On Thu, 25 Feb 2021 17:36:22 +0000, Steve Walker
wrote:

On 24/02/2021 14:37, T i m wrote:
On Fri, 19 Feb 2021 13:38:54 +0000, Steve Walker
wrote:

snip

Relying on humans (human error) or automation (black Friday) to have
total control of something very dangerous is a bad thing, and why I'm
guessing it takes two keys and the presence of some important people
to launch a nuke (not that Trump being involved makes that any less
safe for the rest of us).

Except of course that nuclear power plants have the normal, programmable
(and therefore fallible and hackable - despite being locked down)
control systems backed up (for the critical safety systems) by multiple,
dedicated, hard-wired safety systems, which in turn are backed up by
fail-safe mechanical systems. And the designs, calculations and
permutations are pored over by the Nuclear, Process, Mechanical and
Control Engineers that do the designs, the Safety Engineers that oversee
the documenting of the safety systems, their equivalents in multiple
companies working on the project and finally the Office for Nuclear
Regulation.

Yet 'accidents' have happened?

The same processes apply to aeroplanes and space rockets yet they
still hit the ground and other planets pretty hard?

Both planes and rockets have the same problems: control system failure,
pilot error, mechanical fault can all lead to loss of control, which
tends to mean plummeting from the sky (or with rockets, exploding). They
also tend to have everything under software control (albeit multiple
systems).

Ok.

Nuclear power plants have hard-wired, mechanical or passive safety
systems, that continue to work, even if the control systems go wrong or
operators do things that they should not (similar would be too bulky and
heavy for aircraft).

Understood, but it was the concept of being 'pored over by the
Nuclear, Process, Mechanical and Control Engineers' I was really
reflecting ... yet they have *still* gone wrong.

Modern nuclear power plants can have have everything fail (loss of
power, plus backup power) and will remain sitting there passively cooled
- unlike aircraft and rockets where gravity has other ideas.

Sure, it's pretty obvious they weren't all directly equivalent (but
see my point above). And they have still gone wrong.

Safety has moved on massively from the early days where people could
operate the wrong valve.

I'm sure it has, in some countries. Maybe I'll be happier when all the
'old' systems have been turned of and fully decommissioned.

So you are beginning to accept that *NEW* reactors are not the risk that
*some* old ones were?

What do you mean 'beginning to'?

Please don't be like the others by 1) putting words into my mouth and
2) assuming that I wouldn't realise that such things are likely to get
safeR as time goes on.

It seems like you are trying to prove black is white and so will
obviously fail.

Where the risks are high, of course the systems to minimise the risks
should be commensurate but history has proved (so I don't need to)
that that's not always been the case.

If you are saying that now days there is no risk what_so_ever of the
latest design nukes ever 1) going wrong and 2) that then becoming more
of an issue than say a windmill falling over or a solar panel inverter
exploding then fair enough.

If you aren't saying that, but are saying that 'the real world chances
of a modern nuke ever causing any 'nuclear' problems are extremely
low' ... then we are in likely in agreement.

Cheers, T i m