Private Controls Of Nuclear Plants

[Godless Capitalist]

Perfectly safe.

Nothing can ever be perfectly safe. In fact, the idea of perfect safety is what is behind some of the current excessive regulations on nuclear power plants, levels of chemicals in food, etc, etc.

In my opinion government intervention would only be justified in situations equivalent to someone walking around with a live grenade with the pin out; ie where there was demonstable and immediate risk of harm to others.

[Capitalism Forever]

Nothing can ever be perfectly safe.

http://forum.ObjectivismOnline.com/index.php?showtopic=6079

[Godless Capitalist]

^^Interesting discussion but I don't see how it applies to situations like nuclear power plant safety. No matter how well engineered the plant is and how well trained the workers are, there is always some possibility of an accident that could harm someone. The possibility might be very small, but I don't see how it could ever be zero.

So then the question becomes what level of possible harm is equivalent to unreasonable negligence and thus justifies some sort of government intervention? Perhaps the answer is that it would be left to the elected legislature ...

[Rationalis]

That exact scenario may be a myth, but an accident at a nuclear plant can certainly cause extensive damage to those nearby. Chernobyl was no myth. In that sense it's a "time bomb" metaphorically if not literally.

Chernobyl, was a special case and as was posted in a thread earlier, there is a distinct difference in the structual designs of the fuel rods of nuclear plants in the United States and those of the Soviet Union (such an accident would never have happened in a plant using the American design). Now, in a completely free market economy safety itself would be a valued commodity and as such, any of this "faulty" power plant's competitors would simply have to advertise the increased safety and reliability of their own product, and backed by reliable scientific evidence(company that tests plants according to set standards and gives ratings?), would quickly drive it out of business.

Edited March 13, 2006 by Rationalis

[Capitalism Forever]

there is always some possibility of an accident that could harm someone.

Prove it.

[Godless Capitalist]

^^In order for the possibility of an accident to be 0, you would have to guarantee that no human being involved in designing, building, and operating the plant would ever make a mistake. You would have to make this guarantee even of future workers who have not even been hired and trained yet. You would also have to guarantee that the plant could not be damaged by some unforseen natural event such as an earthquake. (As yet scientists cannot predict earthquakes, and they do sometimes occur in regions thought to be geologically stable.) So basically you would have to somehow have to have complete contol of uncontrollable factors such as human behavior and geology, which is impossible.

Even assuming one could somehow eliminate all risk (for example, by requiring that the plant be built on the far side of the moon), the cost could be astronomical. It would not be reasonable for government to insist that a plant be made 100% safe at a cost substantially higher than 99% safety. A priniciple like this applied to all human activities would bring society to a crashing halt.

(changing topics) I agree with Rationalis about Chernobyl. Remember, though, that the Three Mile Island accident did expose the public to some small amount of radiation, which if I remember correctly is expected to lead to one premature death from cancer.

As for market forces, the potential risk from a power plant falls on those who live near it, who may not be the same people who are buying the power. I don't have the slightest idea where my electricity is generated or how safe the generating plant is. So consumer pressure alone will not resolve the issue.

Edited March 13, 2006 by Godless Capitalist

[mrocktor]

Let's make it a bit more precise. What the plaintiff has to prove is that the defendant is engaging in an activity that will cause harm to the plaintiff (or the person on whose behalf the plaintiff is acting) unless the defendant takes active safety measures in addition to the safety measures already known to the plaintiff.

I actually agree with this implementation with the following change: "in addition to the safety measures already in place". Whether the plaintiff knows about them or not is immaterial, a business (or a person) is not under obligation to divulge information about their own activities so the fact that the plaintiff may be uninformed cannot be used against the defendant.

I still think you are wrong about the principle of the thing though. Unlike the discussion about certainty, in this case it is objectively impossible to be 100% safe. All parts can fail (tubes, sensors, controllers etc.). These parts have known failure rates (typical of systems components) or are designed not to fail during a certain period given a certain usage cycle (structural components).

The thing is, failure rates are *averages* and structural tolerances *are statistical*. It is an objective fact that these parts can, and do, fail. What is done is to make sure that no single failure or reasonably probable combination of failures can lead to a catastrophe - however the probability of catastrophe is ALWAYS larger than zero.

This coming from an aviation safety engineer (I don't mean this as an appeal to intimidation, just for background on the technical part of the discussion).

mrocktor

[Capitalism Forever]

What the plaintiff has to prove is that the defendant is engaging in an activity that will cause harm to the plaintiff (or the person on whose behalf the plaintiff is acting) unless the defendant takes active safety measures in addition to the safety measures already known to the plaintiff.

I actually agree with this implementation with the following change: "in addition to the safety measures already in place". Whether the plaintiff knows about them or not is immaterial

But it would be rather difficult for the plaintiff to prove the inadequacy of the safety measures he doesn't know about, wouldn't it?

a business (or a person) is not under obligation to divulge information about their own activities so the fact that the plaintiff may be uninformed cannot be used against the defendant.

Normally, businesses are not under obligation to divulge information, but if they engage in hazardous activities, then they put themselves under an obligation to divulge enough information about their safety measures.

The thing is, failure rates are *averages* and structural tolerances *are statistical*.

As long as the statistics are objective, it is fine to use them in determining whether the system is safe or not.

[Godless Capitalist]

The point, though, is that the choice is not between "safe" and "unsafe." The choice s between "more safe" and "less safe." So the question remains: how to draw the line between adequate and inadequate safety precautions? How do you determine when a plant poses a risk to the public that rises to the level of being equivalent to an initiation of force?

[Capitalism Forever]

The choice s between "more safe" and "less safe."

Safety is the absence of danger. As long as danger is absent, I don't care whether it's "more absent" or "less absent."

[mrocktor]

Normally, businesses are not under obligation to divulge information, but if they engage in hazardous activities, then they put themselves under an obligation to divulge enough information about their safety measures.

I disagree. They put themselves under an obligation to take all the measures necessary not to violate rights. Period.

Safety is the absence of danger. As long as danger is absent, I don't care whether it's "more absent" or "less absent."

That is equivalent to saying "I want the risk to be zero, I don't care whether it is more zero or less zero". I have explained that zero is not possible. I'll concretize the issue further:

Flying planes is a risk. What would you consider an acceptable risk of death on an airplane flight, per hour of flight time:

1- one in a million

2- one in a billion

3- one in a trillion

4- zero

Remember, we are not talking about unavoidable risks (meteor strikes and such) or arbitrarities (being attacked by space aliens) - we are talking about things that are known to happen (bearings sieze, computers burn out...).

mrocktor

Edited March 16, 2006 by mrocktor

[Maarten]

If in a certain situation you have no specific arguments to say that there is a good chance of an accident happening, then arguing that it could happen is completely arbitrary, and should be regarded as such. If someone builds a nuclear plant of a certain type, and you can prove there is a flaw, or you can prove that someone is intentionally disregarding normal procedures then you have a good case, otherwise it's just hot air being thrown around.

[Michero]

I just had a discussion with a friend of mine about regulations (if they should be done by the state or by private institutions). We discussed doctors and everything and for everything he was bringing forth, I had a counterargument ready why private is better.

But then he started talking about Nuclear plants. And here I was lost. His argument was like this:

"If there were only private controls for nuclear plants..."

I'm not an economist and I may be way off, but don't most privately owned businesses get their money from investors? I mean, yes they make money from their customers, but really don't they need investment capital in order to continue operations?

I'm assuming that if everything were private there would be some type of consumer report (it is in the best interest for a plant to allow investigations to coax investors) and the consumer report would say if a plant was being run poorly. In that case I think investors would be very unlikely to invest in the project--if they are at all rational and thinking ahead. If there is one incident, not even accident, just one it could ruin the entire industry (for the same reason any rational seller of nuclear material would be very reluctant to sell to a poorly run plant--if people lose faith in nuclear power the materials dealer goes out of business too). Take a look at 3MI. That was actually a success story, but it scared people off of nuclear pretty much up until now.

If the consumer report investigators found egregious offences, for example the automatic safety measures were continually being overridden or regulations were not being followed I think a regulatory body could go in and investigate. I see it being the same as a building that's falling apart and may threaten houses near it-- if the people running the plant are threatening the lives and property of people around them, then the government may be able to step in and shut them down until they can raise standards.

Edited March 17, 2006 by Michero

[mrocktor]

If in a certain situation you have no specific arguments to say that there is a good chance of an accident happening, then arguing that it could happen is completely arbitrary, and should be regarded as such. If someone builds a nuclear plant of a certain type, and you can prove there is a flaw, or you can prove that someone is intentionally disregarding normal procedures then you have a good case, otherwise it's just hot air being thrown around.

A perfectly designed and operated plant can fail, and it's not arbitrary. Let us say that overheating of the core is our major concern. The temperature of the core must be monitored in order to control it effectively. Temperature sensors can fail though. Usually MTBF is used to measure a component's reliability, it stands for Mean Time Between Failures - a statistical average of how long it takes for this part to fail. Some parts fail with one hour of usage, others fail after years.

Let's say the temperature sensor's MTBF is 1000 hours. If we use three of them that gives us a probability of one in a billion that all of them will fail at the same time, for each hour of use. We can use four, six, twenty or a million redundant sensors - the probability of all of them failing will NEVER reach zero.

We can choose to implement safeguards of other natures to protect against the fragility of the sensors. Relief valves, fuse plugs (things that melt at a given temperature and interrupt operation). Every solution implemented, however, has a possibility of failure - a *real* possibility not an arbitrary one.

In the end it comes down to having enough fail safe capability, enough protections and enough fault containment to make the catastrophe vanishingly improbable. The probability *is not zero*.

mrocktor

[Maarten]

The chance of you getting hit with lightning is also not zero. Does that mean you should avoid walking outside?

I see no reason why a government-run nuclear plant would be safer; as far as I can tell all the arguments you can have speak in favor of private ownership, because the person who owns it actually invests his own money. A government official is merely using someone else's cash, and therefore has a much lower attachment to it operating as safely as it can.

Edited March 17, 2006 by Maarten

[Capitalism Forever]

That is equivalent to saying "I want the risk to be zero, I don't care whether it is more zero or less zero".

Something like that, yes. (In case you wonder, I was joking.)

Flying planes is a risk. What would you consider an acceptable risk of death on an airplane flight, per hour of flight time:

Acceptable, for whom? The passengers? Each passenger can decide whether or not the risk of the flight is acceptable for him and stay off the plane if he finds it too much. The relationship between the passenger and the airline is either mutually agreed upon (if he flies) or absent altogether (if he doesn't fly), so it is not comparable to the relationship between the nuclear plant operator and the resident next door.

Now, the relationship between an airline and a person living under the flight path IS comparable, and the same principle applies as with the nuclear plants. The airline can't tell the owner, "Don't worry, our planes will only crash on your house once out of every x flights, so you can survive our flights for n years ... on average." Regardless whether n = 1 or n = 1,000,000, this is a threat of force against the owner--a violation of his rights.

What the airline can do is ask the owner: "Would you be prepared to accept a plane crashing on your home every 10,000 years on average in exchange for a sum of $10/year?" If the owner says no, the airline will have to make a better offer, improve the safety of their planes, or fly elsewhere.

Usually MTBF is used to measure a component's reliability, it stands for Mean Time Between Failures - a statistical average of how long it takes for this part to fail.

Parts do not fail because God throws dice; they fail because of the nature of the materials they are made of makes them fail. You can establish the minimum time to failure of a given part, and as long as you don't exceed this time, there is no danger of the part failing.

[Godless Capitalist]

^^Engineering doesn't work that way. It's impossible to completely control the nature and quality of the parts that go into a machine, so that you could absolutely 100% guarantee that a part will not fail within a certain time period.

Your idea of paying people to accept risk seems plausible; most likely it would encourage operators of potentially dangerous facilities to locate them in remote unpopulated areas.

[Capitalism Forever]

It's impossible to completely control the nature and quality of the parts that go into a machine, so that you could absolutely 100% guarantee that a part will not fail within a certain time period.

You don't need to be omniscient in order to know something.

[mrocktor]

The chance of you getting hit with lightning is also not zero. Does that mean you should avoid walking outside?

Exactly, I was arguing the invalidity of "perfectly safe" as opposed to "acceptable level of risk".

Parts do not fail because God throws dice; they fail because of the nature of the materials they are made of makes them fail. You can establish the minimum time to failure of a given part, and as long as you don't exceed this time, there is no danger of the part failing.

I'm going to leave your discussion of passenger/airline and airline/landowner aside becasue I agree with you on that. I was trying to use an analogy where there is life at risk protected by engineering (which is the case with airplanes - which I work with - and nuclear plants - on which the same systems engineering methods are used).

I agree with you that epistemologically it *is* possible to be certain about safety. In practice though it's not always *viable* to do so. Statistical treatment of failure is used exactly because inspecting all the relevant components to the level we *know* is needed to be *sure* they will not fail for a "minimum time" is not practical. Imagine running microscopic analysis of all components of every airplane after every flight. So, in practice, we are *not* sure that a given plane will not crash or that a given nuclear plant will not leak - but we are sufficiently confident that that is so.

mrocktor

[Capitalism Forever]

I'm going to leave your discussion of passenger/airline and airline/landowner aside becasue I agree with you on that.

Great!

I agree with you that epistemologically it *is* possible to be certain about safety.

Wonderful!

Imagine running microscopic analysis of all components of every airplane after every flight.

Remember that safety is simply the absence of danger. If a part of the plane was designed to last for a certain minimum-time-to-failure under normal operation conditions, then, if the conditions of the flight were within the normal range, there is no need to re-examine it; the scientific facts that made the designer conclude that the part will last for time t are still valid and, if there is enough of t left, can still be used to conclude that the part will last for the next flight. (Of course, any regular inspections that the designer prescribes ought to be performed.) Danger arises only if you have a reason to doubt that a given part will last for the next flight, i.e. if the minimum time to failure has been exceeded or the facts on which the designer's prediction of the minimum time to failure is based no longer apply.

[softwareNerd]

I'm assuming that if everything were private there would be some type of consumer report ... ...

There would surely be some type of third-party check. Today, in various industries, there are businesses that check up on other businesses. Auditors check the accounts, software QA firms check up on software developers, some firms check up on exporters' goods on behalf of an importer before they leave home port. Apart from investors who would be interested in plant safety, the insurance company that provides cover would also want to check up on the plant.

It is when the government is involved that the checking is no longer a matter of self-interest of the checker. Instead, the checker is a bureaucratic institution that's often more concerned to be seen to be checking than actually doing the checks.

[Maarten]

Suppose we were to agree that control of nuclear plants should be in government hands, on the grounds that it is too dangerous an instrument for normal people to play around with (because there might be an irrational person there). Wouldn't this make a very arbitrary split in the private ownership case? How are we to determine what is acceptable danger and what isn't? One could argue that oil refineries are also very dangerous, I wouldn't want to live next to one if it explodes due to an emergency, or just mismanagement.

I don't see any good grounds for drawing the line somewhere without making it a totally arbitrary decision based on the whim of some bureaucrat or another. I think we all agree that that is not what we want. You can't treat these cases merely on a case-by-case basis, without them having influence on the underlying principles.

I still don't see why a nuclear power plant would be safer if the government operated it, the danger will always remain. I think, however, that these days these plants are so safe that there shouldn't be much cause for concern.

[Capitalism Forever]

I still don't see why a nuclear power plant would be safer if the government operated it

Indeed it wouldn't; in fact, it would probably be very unsafe--consider the example of Chernobyl!

But who on this thread has argued for government-operated power plants?

[Maarten]

I thought the discussion was whether or not they should be owned by private individuals. Well, if we conclude that they shouldn't be who else can operate them but the government? I'll go reread the thread then to see what the actual discussion was about. Sorry if I misinterpretated what this was about.

[Capitalism Forever]

I thought the discussion was whether or not they should be owned by private individuals.

No, no, even the original question was only about whether or not they should be regulated by the government. The subsequent discussion that I triggered has been about whether one private individual has a right to verify the safety of a nuclear plant owned by another private individual (and about whether safety is possible at all).