Why Koeberg nuclear project was a success - expert

Cape Town – The construction and operation of Koeberg nuclear power station was a success, a former top-level official of the project said in a rebuttal to a former colleague, who criticised it for going overtime and over budget.

Fin24 user Martin Alberts – writing under a pseudonym as he is still in the industry – wrote an opinion story on Fin24 about his experience during the project, concluding that “nuclear will be the death of this country”.

READ: Ex-Koeberg nuclear plant worker lashes out at SA plans

See all Fin24's stories and opinion contributions on South Africa's nuclear plans

In response to Alberts, former Atomic Energy Board member Geert de Vries wrote an opinion piece, defending what he believed was a successful project.

De Vries begins by introducing himself, before tackling the critique of the Koeberg project. He then discusses South Africa’s failed Pebble Bed Modular Reactor project, before discussing the Fukushima nuclear and tsunami disaster in Japan.

De Vries writes…

I left the Atomic Energy Board at the end of 1976 as chief scientist (physicist) and joined the French consortium (of four companies) that had in August 1976 started construction of Koeberg.

I worked in the Johannesburg office, the hub in SA that communicated daily with the Eskom head office at MegaWatt Park, the Paris head office and the Cape site office. As such, I knew and dealt with all 60+ staff of the Eskom HO project team, all top and middle management and other site staff of the four consortium members (about 200) and Eskom site (about 50), and of course Paris.

Over the years I saw and forwarded each and every letter to and from Eskom and between consortium members, attended all meetings with Eskom in Johannesburg and nearly all monthly site meetings, besides many missions in Paris.

I also once acted as consortium site manager for several months while we were looking for a replacement from France. And as a special assignment in 1981, I edited and partly rewrote the 4000+ pages Safety Analysis Report, making me familiar with all aspects of the plant.

Now to the issues raised by Alberts. The misinformation fed to Fin24 readers cannot be allowed to stand.   

Defending Koeberg

Alberts begs "to differ quite strongly that Koeberg was finished on time or under budget”.

I personally dealt with the milestone submissions. We kept graphs of progress in terms of mass (mainly civil work), number and percentage of systems completed, and value passed, that went into the reports to support/justify our periodic milestone invoices to Eskom. These were based on report-backs from equipment manufactured and shipped from France, from site meetings and progress reports and related sources. They were countersigned by Eskom staff.

The two Koeberg units were spaced 15 months apart. December 1982, the month of the African National Congress (ANC) bombing attack, was month 75 of the 80-month contract period for Unit 1 and month 60 for Unit 2. At that time we had passed all earlier milestones successfully, and were on schedule - plus or minus two-to-three months on some of the systems that were not on the critical path - to pass the next milestone, and to complete Unit 1 five months later, around April-May 1983.  

I provided this information to Dr Kelvin Kemm, who used it in his write-up on Fin24, with which Alberts begs to differ.

READ: Nuclear adviser responds to his critics

The attack caused construction to stop. It was followed by the removal of damage and repair work. All in all, it was a year’s delay, which was not caused by the consortium’s construction effort. Unit 1 was completed a year late, on 4 April 1984. Unit 2 followed the schedule, 15 months later in July 1985. The ANC attack successfully derailed the fully on-time and on-budget project.

Alberts also said: “In reality it had very little South African build in it”. There is no logical link between increasing local content and being within time and budget. If anything, local content might cause delays because of learning curves. But more to the point, Koeberg had a surprising amount of local content even though the contract did not specify it and only stated that efforts should be made to promote local content. And this was done.

All civil work was local content and amounted to 20% of the contract value. The French civil contractor invited local companies like M&R and LTA and others to form Koeberg Civil Contractors (KCC). This provided work for up to 4000 local workers during the intensive initial civil construction period, and later to a few hundred.

It meant that much of the cost was spent locally, benefitting local industry and keeping the money in SA. The senior civil engineer in charge of daily operations was a South African Wits-educated civil engineer. Local companies were also involved in all delivery to site, all cranage, manufacture of many non-nuclear vessels and other non-nuclear supplies such as piping and steelwork, stainless cladding of the reactor buildings, all power and non-nuclear instrument cabling, cable trays, the construction of the intake basin with breakwaters, the underwater construction, the design of the essential cooling systems pump station as well as other design work and manufacture that came up due to the changes to the contract (more on that below). To paraphrase Alberts, “it had very MUCH South African build in it”.

Over the years, I did much of the administration of the few hundred change orders and the local and site orders, and my unconfirmed guestimate of the localisation content is that we reached 35-40%. This is better than most countries achieve on their first plant. The French were thankful that our industry was so capable and could play its part. It reduced the hassle factor considerably.

Alberts writes: “The lion's share of the work went to Frametome, Framex, Alstom and a host of French companies such as KCC that were in joint venture with a couple of local companies. This was a pre-condition to France ‘lending’ us the money to build nuclear (SA wanted a German build but was unable to get financing); we had to use a French design (Tricastan was the sister plant which used old technology) and French contractors to build.”  

This string of enormities shows how little Alberts understands the process and the configuration. Obviously, if you buy a French plant from a French contractor, that contractor does the lion’s share of the work. That the contractor would do his own work was not a precondition for lending. And I already dealt with the unexpectedly large local content above.

Alberts lists the consortium partners and, apart from mis-spelling Framatome, Alsthom-Atlantique and Tricastin, he gets it wrong. He omits the civil engineering partner Spie-Batignolles and the project management partner Framateg, includes Framex who was not a consortium partner, and calls KCC a French company while it was fully South African.

Also, it is not clear why ‘lending’ is between quotation marks, as it was a standard financial transaction for infrastructure building, where a utility borrows and pays back later. The only precondition is whether the borrower has the standing to borrow, and can he be expected to pay back. In the 70s, Eskom could.

SA did NOT “want a German build”.

Eskom as sole adjudicator had selected a US-Swiss-Dutch consortium offering a Boiling Water Reactor (BWR) as the number one shortlisted. There was nothing German there. And it was not a problem of getting finance that made Eskom change to the number two shortlisted French consortium that offered a Pressurised Water Reactor (PWR). It was the refusal by the Dutch government to provide a performance guarantee to Eskom concerning that Dutch partner. Without that performance guarantee, Eskom was quite rightly not prepared to engage with number one.

The Dutch government did not provide that performance guarantee because it disliked SA’s apartheid policy and thereby meant to punish SA. This was thus a political decision, possibly reinforced by the fact that in 1976 that Dutch partner had not yet fabricated a reactor vessel of the size for Koeberg, only much smaller ones, and it could place the Dutch government at risk in case of non-performance.

The French government had no such qualms, and issued the performance guarantee, also in the knowledge that its industries would be able to deliver because Koeberg’s units were just two units of the French CP1 series of state-of-the-art-plants already under construction in France.

Tricastin did NOT “use old technology”. Tricastin was a state-of-the-art member of the CP1 series. It was also not a “sister plant”; it was the reference plant for the nuclear portion of Koeberg (there were two more reference plants for Koeberg: Bugey for the conventional part of the plant, and Gravelines for the sea-board pumphouse).  

Alberts writes: “Thus, France laughed all the way to the bank. In terms of the cost of Koeberg, this was always a state secret that was never revealed”. Nonsense. When I asked one of my French senior colleagues how sure they were that the contract amount would cover their costs, the answer was “this is our second export. The first was to Iran and it aborted when the Shah was overthrown in 1979. We are far short there and need to go to court with Iran. This second export contract is also in a far-away country. We just hope we did not overlook something that may well make this contract a loss-maker. We are holding thumbs.” So much for laughing all the way to the bank.

As for the contract amount, it was not a state secret at all. It was in section 20.A.3 of the 180-page contract which I still have, and which every senior manager, also at Eskom, had a copy of because we needed to know it well and work according to it.

The reason it is not broadcast generally is to respect vendor-client confidentiality, a standard practice among civilised contractors and clients. This practice is followed in particular to protect the commercial interests of a vendor who may go out and sell at another price somewhere else where circumstances are different.

Since this information is now 40 years old, I might as well mention that the base price was somewhat below R1bn in 1976. It is not a round amount but a calculated amount.

During construction, the basic contract scope was added to by agreed-on changes that were either requested by Eskom or proposed by the consortium, but always agreed to by both parties.

Numerous changes came about because the French design kept evolving as they gathered experience. Eskom wished that Koeberg would stay identical to the French CP1 series in order to benefit from the fleet effect. Because of these changes and additions to the contract, and not counting Eskom’s own scope of work (roads, potable water supply to the site, staff township, site fence and security, certain buildings on site, etc) the final French portion of the contract was just shy of R1.5bn, which at the time was said to be SA’s largest contract ever.

Alberts writes: “SA had to use the French design”. Yes, after the number one shortlisted supplier disqualified himself by not providing the performance guarantee in the normal commercial manner for large projects, South Africa “took” the French design.

This turned out to be extremely fortunate because we have now got, instead of a one-off BWR, the better and more mainstream PWR, based on the original proven Westinghouse design, which was further improved by Framatome for its CP1 series, which ultimately comprised of 34 units, 32 in France and two in Koeberg.

This fleet approach meant that Koeberg, still today, profits from the lessons learned in France via the technology exchange programme, where France alerts Eskom to issues and fleetwide improvements when they come along. If I remember correctly, Koeberg’s units were numbers nine and 10 in the series.

Next is Alberts’ statement: “… when the ANC bombed the plant and set the programme back by 18 months and R500m.”

As explained above, the setback in time was 12 months. The cost of it was just over R100m (1983-84 rands). Part of this was cost to remanufacture the damaged parts, the remainder consisted of the contractor claim for having to maintain a presence of many hundreds of local and French staff and equipment at site for a year, while waiting for the re-manufacture to be completed, and then for the claim to be settled.

When it became clear that the claim should rather be settled and not be allowed to balloon further by time consuming negotiations between middle managers, the two chairpersons had breakfast together and an amicable agreement was reached between Eskom and the Consortium at top Management level.

The issue that “piping elbows had to be replaced due to lead tags, which were not removed during smelting and casting process overseas”, turned out not to be a real problem. It just produced a lot of hot air and correspondence. Some 10% of about 500 piping elbows were discovered to have discolorations that looked suspicious and needed investigation.

Before the investigation was concluded, the contractor decided to replace the installed ones and not to use the rest. In this manner, the issue barely caused a delay in the construction programme which, as explained above, was on track apart from the one-year delay caused by the ANC bombing.

The resolution of the issue turned out to be that some of the thin ferritic iron identification tags with numbers and letters stamped on them and attached to the various ingots to identify these, had not been removed when the ingots had been melted and formed into elbows.

This ferritic material thus sat in the middle of the stainless and could, in theory, become a corrosion problem. In our nuclear world we replace those things or justify a decision not to replace with extensive studies. Elsewhere, we would shrug our shoulders and carry on. Note that Alberts says they were lead tags. They were not, they were ferritic (iron) tags and the issue was known as ‘ferritic inclusions’.

Alberts said: “Operating manuals were taken from Tricastan and translated by non-technical translators, which resulted in Koeberg having to generate a humungous amount of operating procedures.”  

Half correct. Not manuals, but operating procedures were indeed taken from Tricastin because Tricastin was Koeberg’s reference plant. Manuals came directly from original manufacturers. They were indeed translated by non-technical translators, but were then passed by the French engineers in charge of the relevant systems.

That made them technically correct even while the English was not. Translation is a problem in the best of houses, which is why the country of origin is an important factor. Let us all stop to shudder to think about purchasing a Chinese plant with chinglish procedures and manuals.

So, Alberts’ statement, “thus, there is no way that any so-called expert can say with any conviction that Koeberg came in on time and under budget” is pure nonsense.

I am not a so-called expert, I am a real expert, and I can say the following – not with conviction based on belief, but with conviction from having done the admin: Apart from the clearly identifiable one year delay and extra costs caused by the ANC bombing, there was no construction delay and no over-budget spending. Koeberg was built on the same 80-month completion schedule in South Africa as the other 32 French units that were built in France.

Pebble Bed Modular Reactor: close but no cigar

Now we come to PBMR, where Alberts writes: “The second issue I would like to take umbrage with is our so-called highly competent nuclear scientists and engineers. I hope these are not the same dudes (suckers?) that went and destroyed the Pebble Bed Modular Reactor (PBMR) project.”

I know the project. I worked from 2002 till the end in 2010 at PBMR, writing essential documents. The project was destroyed not by the scientists and engineers, but by the ANC government that stopped the project because it could not distinguish between serious argument from competent engineers and nonsense dished up by greenies who enlisted funding from an overseas foundation and engaged people like Professor Steve Thomas to build a case that PBMR was economically flawed.

Economy is not the decisive factor on first-of-a-kind (FOAK) projects; the first thing is whether the concept works. Few FOAK projects are immediately economical, let alone optimal. Not doing FOAK projects because of an unproven economic picture would mean that there would not now be railways, aeroplanes, submarines, spacecraft, cellphones, motorcars or anything else.

Economic feasibility and optimality are only sorted out after the first prototypes show it can work. Stopping this very promising project (a 400 MWth design was complete and the 200 MWth design was at 80% when the project was killed) means that SA went from frontrunner in the Small Modular Reactor (SMR) stakes to non-participant.

Other nations now run with our technology, which will flood the world in 20 years from now when the market for small units in remote places and/or off-grid opens up. The engineers who became redundant after PBMR was killed have all found immediate employment in other industries in SA and overseas. If they were only “so-called” competent they would not be so employable and highly sought after.

Alberts is wrong when he says: “Literally billions were spent on a project driven by nuclear scientists, engineers and good old Eskom as the Project Office. We designed and redesigned and redesigned until we had designed ourselves out of existence. We placed orders and paid (in full) for components from Spain, Germany, etc and to this day nothing has ever been delivered. The project is in tatters and the intellectual capital has taken flight.”

Design proceeds in back-and-forth steps. There was only one major redesign, started around 2006, to the step-down from 400 to 200 MWth. It was caused by Eskom retracting its letter of intent to purchase the first set of modules.

PBMR was then without a client, looked at the world market for SMRs and chose 200 MWth, the size most in demand. Eskom withdrew because it had been weakened in many respects. It had suffered a capital shortfall by 2006 by selling its electricity since 1990 with annual tariff increases that were consistently below the inflation rate in an effort to be the world’s cheapest.

The government had blocked Eskom in 1997 from starting on the next power station that was needed by 2007, which meant that Eskom was facing expensive oil-based replacement power. Eskom had lost many of its competent engineers since 1994 due to transformation, and had to outsource its maintenance.

By 2006, Eskom was close to running on empty and could no longer afford to purchase PBMR modules. So PBMR had to redesign. And the intellectual capital did not take flight; it was chased from the premises.

I do not know whether PBMR ordered anything from Germany, but the reactor vessel was reportedly delivered from Spain and lies forlorn in the Richards Bay Harbour. Indeed, the project is in tatters. We can thank our ignorant politicians and stupid anti-nuclear activists for this destruction of capital and effort.

Fukushima: False radiation fears to blame

Let us now turn to Alberts’ paragraphs on Medupi. I will not address the remarks about the Koeberg weld debacle, uranium leakage, nuclear waste, but his quotation from a physician about Fukushima merits response.

It says: “Steven Starr in Physicians for Social Responsibility reported that the estimated total economic loss from the Fukushima incident ranges from $250bn to $500bn. As for the human costs, in September 2012 Fukushima officials stated that 159 128 people had been evicted from the exclusion zones, losing their homes and virtually all their possessions.”  

It has become customary in the press not to distinguish between:

a) tsunami victims that drowned - approx 19 000    

b) tsunami victims that survived, but lost homes and were put up in shelters - approx 200 000    

c) nuclear victims that died due to radiation - zero    

d) nuclear victims evacuated from exclusion zone and put up in shelters - approx 90 000    

The numbers are not precise, but these are the orders of magnitude. Here (b)+(d) together was reported as 290 000 (ref below), hence (b) is around 200 000

READ: Nearly 290 000 people still living in shelters 2 1/2 years after Tohoku disaster

I cannot comment on a precise count like 159 128 and the billion dollar damage amounts. Every estimate and every study   and there are many   differs from the next one, showing that the bookkeeping of chaotic situations is essentially impossible. It is also regularly reported that about 1600 people died from disaster-related stress following the living in shelters, lost health care, etc. Since (d) is nearly one-third of (b)+(d), the stress-related extra deaths due to the nuclear evacuation (d) is around 500.  

Category (d) carries the label “nuclear damage”. We now get to the crux of the matter. It is NOT nuclear damage, but sociological damage. These 90 000 evacuations should not have taken place at all. They are forced on people because of wrong legal regulations that declare totally innocuous radiation levels as dangerous. Even today, one still reads of the “highly contaminated” areas. There were none, repeat none.

The normal background level is 3.6 mSv/year. Johannesburg at high altitude has about 5 mSv/y. In Brazil and Iran, there are areas where people have lived since history began in areas with respectively 180 mSv/y and 260 mSv/y. Workers in nuclear plants are allowed 20 mSv/y (it used to be 50).

No effect has ever been detected in anybody below 100 mSv/y. The first effects start to be noticed at 1000 mSv as an acute dose, say in a short period like one day or less. If 1000 mSv are spread out over a year then, there is no effect because the body copes with it.

By 2000 mSv, a single dose to the whole body serious illness occurs, and a 4-5000 mSv whole body single dose is lethal. The dose delivered during medical irradiations of tumors is of the order of 50 000 to 100 000 mSv per occasion, but to the tumor only (the rest of the body is screened) and after enough repeats it kills the tumor.

The nuclear industry is saddled with regulations that prescribe a dose-rate limit to the population of 1 mSv/y (above background), which is 1/20th of what plant workers are allowed, and which is less than 1/1000th of the level where there are still no effects.

If one makes a comparison with rain, it would be like this: the public is allowed to go out in a drizzle that does not yet show up as a wet street, but is not allowed to walk outside when the street starts to show a wet street. Plant workers are allowed to be outside when it rains a bit stronger, but not when it pours. And even pouring rain harms nobody. It becomes lethal when it hails.  

The so-called “highly contaminated areas” referred to around Fukushima were all of the order of 1 to 10 mSv/y above the background of 3.6 mSv/y. The law   based on recommendations dating back 70 years when people did not yet know the effects and were overly cautious   says that people have to be evacuated at 1 mSv/y above background.

So obeying this law has added another 500 unnecessary deaths to the already high number. The death toll is thus not nuclear-caused, it is the result of wrong over-zealous legislation that in turn is the result of exaggerated fear ingrained in ignorant populations to the extent that proposals to return to rational legislation keep falling on deaf ears. So, people die for insisting on being safe.

Alberts said: “Will highly competent scientists, engineers and government bear the brunt of such an incident? I think not, but the public will.” 

How ironic. Real radiation scientists have known for a very long time that the radiation limits are set at the wrong level and cause more harm than good. There is much peer-reviewed literature on the issue. But they have not managed to convince ignorant lawmakers to stop this nonsense. Besides, there is a trillion dollar radiation protection industry that has great interest in perpetuating the status quo.

Alberts should rather ask “Will incompetent officialdom, lawmakers and government bear the brunt of such an incident? I think not, but the public will.” Yes, indeed, the public does bear the brunt of its own ignorance.

ALSO READ: Nuclear site: Fin24 community responds to Dr Kemm

ADD YOUR VOICE: Send your views now.

* Alberts has been notified of this rebuttal and has been given an opportunity to respond.

Disclaimer: All articles and letters published on Fin24 have been independently written by members of the Fin24 community. The views of users published on Fin24 are therefore their own and do not necessarily represent those of Fin24.

Brent Crude
All Share
Top 40
Financial 15
Industrial 25
Resource 10
All JSE data delayed by at least 15 minutes morningstar logo
Company Snapshot
Voting Booth
Please select an option Oops! Something went wrong, please try again later.
I'm not really directly affected
18% - 1590 votes
I am taking a hit, but should be able to recover in the next year
23% - 2079 votes
My finances have been devastated
34% - 3111 votes
It's still too early to know what the full effect will be
25% - 2250 votes