In response to much feedback I received, on this blog and elsewhere, here is another article on nuclear power and Kudankulam, in today’s Hindu.

I’ll sit back and watch the discussion now, and let the real experts speak!

Advertisements

In response to much feedback I received, on this blog and elsewhere, here is another article on nuclear power and Kudankulam, in today’s Hindu.

I’ll sit back and watch the discussion now, and let the real experts speak!

Advertisements

*Posted by Rahul Siddharthan on October 5, 2012*

https://horadecubitus.wordpress.com/2012/10/05/another-and-probably-last-article-on-nuclear-power-and-kudankulam/

## suresh

/ October 8, 2012On risk, there is one thing that bothers me. Let x be the probability of a catastrophic accident occuring at an individual nuclear plant and let there be n nuclear plants in all. (For simplicity, assume that x is constant across plants.) We can surely assume that x is very small but not zero. Then, the probability that all plants are safe (assuming independence) is (1-x)^n. Hence, the probability that a catastrophic accident occurs somewhere is 1 – (1-x)^n. When n is very large, this probability will be close to 1 even if x is very small. Even if not close to 1, the probability can be significant. For instance, if x = 0.005 and n = 100, the probability that a catastrophic accident occurs somewhere is 0.39 which is surely significant.

This is not an argument for not having nuclear plants per se. But given that the fallout from a nuclear accident can have global consequences (unlike other forms of power generation) it seems to me that the decision on the number of nuclear plants cannot be left *entirely* to individual nation states. If that is done, then we (as a global entity) will end up having too many nuclear plants, as would follow from a standard “tragedy of the commons” type argument.

Or am I missing something?

## suresh

/ October 8, 2012I apologize for following up my own post but just wanted to add two observations. First, what I said above also applies within a country. When we consider whether or not to add a nuclear plant, it is not the riskiness associated with the plant itself that is relevant; rather, it is the total risk that there will be a catastrophic accident in the country. The latter may well be significant even when the former is very small.

The second is about the additional riskiness associated with adding a new plant. If one considers the function y = 1 – (1-x)^n and treats n as a continuous variable, then we can obtain the additional riskiness associated with adding a new plant by taking the derivative with respect to n. So we get dy/dn = -(1-x)^n(ln(1-x)). This is a decreasing function of n. (That is, the second derivative is negative.) Hence, once we have enough nuclear plants, adding one more may not make us that much worse :-) (There are 433 nuclear plants in the world currently including 20 in India if Wikipedia can be trusted.)

## Anonymous

/ October 11, 2012Suresh,

You are absolutely right, but you are too theoretical.

The probability x is a function of time. The probability of a catastrophic event happening in one year is smaller than that happening in 100 years. When you work out what that would be for 433 plants, it would be pretty close to 1, as history has shown. We are already doomed.

In reality, risks are managed and are weighed against the associated benefits.

## Rahul Siddharthan

/ October 11, 2012Suresh – interesting argument and sorry for the delay in answering (too much feedback). Yes, a nuclear accident can have global consequences, but in over half a century only one (Chernobyl) has had consequences outside the country where it occurred. And safety standards have improved a lot since then. Other than Fukushima, all the significant leakages of radiation occurred before Chernobyl — in fact, after the 1950s, the only other major event was Three Mile Island. And neither TMI nor Fukushima had global consequences, or even significant local consequences (no fatalities that can be directly attributed — and if the radiation may have long-term effects, pollutants from coal plants certainly do too).

So, anonymous, how do you get a figure of 1 for the probability of a catastrophic event in 100 years? I agree there will almost certainly be catastrophic events, not just in the next 100, but in the next 10 years — but they need not be nuclear. Just in the past 10 years, we have had the 2004 tsunami, the 2011 tsunami, Hurricane Katrina, Cyclone Thane, earthquakes, fires, a volcano that disrupted air traffic over most of Europe for weeks, etc… and those are just the natural disasters. Perspective is needed…

## Dilawar

/ December 11, 2012Reminds me of ‘Single issue activist’ in calvin and hobbes. http://pinterest.com/pin/71565081549777448/