Schalk Cloete is creating his own 5-part independent Global Energy Forecast to 2050, to compare with the next IEA World Energy Outlook, due in November. Many of his assumptions are different from the big institutions, not least that technology-neutrality will be widely adopted as the best policy, as carbon budgets are exhausted around 2030. There are other big differences too. He starts with wind and solar, two technologies that the IEA and others have consistently underestimated. His projections include six different simulations of a cost-optimal technology mix, looking at how nuclear, gas, coal, hydrogen and batteries affect the prospects for wind and solar. The red flag is grid costs: they will be expensive as wind and solar grows, opening a door for nuclear according to his analysis.
You can read the author’s first article which introduced his methodology. The next three, published here in 2-3 week intervals, will cover fossil fuels; nuclear, biomass and CCS; battery electric vehicles. After the IEA WEO 2019 is released he will compare his predictions with theirs. On his journey, Cloete welcomes comments and feedback from our readers.
Introduction
Wind and solar power have performed very well over the past decade and this trend is set to continue. However, there are vast differences in opinion regarding the future of these variable renewable generators.
On the one hand, there is the more traditional bodies like oil companies and the IEA who continually underestimate wind and (particularly) solar energy growth. The trend in revisions of recent IEA World Energy Outlooks is illustrated below. The two diamond symbols in the figure show that I expect this trend of upwards revisions to continue for several more years, particularly for solar.
On the other hand, there are many who advocate for complete renewable energy dominance within relatively short timeframes. This linked study is a recent example. In this case, 2040 generation from wind and solar PV are projected to amount to about 25,000 and 50,000 TWh of annual generation respectively, i.e. about 4x and 7x greater than my forecast.
As an illustration of the difficulty in maintaining the continued exponential growth envisioned by such renewable-dominant scenarios, historical wind and solar growth is plotted below as an example, using data from the BP Statistical Review. Although I only did this after my forecast was made, simple extrapolation of these trends to 2040 (with accounting for retirements) end up in the same ballpark, at about 8,000 TWh of wind and 6,300 TWh of solar.
Only time will tell where we will end up on this very wide range of possible future outcomes. For reasons to be outlined further below, I believe that wind and solar growth will follow a more traditional energy expansion profile.
But first, let’s look at a couple of graphs from the forecast.
Wind power
As shown below, my forecast for wind capacity and generation falls roughly in between the New Policies Scenario (NPS) and the Sustainable Development Scenario (SDS) of the IEA. From 2030 onwards, my generation forecast gets a bit of a boost relative to the IEA forecasts by assuming higher wind capacity factors as large numbers of old turbines start to retire and more offshore wind is installed.
Expressed as a fraction of electricity, my forecast is closer to the New Policies Scenario due to my more optimistic assumptions about electricity demand growth. When considering primary energy, I end up between the NPS and SDS scenarios due to my expectation of slow primary energy demand growth, particularly for oil.
Solar power
I’m more optimistic about solar. As shown below, my forecast is higher than all the IEA scenarios, both in terms of capacity and generation. I expect solar PV cost reductions to continue outpacing the IEA assumptions. Additional tailwinds will come from the gradual shift of global growth to sunnier regions.
In terms of electricity and primary energy shares, my forecast ends up slightly below the IEA SDS. The SDS is very optimistic with respect to energy efficiency, leading to relatively low electricity and primary energy growth. Although I am optimistic about energy efficiency and reduced energy consumption via intelligent lifestyle design, I do expect that the developing world will continue to demand robust energy and (particularly) electricity growth.
Wind and solar performance in a cost-optimal mix
To better understand the dynamics of wind and solar, we’ll look at some results from a simple power system model that optimises investment and hourly dispatch of 12 different technologies. Details about this model can be found in this working paper (authors Schalk Cloete and Lion Hirth). However, I have reduced the wind, solar, battery and electrolysis costs to €1200/kW, €500/kW, €170/kWh and €400/kW, respectively. These wind and solar costs are about 15% and 21% lower than the IEA 2040 cost projections for Europe used in the article linked earlier. Battery and electrolysis costs are 7% and 13% lower, respectively.
The model is loosely based on Germany and uses wind and solar performance representative of Europe in 2040, with capacity factors of 30% and 13% respectively. It has the simple objective of deploying the available technologies in such a way that total system costs (capital, operating costs, fuel costs and emissions taxes) are minimised.
Six different simulations were completed:
- All in: This case includes all available technologies: nuclear (€5000/kW capital cost), coal, gas combined cycle (NGCC), gas open cycle (OCGT), onshore wind, solar PV, coal and gas with conventional post-combustion CO2 capture and storage, hydrogen-fired combined and open cycle plants, battery storage, and PEM electrolysis.
- No nuclear: A case where nuclear is eliminated as an option.
- No CCS: A case where nuclear and CCS are eliminated as options.
- No batteries: A case where nuclear, CCS and batteries are eliminated as options.
- No electrolysis: A case where nuclear, CCS, batteries and electrolysis are eliminated as options.
- Added grid costs: the “All in” case with €10/MWh of grid-related costs added to wind and solar.
Optimal capacity and generation mixes of these six cases are shown below with a CO2 tax of €100/ton and a discount rate of 7%. In addition, the black lines on the two plots show the levelised system cost in the capacity graph and the system CO2 emissions intensity in the generation graph.
The first three cases in the above figure clearly show that the elimination of nuclear and CCS as options substantially increases the optimal wind and solar market share. However, this comes at a moderate increase in cost and a large increase in emissions. The emissions come from the need for sizable unabated NGCC power generation to back up wind and solar power in the “No CCS” case.
The “No batteries” case shows that batteries can play a significant role in increasing the optimal solar share, whereas wind share is unaffected. This is understandable since batteries are best suited for short-term energy storage following the daily generation pattern of solar. A similar small, but significant effect is shown in the “No electrolysis” case. It is therefore clear that these energy storage technologies can significantly enhance wind and solar performance.
Expensive wind and solar grid costs will favour nuclear
However, as shown in the final case, when the added grid-related costs of wind and solar are accounted for, the optimal mix consists almost exclusively of nuclear power. These added costs arise from the spatial variability of wind and solar power, requiring large grid expansions to bring the produced electricity to demand centres and capitalise on the seasonal complementarity of wind from the north and solar from the south.
As Germany is currently illustrating, in addition to being quite costly, this grid buildout faces other problems as well. Much like nuclear power, grid expansion faces strong public resistance that causes large cost escalations and multi-year delays. This is a major problem when a rapid transition is targeted.
If one recognises these practical challenges with integrating high shares of wind and solar and entertains the possibility of true technology-neutrality by the end of the 2020s (as assumed in this forecast), it is easy to envision scenarios with relatively slow wind and solar growth. However, the momentum behind wind and solar is strong and their impressive cost reductions continue. Therefore, I have opted for middle-of-the-road numbers in this forecast.
So, what do you think? Am I too pessimistic? Too optimistic? Do you agree that solar will overtake wind around 2030? Let me know in the comments section below.
***
Schalk Cloete is a Research Scientist at Sintef.
Erik says
I see you have cost of solar set to 500€/kW
Is that cost constant in the model until 2050?
I do think that solar will continue to accelerate. Cost will continue to drop, both silicon, but also organic cells. I saw a research group that had made a cell a few layers thick, meaning that there is no lower limit to cost.
It will be integral in everything because it adds very little cost to the final product, but a lot of return.
Even today, it’s foolish to build a roof that’s not solar. The return of investment is about 7 to 10 %. Even cell phones can benefit from a layer.
Increase availability of Direct Current products, and add a little behavioral change, and solar will be our main source of power in just 10 years, because for the individual, it will be the cheapest way to get power.
Schalk says
This is a steady-state model optimizing a mix of technologies that can be deployed around 2040. Each technology is assumed to operate for 25-50 years and must offer a 7% return on investment.
Yes, I’m very interested to see how solar develops over coming decades. If you run these kinds of system-scale assessments, you see a rapid value drop with increased market share because of the highly pronounced variability, particularly when balancing generators are low-carbon. But I do agree that the impressive cost reductions will continue and that different approaches will emerge to combat the value decline (although each of them also imposes their own costs). It will therefore be a race between cost and value declines over coming years.
I’d be interested to see some calculations behind your 7-10% return for a solar roof. I hope it does not assume that residential solar is worth residential electricity rates. This is only the case if solar capacity can directly displace grid and generating capacity which, at best, only applies to the first couple of %-points of market share in growing economies where grid and generating capacity is still being expanded. For all the rest, solar only displaces power plant running costs, which is at least 3x lower than residential electricity rates.
Bas Gresnigt says
In the past 20years the IEA always:
– underestimated the increase of solar and wind greatly; and
– overestimated the increase of nuclear greatly (nuclear share in electricity generation continued to decrease).
You assume that the IEA stopped doing that.
But I don’t see a valid reason for that.
Peter Farley says
I think you understate the effect of overcapacity, flexible demand and increasing capacity factors for wind and solar. For example with rooftop solar becoming more efficient and cheaper, standard roof installations are now approaching 7 kW but still mainly face south (north). It is now more optimal for many users to put 8-10 kW or more east-west solar with an undersized inverter. In much of the world that system will generate 8-10 MWh per year, less than a conventional system but more power in the morning and evening and less power in the peak of the day, so less curtailment of exports and less imports early and late. An efficient free standing house can easily run on 5-6 MWh/y, in some cases much less, so most solar houses will be net positive generators
At the same time energy efficiency is slowly driving down household demand and more of that demand can be scheduled into the solar window, washing, water heating, pre-heating/cooling houses etc. That means the house can still export energy at low rates probably for 4-8 hours per day and households will still be better off as less of their energy will be curtailed.
If they still have 15-30% of their generation curtailed they won’t know or care, after all they buy houses which have an occupancy rate about 1/3rd of what they had 60 years ago, their cars are used as you say for about 4% of their capacity so, so a solar system that wastes 30% of its generation is a bonus.
Similarly, large rotor wind turbines generate for many more hours per year, so curtailment during relatively short periods of high winds is less damaging to their economics. Therefore new turbines will still be built. Further they will be built in areas previously considered unsuitable such as southeast China and the USA, where their relatively lower yield will be more than offset by transmission cost savings or fuel costs for nuclear or gas generation. Then there is the incipient boom in offshore generation.
Across the system demand scheduling has huge potential to reverse the old “off Peak” paradigm to charge EVs, heat water, transfer water and sewage and even preheat/cool commercial and industrial buildings and processes when the power is available. Further, as electricity costs fall electrical refinement of metals vs gas and blast furnaces will mean that much heavy industry demand can also be electrified.