fredag 13 maj 2011

Underestimation of tsunami risk at Fukushima

A German report (see below) suggests that the reason for the Fukushima accident was an underestimate of the risk that an earthquake should be followed by a huge tsunami. I submit the question to the forum if the risk was much underestimated before the accident and thus better preparations should have been done. If so, the next question I submit to the forum is what changes will make similar mistakes less likely in the future.

Fukushima was designed to resist a tsunami up to 5.7 m. That meant that levees were 5.7 m, but the "ground level" around reactors was 10 m, so the damage would have been less catastrophic for a tsunami less than 10 m. But the tsunami was 14 m tall! It is known Japan is situated near a fualt line in the bottom of the Pacific, which generates earthquakes, some of them cause huge tsunamis. Why was a huge tsunami considered unlikely? How unlikely should an event be, to considered an acceptable risk?

Some relevant information/discussion about the issue can be found among others on following links:
Huffingtonpost 110327, discussing TEPCO and others considerations about tsunami
Open democracy 110507, article mentioning tsunami risk and Japanese focus on nuclear power
Ingenjören 110510, report from a meeting about the accident at the Swedish Academy of engineering Science
VGBPowertechReport110415, German report about the accident
http://www.blog.aviatetech.com/2011/04/fukushima-i-nuclear-power-plant-lets-know-a-bit-about-it/
LACK OF DISASTER PREPARATION AT THE FUKUSHIMA NUCLEAR PLANT from "fact and details"
Washington post 110527
DN 110601


Japan (including surrounding islands) is hit by a tsunami higher than 10 m around three times a century. Among the ten tsunamis in the world since year 1700 with highest death toll, five occurred in Japan (including the latest). For the particular site I have not seen a risk projection, but Fukushima prefecture is no safe haeven. In retrospect it seems reasonable to suspect that the risk of a tsunami was downplayed. But when something like this happens and old files are mined, whistle-blowers can usually be found. That does not in itself proves they were trustworthy or not listened to. Was it a mistake (I think so) and if so how can it be avoided in the future in other countries? To learn from Fukushima this has to be discussed.

It seems an underestimate of the risk was done half a century ago, when the initial choice of the site was done. Once a lot of capital is digged down in a place, it seems likely that the investment will be vigourously defended. It does not seem to be a very good idea that people representing or associated or paid by the nuclear power industry are the main players in estimating tsunami risk once a site is decided. It looks reasonable it should be more of a national and global responsibility. This failure seems to me also an IAEA failure and not just a Japanese. On the other hand the Japanese should be world leading about tsunamis considering its relevance for the country.

Actually, soon after the tsunami, updated plans were presented to authorities for two additional reactors (7 and 8), which seemed to show a lack of will to learn from the disaster. After outrageous response from authorities the plans were withdrawn.

Even if the risk was underestimated when the reactors was initially constructed, it would have been possible to modify some security measures to a higher wave without abandoning the reactor site (like having some mobile reserve generators parked at higher ground as well as a way to hook them up). The tsunami levees were actually raised from 3.1 m nine years ago and an "administrative catastroph centre" built, so something was done to improve safety the past decade.

The authorities must understand that it is important to guard a big investment. It is also psychologically important as the operator is more likely to point at weaknesses if it is felt that it can be discussed how to solve them. We should not be too fundamentalistic when discussing the tolerance level and understand that improvement take time. Events similar to Chernobyl and Fukushima will unavoidable happen again and must be seen as tolerable if they do not happen too often.

But notable switchboards allowing available generator trucks to eliminate the black-out were not out of reach from the tsunami. There was no plan B if a higher tsunami appeared.

Probably some experts on tsunami in Japan had data to show that the risk of a higher tsunami was considerable. If it had been wider known what the reactor was designed for, the level of preparedness may have been higher. Is the information available to the public what reactors are designed to stand? I looked at the Swedish reactors on the net. Nothing was written about "design height" and from photos they do not seem much elevated above the water surface.

I wrote to Ringhals (Swedish reactor site) and got a fast reply. The reactor site can stand a water level 3 m above the average water level. A tsunami may come from the Atlantic Ocean, but it will hardly be above a meter when it reaches Ringhals. That seems to give room for a factor 3 underestimation. As the Japanese failed their estimates with only a factor 2.5, Ringhals may be safe.

If I detected information from before the accident that a more than 10 m tsunami in connection with a major earthquake is unlikely to occur once in 20 000 years at the site, and a more than 6 m but less than 10 m tsunami less than once in 5 000 years, probably I would regard the accident as an unpredictable act of God with little consequences for the safety culture, but currently I guess the risk could have been foreseen to be higher than that.

The Japanese problem with preparedness for rare events is not just the tsunami. The reactors were exposed to larger forces by the earthquake than they were designed for. They were still mainly unharmed, but this was good luck, and maybe also outdated engineering, which was less able to follow the design parameters more economically optimised than will be followed in new reactors. That the electric main grid was cut off contributed to the catastroph but has little with the reactor site to do. Was it assumed to happen more seldom than once in 10 000 reactor-years that Japanese reactors are exposed to larger earthquake forces than they are designed for? They are said to be designed to stand 8.2 magnitude earthquakes, but this was 9. Three earthquakes larger than 9 has occured on the Earth the last 60 years besides the Japanese one. Were there good reasons to assume that Japan would be spared from the strongest earthquakes?

10000 years is around the "functioning reactor"-years accumulated in the world today. Seems reasonable to tolerate one-two "INES7" reactor incidents in this time, but now it has been 4 (counting Fukushima as 3 catastrophe reactors). Hope the risk will be reduced by a factor 2-4 in the future.

Swedes became very aware of tsunamis Xmas 2004, when a tsunami killed a lot of Swedes in Thailand, I wrote an essay on that when it happened. This large tsunami was probably a world-wide alert. So long time has passed since that so I would have expected a stronger respons for Japanese nuclear plants.

Easy to sit in Sweden and speculate about tsunamis and earthquakes in hindsight. The risk for such large events causing a major disaster is low here. The more constrained lesson from Fukushima is to avoid placing reactors where tsunamis can be expected, in particular if following earthquakes, and if the risk is not neglectable, to have installations increasing quality of failure management following tsunami.

Harrisburg and Chernobyl were mainly caused by the behaviour of operators, which is impossible to predict. This has been countered by more forgiving reactors. Fukushima is different, it was a predictable foreseeable risk, which probably was underestimated. It is administrative matters about riskmanagement, which must be reviewed, e.g. how responsible bodies are set up and controlled, which is the likely most important lesson of Fukushima. This is a lecture for all countries and concerns also Sweden.