The case for nuclear power in South Africa

There has recently been a meeting in Stellenbosch where some energy modellers tried to plan South Africa's energy future.  They had some interesting concepts, not least of which was that the most important driver of any plan should be a reduction in our carbon emissions.

If you or I tried to make a plan, we would probably recognise that what South Africa needs, and needs as soon as possible, is enough power at as low a price as possible.  It goes without saying that the power should be reliable, available on demand 24 hours a day, and that the source of power should have a reasonably long life, so that we had some hope of achieving a return on the considerable investment needed to provide us with the power we need. 

Most people would think, if they stopped to consider a basis for planning our energy future, that this was a reasonable set of requirements.  However, they may also be moved by the fact that our President offered massive reductions in our carbon footprint at Copenhagen in 2009. Surely this should also be taken into account?

If they were so moved, they might also remember that the President’s generous offer was conditional upon large-scale aid from the developed nations, those nations that have done most of the carbon emitting for the past few generations. There is not one sign of any such aid.  There have been several conferences since Copenhagen, and none of the meetings has made any progress towards making aid commitments.  Accordingly, our armchair planners would probably reject reducing our carbon footprint as an additional requirement.  They would say that it is a constraint that can wait until we see the money.

However, the Stellenbosch planners believed that we should still strive for lower carbon emissions.  They believe lower emissions can be achieved while still producing cheap reliable power by using renewable energy such as wind or solar power.  They are wrong.

It is true that the costs of renewable power have fallen in recent years.  However, that does not mean that renewable energy is economic.  Remember, we need sufficient reliable power at as low a price as possible.  Our experience to date with wind power alone is sufficient to cause major concerns.  Eskom installed three large wind turbines at Klipheuvel.  One of them has failed already, and the other two have produced about 10% of their potential output. 

The Darling wind farm was launched in 2008 to great fanfare.  The four turbines have the potential to generate 5.2MW.  The power is bought by the City of Cape Town, which recently reported that the farm had generated 30.8GWh since inception.  Its average output is therefore about 720kW, less than 14% of its potential. It cost R70 million to construct. I have a 2litre diesel car that has 120kW output. It cost just over R100 000 in 2008. The wind farm produces six times more power, but costs about 700 times more.

In Europe, North America and China, all of which have been sold tens of thousands of wind turbines, the story is similar.  This is not yet an established technology.  You need ordinary power to ensure continuity of supply when the wind doesn’t blow.  You need to be able to dump the wind power somewhere when the wind does blow and there is no demand for the energy. Denmark, the home of wind turbines, has to send over half of what it generates annually out of the country at a loss.  The result is that Danish power is presently the most expensive in Europe.

Surprisingly, the technology is comparatively untested.  The large units presently being favoured have only been in operation for about seven years.  Mechanical and electrical failures are still very much more common than the tried and tested conventional alternatives. A recent European report has shown how output falls as the installation ages, until after about 12-15 years it has to be replaced because maintenance costs exceed revenue. In many places there is a realisation that, in the haste to set up a wind farm, they forgot how much it might cost to take it down.

Siting is critical, and the technology of selecting the optimum site has improved in recent years. However, there remain mistakes – one US wind farm was recently hit by a tornado.  Skittles, anyone?

Wind power is the cheapest of the renewables.  The lowest offer under our Integrated Resource Plan IRP2010 was 89c/kWh.  This is impressive, but it has to be remembered that it is private capital at risk.  There is no guarantee that the project will yield investors a reasonable return.  The wind may not be quite as good as the developers had hoped, or the turbines may need somewhat more maintenance than assumed, or maintenance technicians may cost more than budgeted. 

And while 89c/kWh may be impressive, it has to be seen against Eskom’s present costs of 32c/kWh, projected Medupi costs of 53c/kWh, and present independent power producer costs of 77c/kWh.  The recent Eskom price increase includes a significant allowance for the extra cost of the renewables that IRP2010 will bring.

The other renewables are even more expensive.  Solar PV costs have also fallen, but are still over R1.20/kWh at the MW scale.  Solar thermal power is really expensive, around R2.50/kWh, but it is a really unproven technology – last year it made about 0.01% of the world’s electricity.

So what does this leave us? We need sufficient, reliable power at as low a price as possible. The best bet by far is natural gas.  The average global economy gets about a third of its primary energy from gas.  We get about 2%.  There is a strong indication of gas in the Karoo, but it still requires exploration.  We should not be dithering.  There is gas in Mozambique, and, reports to the contrary, it is clearly economic to pipe it into South Africa to generate power.  Sasol is already doing that on a 200MW scale at Sasolburg. Liquefied natural gas is a widely traded commodity, and all it requires is a regassing station.  Gas turbine power stations are cheap and can be built rapidly.  Siting them at the coast would allow sea water to be used for cooling.

Gas would even contribute to a reduction in carbon dioxide emissions.  In the US, thanks to shale gas, the price of natural gas has fallen by a factor of four, coal-fired power stations have been shut down, and the US now emits less CO₂ than it did in 1990. Even Canada has been affected – the increase in energy supply in the US has meant that Canada has been saddled with excess oil.  Recently you could buy tar-sands product at $53/barrel.

But we don’t yet have our own gas, and we are clearly running out of coal, if only because the rest of the world has started to value the rubbish we are presently burning.  Coal mines can get better prices on the international markets than Eskom is prepared to pay.

That leaves us with nuclear.  There is a belief that nuclear is expensive.  It is not much more expensive to build than conventional thermal plant, and with low fuel costs and a long life, the costs of electricity are low.  Indeed, today Koeberg produces the cheapest electricity in South Africa.

Part of the erroneous belief in the high cost of nuclear power arises from the fact that it comes in large units, 900MW upwards.  But that is reliable, continuous power. Koeberg will shortly be delivering 2100MW 80% of the time.  To do that with Darling wind farm technology would require over 9300 wind turbines and would cost about R200 billion today.  For that money you could build quite a few Koebergs.

The case for nuclear in South Africa could hardly be stronger.  Models that show otherwise are driven primarily by the reduction in carbon dioxide emissions, not by our need for cheap reliable power. They rely on assumptions about the possible reduction in renewable energy costs that have little basis in reality. There seems to be a belief that policy should determine our energy future.  The cost of power and the reliability of its supply are far more important. That should be the basis for the next IRP.