to the grid? And what to do with the lost tax revenue – must you pay an additional PV self-consumption tax? This is all a question of legislation.
Talking of politics, will a 300 GW/a scenario only work if there are no subsidies for fossil fuels or for nuclear generation?
I am quite sure that the subsidies for these industries will come to an end because the cost is so high that the societies cannot afford them. As we see currently in Japan, everyone is now aware of the consequence of the subsidy for the insurance of nuclear power plants. The Japanese government has to pay the cost and it’s very high. TEPCO was the fourth largest utility in the world and it went bankrupt with the Tsunami [and Fukushima-Daiichi meltdown]. The Japanese government had to pay €10 billion in this summer just to keep TEPCO out of bankruptcy. The full costs are currently estimated at up to €100 billion. And this is a kind of subsidy.
Do you think for certain markets FITs will be needed in the coming years or now, based on the competitiveness of PV, that is sufficient?
I am quite sure that we need such mechanisms; because it really depends on which market or country you are talking about. If you are in a liberalized country and energy market, then without regulation it will never work. This is because the merit order effect is against such investments, because you will invest in PV power plants, they will reduce the price for electricity, this will reduce the profitability of the balancing power plants, and the entire system will collapse. But in a country in the MENA region, for example, there is no problem at all because in the very end all the investment and all the money that is needed to generate the electricity is added up and it is more or less divided by the consumption and everyone has to pay that full cost price. And we had that before the market liberalization in Germany and in Europe, and I’m quite sure we have to come back to a full-cost price system. And the FIT is a full-cost price system.
pv magazine’s sister publication photovoltaik magazine, is investigating the idea of whether a goal of 200 GW in PV capacity can be installed in Germany, how different is that to the idea of 300 GW/a globally by 2025?
The 200 GW for Germany in total and 300 GW/a globally are coupled together. This is because if you see that there are 82 million people in Germany, which makes up roughly 1.2 percent of the (future) global population, then if you multiply that 200 GW by 60 - to include the different irradiation levels in the world – then you end up with 12,000 GW, which is the basis for 300 GW/a PV. So these numbers are closely coupled.