Shortly after midnight on 28 October 2021 some test results became available which found excessive levels of Iodine-131 in the ventilation system at the Koeberg nuclear power station. This was likely a result of a test which had been done in the afternoon, combined with faulty filtration systems that meant the Koeberg Security Central Alarm Station (CAS) had to be evacuated.

While it is reassuring that Eskom were testing this safety system, no standard test procedure includes evacuating all staff from the security nerve centre of a nuclear plant. Clearly there was something significantly wrong.

While Eskom has not released much by way of detail, it is likely to be a result of a combination of three factors.

Firstly, Koeberg was built starting in 1974 based on a 1960’s design. Modern designs would have a very different ventilation configuration and filtration test procedure.

Secondly, Koeberg is in a harsh marine environment with the base of the containment domes 10m below sea level. In Eskom’s own words when describing the corrosion of these domes: “It is clear that the original designers did not fully comprehend the severe environmental attack which the structures would be subjected to.” (see the Eskom report included in this post). It may be that the design of the ventilation filtration system is just not up to the job in that environment.

Thirdly, and most worrying, are questions around the competency of Eskom in terms of plant maintenance, and the ability of the company to retain skilled staff. With about half of Eskom’s capacity offline, there is undeniably a widespread problem within the company. Incidents such as one of the Medupi generators exploding due to a shoddy safety culture only adds to the concern. The COO of Eskom announcing that he had discovered a ‘horrifying’ exodus of skilled staff from Koeberg specifically is yet another warning sign of problems to come.

When faults start becoming more frequent there is the risk that more than one system breaks down at the same time. This is called the Swiss cheese model of safety. For example, if the main reactor vessel developed a leak, the containment dome would contain the pressure and the radioactive material. On the other hand, if the containment building was corroded and cracked, it would not be a problem by itself.

However, it would be a very real problem if both these things happen at the same time. That could result in the release of a radioactive plume and the evacuation plan being implemented, with hundreds of thousands of people being required to leave their homes. The last time this evacuation plan was tested, the public were not allowed to observe as usual, and so it is difficult to have confidence that it would work given the current population levels.

It is worth noting that all standard home and business insurance policies exclude any loss of damage due to an accident at Koeberg. The insurance companies have done the maths and found that Koeberg is not worth the risk. It is time Eskom comes to that realisation and stops spending billions on patching up this plant which is rapidly approaching its end of life.

We need Koeberg to be running right now, and the extended shutdowns required for refurbishment are likely to cost the country billions due to additional loadshedding. Given the history, the chances of Eskom completing major work at the plant on time and on budget are remote. By 2024 other repair work to the fleet will be complete, and new generation and storage capacity will have come online. We will not miss Koeberg after 2024, but we will miss it over the next two years. For now the plant should be allowed to run undisturbed by major refurbishing, and allowed to shutdown gracefully as planned in 2024.