The Koeberg nuclear plant is the only nuclear power plant in Africa. Construction began in the 1970s and the plant came online in 1984. It operates under a licence from the National Nuclear Regulator (NNR) which expires in July 2024. In 2010 the Eskom Board approved spending R20bn on refurbishing Koeberg to extend its life by 20 years, provided that it received approval from the NNR. That approval has not as yet been given to Eskom by the NNR.

Currently Koeberg provides 3.5% of national nominal generating capacity, although over the past few years it has been plagued by problems. For most of 2022, unit 2 has been offline and the plant has run at half of its nominal capacity.

R20bn in 2010 is how much in today’s money?

Until the presentation to parliament on 27 September, Eskom maintained that the refurbishment would still cost R20bn. This made no sense, as the exchange rate in 2010 was R7.50 to the US dollar. In 2010 R20bn was equivalent to $2.67bn. Adjusting that figure for US inflation from 2010 up to 2022  gives a figure of $3.63bn in today’s money. Using the exchange rate on 28 September 2022 of R18.00 to the US dollar this is equivalent to R65.3bn.

On top of this, the Eskom presentation includes R1.5bn in extra costs in the form of interest paid due to the delays. Framatome also invoiced Eskom R1.1bn as a penalty for wasting their time over the attempted steam generator replacement debacle in early 2022, as reported in the Daily Maverick.

These three figures total R67.9bn.

Is that a realistic estimate?

Large scale Eskom projects typically overrun costs by two or three times. While the R67.9bn above results from a simple calculation, guessing the overrun is speculation. If the cost doubles from the initial estimate, that would make the eventual cost over R130bn.

To put that into context, using the latest costs, R130bn could purchase about 8GW of solar PV. That would on average generate more power than the Koeberg plant running at full capacity.

But the above still excludes several costs. In order to refurbish the plant, which includes replacing three steam generators in each unit, it needs to be taken offline for extended periods. Eskom estimated that it would take five months, although their first attempt resulted in unit 2 being offline for 8 months, and they didn’t manage to actually install the new steam generators. Eskom plans to try again on unit 2 in September 2023, and to attempt the refurbishment of unit 1 starting in December 2022.

These long term outages result in direct costs to Eskom, as well as indirect costs to the economy due to load shedding.

The costs of long term outages of Koeberg

Eskom has made hourly historical data available via their data portal and Clyde Mallinson has used this information to work out precisely in which hours there have been load shedding, and when Eskom has used diesel to run its OCGT.

This has enabled Mallinson to calculate what the effect has been of having unit 2 of Koeberg offline for eight months of this year. According to these calculations, as a direct result of the extended outage Eskom has used an additional R5bn worth of diesel due to the 8 month outage.

In addition to the human suffering caused by load shedding, it also has a significant effect on the economy. According to Mallinson’s calculations, 1650GWh of avoidable load shedding have occurred as a direct result of the extended outage of unit 2.

There are various views about how much each kWh actually costs the economy, with figures varying from R10/kWh to R100/kWh. Mallinson uses a conservative figure of R20/kWh which means he estimates the cost of the unit 2 outage to the economy in 2022 has been R33bn. Using the official figure from the IRP 2019 of R87.85/kWh makes that figure R145bn.

Eskom plans to take unit 1 offline in December 2022 to attempt the refurbishment work, and then to retry the same for unit 2 in September 2022. If these two outages do not take longer than planned, are only six months, based on Eskom’s predictions of load shedding, Mallinson estimates the cost to the economy will be R175bn due to these two upcoming outages.

What is the total?

Adding up these costs for the three outages and the conservative R67.9 billion in direct costs, the total cost to the Koeberg refurbishment will be R281bn.

If the R67.9bn engineering costs were to inflate to R130bn, as it has for projects such as Medupi, the total will be in the region of R345bn.

Eskom has claimed that Koeberg will produce the cheapest electricity of all its generating units, although they have not actually released an updated business plan with figures to justify that. Eskom also dodged a question about their business plan in parliament from Kevin Mileham recently by terminating the session before questions could be answered. However, if all the above figures are taken into account, in terms of cost to the country, if the refurbishment of Koeberg goes ahead it will produce the most expensive electricity in the history of South Africa.

Won’t we need Koeberg after 2025?

The ageing coal fleet will continue to experience breakdowns and the only solution is to build large amounts of new capacity. There is a growing realisation of this fact and some projects have recently had contracts signed. Renewable energy projects can come online within two years, and by 2025, if the right decisions are made, we will see significant new generation capacity come online from 2024 onwards. 

Conclusion

South Africa needs Koeberg now in 2022 and 2023 at the height of the current load shedding crisis far more than it will in 2025 and beyond.

It is time for Eskom to seriously consider abandoning the refurbishment and leave the plant running until July 2024. That will avoid the risk of further embarrassment due to bungled attempts to manage the project, save Eskom the balance of the engineering costs, and also avoid unnecessary load shedding resulting in losses to the economy measured in hundreds of billions of Rand.

---