Monday, 20 May 2013
masthead+quote+image
Advanced search

Japan earthquake shuts nuclear facilities

The massive earthquake that struck off Japan’s northeast coast today has triggered the automatic shutdown of four nuclear power stations, with reports of a non-reactor fire in one and problems with the cooling system in another.

However, as yet there has been no reported leakage of radiation from any of the reactors and at this stage experts are not predicting any major discharge that would be hazardous to human health.

According to the International Atomic Energy Agency (IAEA) the Fukushima Daiichi, Fukushima Daini, Tokai and Onagawa nuclear power stations all automatically shut down soon after the 8.8 magnitude quake hit at 2.45pm local time. A fire broke out in the turbine building of the Onagawa plant while the Fukushima Daiichi plant reported problems with its cooling system.

Nuclear power plants have a series of strategically-placed sensors that detect peak ground acceleration (PGA) — which describes surface movement at particular places, rather than the cruder measure of the magnitude of the earthquake itself. Over a certain PGA threshold the sensors will trigger an automatic shutdown of the reactors.

‘Although this earthquake was massive, it was some way off-shore and the levels of shaking on land should not have been as high as in 2007, where the earthquake was smaller in magnitude but directly below the nuclear plant itself,’ said Julian Bommer, Professor of earthquake risk assessment at Imperial College London.

Geohazards are a key criteria when citing a nuclear power stations — plants are very unlikely to be built on soft or unstable soil. Then there will be a very detailed assessment of the potential seismic hazard — in other words, the probability of different levels of ground shaking at the site

Prof Julian Bommer, Imperial College London

In 2007 a 6.8 magnitude earthquake struck just 16km from the world’s largest nuclear power plant — Tepco’s Kashiwazaki-Kariwa facility. This caused a fire on an electric transformer and the leakage of 1.5 litres of mildly radioactive water, but the reactor was ’absolutely unscathed,’ according to Bommer.

Based on this event, Japan’s Ministry of Economy Trade & Industry (METI) set up a 20-member Chuetsu Investigation and Countermeasures Committee to investigate the impact of earthquakes on power stations and potential counter-measures.

‘Geohazards are a key criteria when citing a nuclear power stations — plants are very unlikely to be built on soft or unstable soil. Then there will be a very detailed assessment of the potential seismic hazard — in other words, the probability of different levels of ground shaking at the site,’ said Bommer.

He explained that nuclear reactors employ the same sort of engineering techniques and designs used in other structures that are built to withstand earthquakes — but with considerably higher tolerances built-in.

‘The starting point would be a very high ground acceleration with a probability of around one in 10,000. Then the reactor will be designed for that, but with a large safety margin — so even if that ground acceleration occurred, the probability of there actually being any elastic behaviour in the structure is about one in a 100,000. And because there are further redundancies in the system, the probability of going from elastic behaviour to release of radioactivity from the core is more like one in a million,’ he said.

Readers' comments (6)

  • The probability of occurrence is one every 10000 years (the article cites 1 in 10000, but doesn't specify "years", which is important). Secondly, the article states that the probability of elastic behavior is about one in 100,000. The author probably meant to say "plastic", not "elastic". Nuclear category I seismic structures are designed so that, in a 1 in 10000 year seismic event, the structure's response remains in the elastic range. Although, even though the structure is design to remain elastic, it is designed and detailed so that is capable of performing well in the plastic range. This assures that, in a beyond design basis event (beyond 1 in 10000 year seismic event), the structure will be capable of absorbing this additional energy and will achieve collapse prevention/life safety functions.

    Unsuitable or offensive? Report this comment

  • If nuclear facilities are designed to withstand a 1 in a million year event, does this mean that if 500 facilities are in existence then the average expectation would be for a major disaster every 200 years?

    Unsuitable or offensive? Report this comment

  • I understand earthquakes top the list of geohazards considered in design of nuke facilities. The chances of one occurring whether in years, days, or minutes, who knows what, may not be confused with the measures of severity that can affect plant reliability.

    Unsuitable or offensive? Report this comment

  • In response to Anon.
    No it doesn't work that way, if you had an event big enough to cause these power stations to fail then it would likely cause all or a large number to fail, however that size event wouldn't happen again for 10,000 years, meaning that an identical power station built the following day would be safe for that time period (well beyond it's operational and decommissioning life), however "on average" 1 station failed every 200 years, but in fact it was 500 all at one time. Otherwise you could argue that on average 800,000 people died in wars last year, technically this is true if you consider 80million people died in WW1 & WW2, so teh average per year over the last 100 years is 800,000 per year, but that doesn't mean it happened, be careful with statistics they can be misleading.

    Unsuitable or offensive? Report this comment

  • Yes the UK civil nuke facilities are designed to survive a 1 in 10000yr seismic event. However, this 1/10000 frequency doesn't line up with the requirements for reactor lines of protection which are set at 1/1000. Can anyone advise me how we arrived at the seismic hazard needing to be assessed at 1/10000 occurrence?

    Unsuitable or offensive? Report this comment

  • Yes the UK civil nuke facilities are designed to survive a 1 in 10000yr seismic event. However, this 1/10000 frequency doesn't line up with the requirements for reactor lines of protection which are set at 1/1000. Can anyone advise me how we arrived at the seismic hazard needing to be assessed at 1/10000 occurrence?

    Unsuitable or offensive? Report this comment

Have your say

Mandatory
Mandatory
Mandatory
Mandatory

My saved stories (Empty)

You have no saved stories

Save this article

Digital Edition

The Engineer May Digital Edition

Poll

Forward-looking flying car specialist Terrafugia has unveiled a new autopilot-equipped STOVL concept which it says could be on sale in 8-12 years. But will the science-fiction staple of the flying car ever take off?

Previous Poll

NASA chief Charles Bolden says that the agency is moving forwards with plans for a manned Mars mission, but there are significant ‘technology gaps’. Which of these is likely to be the most difficult to overcome?

Read and comment on the results here