The next major nuclear accident

When in March 2011 reactors at the nuclear power plant of Fukushima Daiichi exploded one after another, TV watchers were not the only ones to be in shock. Nuclear experts, too, asked themselves: why again so soon? It had been only 25 years since a reactor at Tchernobyl had spit huge amounts of radioactive matter into the atmosphere. According to calculations, nuclear catastrophes should not occur so often. Where was the mistake?

What is the likelihood for a grave incident to occur at a nuclear power plant? – this is what nuclear experts all over the world have been trying to determine with help of a method called “Probabilistic Safety Analysis,” in short: PSA.

To do this, they start out small: what are the chances that vent A will get blocked in a power plant? What are the chances then, that the pipe in front of it will burst? Or will the water simply flow through Vent B? And what if this vent too becomes blocked?

Whole trees of events are created in this way, describing every possible damage. You end up with a number: once in x years in this particular power plant a damage will occur to the nuclear reactor. 200.000 years is not an unusual resulting value.

But considering that there are no less than 500 reactors in the world, according to these calculations, such a nuclear core damage should occur only every few hundred years. Not every 25 years.

Many risk experts are not surprised. From their point of view, it is logical that with PSA, chances of an incident are underestimated – simply because it is not possible to think about everything that might possibly happen. “It is not surprising that Probabilistic Safety Analyses did not foresee a series of incidents in the history of the civil use of nuclear energy. It has been shown that they are often based on unrealistic assumptions ", Spencer Wheatley, Benjamin Sovacool and Didier Sornette write in the specialized review Risk Analysis.

The three risk researchers from University of Sussex and ETH Zürich have therefore applied a different method, in order to find out when the next big incident will be likely to happen: they have examined how often incidents have occurred in the past, how serious they were, and how things have developed over time.

With this approach, they bumped into a problem right from the beginning: the difficulty to get at data from incidents. Neither the International Atomic Energy Administration (IAEA), nor any other international organization, is making a complete record of past incidents available, together with with trustworthy indications about the damages which resulted.

"The operators are certainly holding any amount of data about past incidents, but they keep them secret,” Sovacool says. With organizations such as IAEA, he sees a conflict of interests: "The task of IAEA is, among others, to promote nuclear energy. This throws a shadow over its role as a controlling organ."

The IAEA asserts that operators and controllers exchange over incidents: by means of the International Report System for operations (IRS). There, “reports about unusual incident which are relevant to security are collected, " according to IAEA spokesperson Susanna Lööf. The reports are purportedly analysed, and the findings about safety problems are transmitted to representatives of the member countries, of the nuclear industry, or of relevant scientific research institutes.

However, this databank is not public: “The IRS data may contain industrial secrets. They could also be used by third parties for planning disruptive actions, " these are the motives given for the secrecy by Anja Lutz from the German federal administration for the protection from radiation (BfS).

But the risk experts are not in the least interested in industrial secrets. They would be just content with just a complete listing of the incidents, with reliable indications about the amount of damage.

And indeed there exists at the IAEA the so-called INES-Scale. Attributing evaluations between zero and seven, it indicates the incident’s level of seriousness. The IAEA has introduced the INES-Scale precisely in order to be able to easily express the seriousness of an incident.

But the IAEA doesn’t have such high confidence in its own scale, either: "Considering that INES was conceived for fast communication response and for the safety relevance of incidents, the INES values have only relatively limited relevance for other purposes," says IAEA-Spokesperson Lööf. INES-values for past incidents can be called up on the IAEA-Website only for the past twelve months.

This is too little to conduct empiric analysis. Wheatley, Sovacool and Sornette have therefore elaborated their own data-bank of incidents. As they have had to rely, among others, on newspaper articles, their information is at a relatively hight risk of error.

Nevertheless, the three researchers are convinced: the secrecy-mongering of the nuclear economy and of its controlling administrations lead to an excessive trust into the safety of nuclear power plants, because there is no overview over all that is actually going wrong. This perception does not least influence political decisions. Wheatley reaches a strikingly different conclusion: “The level of risk of nuclear energy is according to our analysis extremely high.”

Now this, in turn, nuclear experts won’t let stand unchallenged. They criticize the three risk researchers’ approach of merely studying the frequency of incidents for being like comparing apples and pears: nuclear power plants, after all, are becoming ever safer, their argument is running; therefore incidents of the past don’t give any indication about the future.

Moreover, the installations of different countries differ vastly from each other. Anja Lutz finds the approach of Wheatley, Sovacool and Sornette therefore “questionable.”

But the insistence of the nuclear branch of holding on to the PSA-data is just as questionable. After all, these do not chime with the distance of only a few years separating the incidents in Tchernobyl and Fukushima. Was it truly only a mere coincidence?

At least the Cologne based GRS Company for the safety of installations and reactors is looking critically at Probabilistic Safety Analysis. After the MCA (Maximum Credible Incident) at Fukushima,  GRS scientists have had a look at what had gone wrong according to the PSA for Fukushima. In their 2015 study, they arrived at the conclusion: “The existing PSA for nuclear power plants do not sufficiently take into consideration rare events and their combination."

The methods of PSA must be revised, and this is not becoming any easier. "PSA-scientists are seeing themselves confronted with a series of complex and difficult problems."

This is precisely the result to which Wheatley, Sovacool and Sornette have also reached: incidents of medium importance have over time become significantly rarer; Probabilistic Safety Analysis has succeeded, from all appearances, in identifying weak spots in reactors, and to correct them as a consequence.

But the likelihood of catastrophes such as those of Tchernobyl and Fukushima continues to be systematically underrated. Considering that 65 reactors are presently at the building stage all over the world, this is not good news. 

Maximilian Schäfer

Der Spiegel, July 6, 2016

Translated from the German by Anne-Marie de Grazia

Go to the original article

(March 27, 2016)

A security guard who worked at a Belgian nuclear medical research facility was murdered two days after the Brussels bombings, it emerged yesterday (Saturday), deepening fears that Islamist terror cells are plotting attacks against nuclear installations.

Didier Prospero, a guard with the G4S security company, was shot dead at his home in the Froidchapelle district of Brussels on less than 24 hours after Belgian authorities stripped several workers of their security passes at two nuclear plants this week.

The circumstances of 45-year-old Mr Prospero’s death remained murky last night, with conflicting reports over whether or not the murder was linked to terrorism, or if his work security pass had been stolen.

News of the killing emerged as Belgian prosecutors announced on Saturday they have charged three men with terror offenses over the suicide attacks, as organisers cancelled a solidarity rally at the government's request because police are too stretched to cope.

[...]

...Mr Prospero was found dead in his bathroom by his three children when they returned home from school on Thursday afternoon. He had received four gunshot wounds. His sheepdog Beauce was also killed and lay next to him.

Belgian prosecutors maintained last night that the murder was the result of a burglary gone wrong, but provided no detailed explanation as to why Mr Prospero, a respectable family man, should suddenly have been murdered in a freak burglary.

The killing comes after a string of security scares and breaches around Belgium’s nuclear infrastructure and the discovery last November an Islamic State cell in Brussels had kept a top Belgian nuclear scientist under video surveillance.

The report of Mr Prospero’s murder heightened concerns that the Brussels bombers were plotting to build a radioactive “dirty” bomb — but apparently shelved the plan after security was stepped up at Belgium’s nuclear plants this month following intelligence warnings.

Belgian authorities have played down the risk posed by jihadists to its nuclear facilities in the past.

Last November 10 hours of surveillance footage of a top Belgian nuclear scientist was discovered in a house belonging to a known jihadi, but the existence of the footage was only acknowledged by Belgian authorities on February 18 after it was leaked to a Belgian paper.

The film is believed by security forces to have been taken by Ibrahim and Khalid el-Bakraoui, the brothers who the authorities say were suicide bombers at the Brussels airport and subway station. They are understood to have removed a hidden camera from bushes outside the official’s house.

After the news broke Jan Jambon, the Belgian interior minister, rejected a proposal to deploy troops saying that “nothing indicates a specific threat to nuclear power plants”, but two weeks later, on March 4, changed his mind and deployed 140 soldiers to guard five nuclear facilities.

This is not the first time that fears of a terror threat to Belgium's nuclear power plants have been raised.

In 2013, an engineer from Doel 4, one of the nuclear reactors of a power plant near Anvers, was sacked over concerns that he had been radicalised after he refused to shake his superior's hand. The employee was later identified as the brother-in-law of Azzedine Kbir Bounekoub, a jihadist involved with Sharia4Belgium, who left Belgium to join Isil in Syria in 2012 and had frequently called on Isil sympathisers to launch terror attacks in Belgium.

In the hours following the Brussels bombings, two nuclear power plants were evacuated.

In another disturbing incident, a turbine at the same Doel 4 reactor was sabotaged in 2014. Someone deliberately turned security cameras the other way and then emptied 65,000 litres of oil used to lubricate the turbine. "Then they put the lid back on to make everyone think all was well," according to Eloi Glorieux, a nuclear expert at Greenpeace Belgium.

The incident has never been elucidated and there have been no arrests. The federal prosecutor is "seriously considering" the theory that was linked to terrorism, according to the French newspaper Libération.

The nuclear plant in question is within range of a string of petrochemical plants and one of the most densely populated regions in Europe, with 1.5 million people in a 30km range.

"This sabotage could have sparked a true catastrophe," said Jean-Marc Nollet, Belgian MP and head of the Greens' parliamentary group. "We were lucky in a way."

Jean-Marc Pirotton, union representative at the Tihange nuclear plant said that the plants were very well-protected, there were vulnerabilities. "Belgium nuclear plants were not designed to withstand a big plane crashing into them, only the planes around at the time they were built," he told Libération newspaper.

Protection of the site is run by external security contractors, whose agents are unarmed and not allowed to arrest suspects - only to alert police; Half of the plants' staff are external and their security vetting, according to unions, is not nearly as thorough as those working for Electrobel, Belgium's nuclear operator.

The country's creaking nuclear plants and seven reactors have been causing safety concerns with neighbouring countries after a series of problems ranging from leaks to cracks, as well as the reported sabotage.

Despite the concerns, the Belgian parliament voted last year to prolong their life for another ten years.