Severe Accidents: Singularity of Nuclear Disasters?

Abstract

The risk of nuclear accidents has proven to be low in absolute and relative terms. Nevertheless, the high-energy density fission process and current technology make today’s reactors vulnerable to very severe albeit very rare accident scenarios. The three disasters experienced demonstrated the importance of site conditions, containment systems and severe accident management measures. Not being triggered by a single or combined technical failure in a classical sense, disasters experienced are partly explained as a product of five hierarchical levels of individual and societal human factors. The potentially severe consequences, including costs, of nuclear accidents have played a decisive role in the development of the nuclear power sector, and dominate nuclear risk analysis. However, in terms of cost or loss of life, these accidents are not singular—as other industrial and energy generation sectors have comparable severe accidents.

Despite this, there is a widespread and exceptional human dread of low level radiation exposure that prevents us from facing the real issue of how to improve world prosperity while burning less fossil fuel. To deal with this pragmatically, and operating on the principle that “nuclear power plant safety requires a continuing quest for gain in excellence”, we identify enhanced requirements to take the dread out of nuclear, and to rely less on social stability and long term husbandry of wastes.

Annexes

5.1.1 A.1 The International Nuclear and Radiological Event Scale (INES)

The INES was introduced in 1990 by the International Atomic Agency (IAEA) in order to enable prompt communication of safety-significant information in case of nuclear accidents, and was refined in 1992 and extended to be applicable to any event associated with radioactive material and/or radiation, including the transport of radioactive material. The most recent user manual was published in 2008.⁴⁸ The scale is intended to be logarithmic, similar to the moment magnitude scale that is used to describe the sizes of earthquakes. Each increasing level represents an event approximately ten times more severe than the previous level. Compared to earthquakes, where the event energy can be quantitatively evaluated, the level of severity of a man-made disaster, such as a nuclear accident, is more subject to interpretation (Fig. 5.4).

Fig. 5.4

International Nuclear Event Scale, standard pyramid representation of the 8 levels (INES, IAEA. “The international nuclear and radiological event scale user’s manual 2008 edition.” IAEA and OECD/NEA (2008))

The level of the scale is determined by the highest of scores from three tracks: off-site effects (also called People and Environment), on-site effects (also called Radiological Barriers and Controls), and defense in depth degradation (potential consequences). Effects include health and environmental impact; costs are not considered explicitly. Below, real events are given as examples for INES levels 3–7, scored according to the relevant track.

5.1.1.1 A.1.1 Level 7: Major Accident—Impact on People and Environment

Major release of radioactive material with widespread health and environmental effects requiring implementation of planned and extended countermeasures. There have been two such events to date:

• Chernobyl disaster (Ukraine/former Soviet Union), 26 April 1986. A power surge during a test procedure resulted in a criticality accident, leading to the release of a significant fraction of core material into the environment. See also Sect. 3.4.2.

• Fukushima Daiichi nuclear disaster, a series of events beginning on 11 March 2011, Japan. Major damage to the backup power and containment system caused by the Tohoku submarine earthquake and induced tsunamis resulted in overheating and leaking from some of the nuclear plant’s reactors. Each reactor accident was rated separately; out of the six reactors, three were rated level 5, one was rated level 3, and the situation as a whole was rated level 7. See also Sect. 3.4.3.

5.1.1.2 A.1.2 Level 6: Serious Accident—Impact on People and Environment

Significant release of radioactive material likely to require implementation of planned counter-measures.

• Kyshtym disaster at Mayak Chemical Combine (former Soviet Union), 29 September 1957. A failed cooling system at a military nuclear waste reprocessing facility caused a steam explosion with a force equivalent to 70–100 tonnes (metric) of TNT. About 70–80 metric tons of highly radioactive materials were carried into the surrounding environment. The impact on local population is not fully known, but at least 22 villages were affected with doses that may lead to shortened life expectancy due to higher risk of cancer.⁴⁹

5.1.1.3 A.1.3 Level 5: Accident with Wider Consequences

• Impact on people and environment

  Limited release of radioactive material likely to require implementation of some planned countermeasures. And/or several deaths likely to result from radiation.

• Impact on radiological barriers and control

  Severe damage to reactor core (more than 1% of inventory); release of large quantities of radioactive material with a high probability of significant public exposure. This could arise from a major criticality accident or fire.

  • Windscale fire (UK), 10 October 1957. Annealing of graphite moderator at a military air-cooled reactor caused the graphite and the metallic uranium fuel to catch fire, releasing radioactive pile material as dust into the environment.

  • Three Mile Island accident near Harrisburg (USA), 28 March 1979. A combination of design and operator errors caused a gradual loss of coolant, leading to a partial meltdown. See also Sect. 3.4.1.

  • First Chalk River accident (Canada), 12 December 1952. Reactor core damaged.

  • Lucens (INES level 4–5) partial core meltdown accident (Switzerland), 21 January 1969. A test reactor build in an underground cavern suffered a loss-of-coolant accident during a startup, leading to a partial core meltdown and massive radioactive contamination of the cavern, which was then sealed.

  • Goiânia accident (Brazil), 13 September 1987. An unsecured cesium chloride radiation source left in an abandoned hospital was recovered by scavenger thieves unaware of its nature and sold at a scrapyard. 249 people were contaminated and 4 died.

5.1.1.4 A.1.4 Level 4: Accident with Local Consequences

• Impact on people and environment

  Minor release of radioactive material unlikely to result in implementation of planned countermeasures other than local food controls. At least one premature death from radiation.

• Impact on radiological barriers and control

  Fuel melt or damage to fuel resulting in more than 0.1% release of core inventory. Release of significant quantities of radioactive material within an installation with a high probability of significant public exposure.

  • SL-1 Experimental Power Station (USA), 1961, reactor reached prompt critically, killing three operators.

  • Saint-Laurent Nuclear Power Plant (France), 1969, partial core meltdown; 1980, graphite overheating.

  • Buenos Aires (Argentina), 1983, criticality accident on research reactor RA-2 during fuel rod rearrangement killed one operator and injured two others.

  • Tokai-mura nuclear accident (Japan), 1999, three inexperienced operators at a reprocessing facility caused a criticality accident; two of them died.

5.1.1.5 A.1.5 Level 3: Serious Incident

• Impact on people and environment

  Exposure in excess of ten times the statutory annual limit for workers; non-lethal deterministic health effect (e.g. burns) from radiation.

• Impact on radiological barriers and control

  Exposure rates of more than 1 Sv/h in an operating area. Severe contamination in an area not expected by design, but with a low probability of significant public exposure.

• Impact on defense-in-depth

  Near-accident at a nuclear power plant with no safety provisions remaining. Lost or stolen highly radioactive sealed source. Highly radioactive source in the wrong or unexpected place without adequate procedures in place to handle it.

  • THORP plant, Sellafield (UK), 2005; leak of uranium and plutonium dissolved in nitric acid into stainless steel containment structure.

  • Paks Nuclear Power Plant (Hungary), 2003; fuel rod damage in cleaning tank

  • Vandellos Nuclear Power Plant (Spain), 1989; fire destroyed many control systems; the reactor was shut down safely.

  • Davis-Besse Nuclear Power Station (USA), 2002; negligent inspections failed to record corrosion through 15.24 cm of the carbon steel reactor head leaving only 9.5 mm of stainless steel cladding intact that was holding back the high coolant pressure of 17 Mpa.

    Table 5.5 Radioactive releases compared (numerical values in TeraBecquerel = 10¹² Bq, % of core inventory) for Chernobyl^a, Fukushima^b, TMI^c, and Windscale^d. Empty values mean nothing measured

5.1.2 A.2 Breakdown of Rough Integral Cost Estimates

5.1.2.1 A.2.1 On-site

Emergency management, mitigation, capital loss, and remediation/clean-up.

TMI

Unit 2 was destroyed within 1 year of operation, having cost USD 2B (Billion) to build, 7 years of downtime were caused at unit 1,⁵⁰ and substantial contamination led to nearly USD 2.5B in cleanup.⁵¹ We find USD 5-10B to be reasonable, although higher estimates exits.⁵²

Fukushima

The six-unit site, taken into operation in the 1970s, was lost. Shutting down the near 50 reactor fleet for 7 years already led to further loss of capital. As of the end of 2017, 5 reactors have restarted and 21 are currently subject to approval to restart. TEPCO estimates for cleanup and decommissioning costs are USD 10-20B.⁵³ Thus, overall USD 20-30B is reasonable, but the more than 30 years planned for decommissioning, and ample uncertainty, leaves room for unforeseen costs.

Chernobyl

In 1986 and 1987, some 440,000 recovery operation workers were employed at the Chernobyl site.⁵⁴ The Ukrainian government claims USD 20–30B on accident management and decontamination so far.⁵⁵ Unit 4, less than 4 years into its life, was lost. Despite contamination, units 1–3 (each producing about 6 TWh/year) resumed operation by the end of 1987. The new confinement structure, built over the remains of Unit 4, cost 2B USD. Also allowing for further future costs, the total on-site cost is USD 25-35B.

5.1.2.2 A.2.2 Life and Health

Health impact and trauma due to radiological exposure, emergency response, relocation, continued displacement, and so on. Details on late cancer fatalities in Sect. 4.3.

TMI

Despite stress due to media and poor communication,⁵⁶ human consequences were relatively negligible due to successful containment. There was a minor release (Table 5.5) from the plant’s auxiliary building to relieve pressure on the primary system, such that less than 1 additional death from cancer was expected.⁵⁷ Total payout for a class-action settlement was around USD 0.1B.⁵⁸

Fukushima

The Tōhoku earthquake and tsunami killed over 18,500 people from effects unrelated to destruction of the reactors at Fukushima Dai-ichi. Long-term displacement of more than 164,000 people, attributed to the nuclear accident is blamed for around 2000 early deaths,⁵⁹ largely effecting the elderly (90% older than 66), costing USD 2–3B⁶⁰ for years of life lost (YOLL). Attested to be of lower order are the long term expected radiological fatalities: Linear no-threshold (LNT)-based estimation allows for on the order of 100 eventual fatalities due to cancer, and therefore an expected YOLL of >1000, and a cost of about USD 1B.⁶¹ Traumatic effects (suffering, isolation, etc.) are more difficult to estimate, with >100,000 people still displaced as of 2016, and increased health and psychological issues identified. Compensation of approx. 10 thousand USD per person per year displaced were specified⁶², gives about USD 11B. This impact is the subject of overall compensation, covered in the following category. Overall here: USD 3–4B + 11B (“trauma”).

Chernobyl

The human impact of the major accident at Chernobyl was immense, and its assessment remains contentious. As discussed in Sect. 4.3, a range of 4000–16,000 eventual radiological fatalities have been put forth in serious studies—giving USD 3.6–11B for YOLL. Chernobyl led to more than 220,000 people being relocated. As for Fukushima, allowing 10,000 USD per person per year displaced, here limited to a 10-year period, gives 22B. Overall: USD 3.6–11B + 22B (“trauma”).

5.1.2.3 A.2.3 Public Economic

Economic costs suffered by the public (incl. loss of property and income).

TMI

Negligible in relative terms, and potentially already covered within the USD 0.1B payout.

Fukushima

Compensation liabilities are currently estimated at USD 50–100 Billion⁶³, also intended to cover life and health impacts due to displacement (above). TEPCO is expected to pass these costs onto consumers in the future with higher rates. The once 1100 square km evacuation area is already less than half of its original size⁶⁴ as Cs-137 decays and is transported by rainfall to the sea.

Chernobyl

A 2600 square km exclusion zone formed, the majority of which is now safe for settlement and economic activity. Total losses (not limited to this category) reported for Ukraine and Belarus are USD 210–224B. For this category, in Ukraine^65,66, there have been about USD 18B in capital losses, USD 20B in social costs (to date), and economic losses due to loss of natural resources and agriculture of USD 70B, totaling about USD 110B. A rough estimate to cover also Belarus and Russia is then USD 150-250B.

5.1.2.4 A.2.4 Replacement Power

The increased cost of provision of electricity, incurred by the utility, but also by the public, through higher electricity prices. According to the NRC⁶⁷, for a generic NPP unit being shut down, the replacement power costs are estimated to be approximately USD 17B, and 40B for a new unit in a pool with above average replacement energy costs—thus posing a major financial risk.

TMI

Reliance on nuclear power to meet demand was high, and the accident triggered a financial crisis for the utility due to high-cost (primarily oil generated) replacement power. The cost of this replacement was later allowed to be passed onto customers through a near doubling of the rate.⁶⁸ The resulting increment in cost of electricity was up to USD 0.9B per year, for up to the 6 years of downtime of both units.⁶⁹ On this basis, a range of USD 5–15B can be taken, with the upper limit more in line with NRC guidelines.

Fukushima

Suspending Japanese nuclear power (250 TWh/year, 30% of annual generation) required importing natural gas at an increased cost of USD 20B (1.9 trillion Yen⁷⁰) per year (excl. potential health impacts). With the uncertainty about when the NPPs will be turned back on, the cost is USD 100B–150B.

Chernobyl

There was sufficient need for power that operation at the (highly contaminated) site resumed within the year. The Ukrainian government claims that the Chernobyl accident led to a loss of 62 thousand TWh, much of which was not replaceable, resulting in a larger economic loss of USD 28B, giving a range of USD 10-30B.

5.1.2.5 A.2.5 Industry Response: A Cost or a Benefit?

The nuclear sector is somewhat unique in the extent to which major accidents drive increases in regulatory rigor worldwide, and costly retrofits. On the other hand, it should make the plants safer and potentially improve operational efficiency.

TMI

The “TMI Action Plan”⁷¹ retrofits implemented substantial improvements in operations, plant design, and regulation.⁷² The retrofits in the US alone were roughly estimated to be between 10 and 60 Billion USD—i.e., between 100 and 600 million USD per unit.⁵³ The costs of delays and retrofits to the >50 operating US units and 40 units under construction, were estimated at more than USD 90B.^73,74 This indicates a total (global) cost in the range of USD 100–200B.

Fukushima

This major accident strengthened scrutiny of nuclear power, triggering many safety initiatives⁷⁵ including the European “Stress Test”, the IAEA “Vienna Declaration”, and other activities.⁷⁶ The retrofits are estimated to cost USD 60B in Japan, USA, and France alone⁷⁷, resulting in a similar figure of USD 0.2B per reactor from Japan.⁷⁸ A Swiss unit (KKM) was required to undertake > USD 0.4B of retrofitting, causing the plant to close. Extrapolating the 0.15–0.3 B globally gives a total of about USD 60–120B.

Chernobyl

Major retrofits to all 19 RBMK reactors were required. The cost for this is unknown but, within a factor of two of the 0.2B number from Fukushima, provides a range of USD 2–8B.

5.1.2.6 A.2.6 Beyond/Limitations

Deeply uncertain costs and potential benefits.

TMI

this accident caused an inflection point in the industry, killing growth of the nuclear industry and leading to a new regulatory-economic regime.⁷⁹ In this regard, TMI may have been the most influential nuclear accident of all.

Chernobyl

As stated by the IRSN⁸⁰, such an event can durably stun/affect a nation and an economy. It is difficult to be convinced that a cost estimate is complete in such a case. In particular, the accident may well have contributed to the downfall of the USSR.⁸¹ It also galvanized the “environmentalist movement” against nuclear power. Other things excluded are the resulting suspension of the Belorussian nuclear program⁸², and contributions to the shutdown of former USSR units.

Fukushima

This accident renewed and strengthened political scrutiny of nuclear power. In particular, excluded above are the potential costs of the German shutdown of 40% of the German nuclear power capacity since 2011.⁸³ These costs are particularly difficult to estimate due to future uncertainties about the potential performance of future nuclear and new renewables. Official cost estimates range from tens to hundreds of billions of Euros⁸⁴, with rates expected to briefly peak at about 50% above baseline.⁸⁵