How Safe is Safe Enough?

Written by Rick Haynes on March 15, 2011 – 8:54 pm
 

As I read about the nuclear issues in Japan, I hear people talk about their wish to re-assess the decision to build more nuclear plants.  The question is how safe is safe enough?
The earthquake in Japan is the 5th largest since 1900, for the entire world.  What would be the emperical probability of this event?
The pacific ocean has a circumference of  84000 miles according to a site found on Google, note that the earth circumference is 25000 miles so it seems reasonable. Japan coast line = 18000 miles, so the pacific ocean refernce is not considering bays and islands.
There have been 5 earthquakes this size in the pacific in the last 110 years.
Assume that the power plant is at risk if the earthquake is within 250 miles.  (the one in Japan was around 80- miles)
Nuclear plants last 50 years.
The estimated probability would be (250/84000) * (50/110) *5  =  0.0067 or 0.67% chance it could happen at any site along the pacific coast.
This quick analysis is conservative, since there is truly more coastline than 84000 if you consider the islands and such.  But it prediced about a 1 in 150 chance that a location along the pacific coast will experience a monster earthquake.  Is that enough margin to design for?  I think so.  When I worked at a location designing nuclear reactors, we planned to protect against conditions that had a 1 in 10^9 probability.

This post is not about nuclear energy but is an example of estimating the risk.