The two recent posts focusing on Peter Lang’s wind study have generated considerable debate, and some very stimulating discussion, among BNC readers. This post is a follow-up, which this time highlights Lang’s analysis of solar power and related problems associated with energy storage.
This is about solar photovoltaics (PV), which generate electricity directly via the photoelectric effect. The other rising player in the solar field is concentrating thermal power from deserts, which use a steam turbine to generate electricity via a temperature differential, in the same fundamental manner as a coal-fired or nuclear power station. I asked Peter whether he was planning to do an analysis of CSP. He told me:
“I’ve had a bit of a look at doing a similar paper for CST, but I wasn’t able to obtain the detailed output and cost figures I need. It seems the researchers are holding the figures close to their chest.”
I’ve had similar advice on this matter from Ted Trainer. He has attempted an analysis of CSP, and I might post up a highlight of this shortly, and describe some of the gaps in knowledge that Ted and others are seeking. Lang says the following on this matter:
There are two technologies for generating electricity from solar energy: solar thermal and solar photovoltaic. This paper uses solar photo-voltaic as the example because energy output and cost data are more readily available than for solar thermal. It is not clear at this stage which is the lower cost option for large generation on the scale required (see here): so any cost difference is insignificant in the context of the simple analysis presented here.
Lang’s ‘Solar Realities’ paper (download the 17 page PDF here) is summarised as follows:
This paper provides a simple analysis of the capital cost of solar power and energy storage sufficient to meet the demand of Australia’s National Electricity Market. It also considers some of the environmental effects. It puts the figures in perspective. By looking at the limit position, the paper highlights the very high costs imposed by mandating and subsidising solar power. The minimum power output, not the peak or average, is the main factor governing solar power’s economic viability. The capital cost would be 25 times more than nuclear power. The least-cost solar option would require 400 times more land area and emit 20 times more CO2 than nuclear power.
Conclusions: solar power is uneconomic. Government mandates and subsidies hide the true cost of renewable energy but these additional costs must be carried by others.
The analysis, which focuses on the Australian national energy market (NEM) but is obviously relevant for other countries, considers electricity demand, the characteristics of solar PV and one possible means of storing its energy (pumped hydropower), capital costs of a system that could reliably meet demand for 1-day through to 90 days, and then an attempt to frame these numbers in perspective with an alternative low-carbon energy source — nuclear power.
The ‘Introduction’ of Lang’s paper sets the context quite clearly, with the following statement:
The paper takes the approach of looking at the limit position. That is, it looks at the cost of providing all the NEM’s electricity demand using only solar power for electricity generation. Looking at the limit position helps us to understand just how close to or far from being economic is solar power.
The key characteristics of solar power that are relevant to this discussion can be summarised as follows:
1. Power output is zero from sunset to sunrise.
2. Power output versus time is a parabolic distribution on a clear day: zero at sunrise and sunset, and maximum at midday.
3. Energy output varies from summer to winter (less in winter than summer).
4. Energy output varies from day to day depending on weather conditions.
5. Maximum daily energy output is on a clear sunny day in summer.
6. Minimum daily energy output is on a heavily overcast day in winter.
Backup for solar power is clearly required — to store energy when being generated at peak times and thus deliver energy during times when nothing is being generated (at night, during cloudy weather, and to ensure sufficient winter supply). For this PV backup, Lang focused on pumped hydro in preference to sodium-sulphur or vanadium-redox batteries, due to pumped hydro’s lower costs (the latter do have some other advantages). He also considered transmission requirements.
One key feature of the analysis was his consideration of the problem of just how much energy to store. To have enough backup to meet the total national energy market demand for a 24 hour period turns out to be a much more costly proposition than creating a larger, long-term storage option.
Seems counterintuitive, doesn’t it? Well, it all comes down to those nasty ‘extremes’ — those few days of the year when solar power will give you almost nothing (yes, even the deserts have cloudy winter days, although the problem would be much worse if we were reliant on a distributed system of rooftop PV which was largely sited in the major population centres along the southern and eastern coastlines).
If you’ve only got enough solar PV storage to maintain continuous power supply for 1 day, then you need to overbuild your installed capacity by a truly massive amount to cover yourself for those days when the 24-hour capacity factor of your national system is not 20%, but 5%, or 2%, or 0.75%. To borrow a suitable analogy, under a small energy storage system, you’ve got no money in the bank to tide you over until the next paycheck comes in.
Please do read Lang’s comprehensive analysis to get yourself clear on the full story involved in this matter. I cannot emphasise enough how critical this information is if you wish to understand the implications of a carbon-constrained world based on renewable energy without fossil-fuel backup.
Lang concludes his analysis with these strong words (summaries from the last three sections):
Solar power is totally uneconomic and is not as environmentally benign as another lower-cost, lower-emissions option – nuclear power. Advocates argue that solar is not the total solution, it will be part of a mix of technologies. But this is just hiding the facts. Even where solar is a small proportion of the total energy mix, its high costs are buried in the overall costs, and it adds to the total costs of the system…
The capital cost of solar power would be 25 times more than nuclear power to provide the NEM’s demand [$2.8 trillion for the least-cost solar solution with backup versus $120 billion for nuclear]. The minimum power output, not the peak or average, is the main factor governing solar power’s economic viability. The least cost solar option would emit 20 times more CO2 (over the full life cycle) and use at least 400 times more land area compared with nuclear (not including mining; the mining area and volumes would also be greater for the solar option than for the nuclear option)…
Government mandates and subsidies hide the true cost of renewable energy, but these additional costs must be carried by others.
As noted above, the solar story is not complete without also looking hard at the situation for solar thermal power. I will address this in due course.