| |
DF3VI > ENERGY 28.05.05 10:39l 32 Lines 1588 Bytes #999 (30) @ WW
BID : S5FDB0II_019
Read: GUEST OE7FMI
Subj: Re^2: Re^2: Hydrogen, ZL4AJS.
Path: DB0FHN<DB0FOR<DB0MRW<DB0SON<DB0SIF<DB0EA<DB0NOS<DB0BI<DB0PRA<DB0II
Sent: 050528/0928z @:DB0II.#NRW.DEU.EU [BCM M'Gladbach] obcm1.06b51 LT:030
From: DF3VI @ DB0II.#NRW.DEU.EU (Patrick)
To: ENERGY @ WW
Reply-To: DF3VI @ DB0OVN.#NRW.DEU.EU
X-Info: Sent with login password
> From: ZL4AJS@ZL4GQ.#95.NZL.OC
> You mean to say there is no such thing as an error-free program?
YES. Apart from small programlets maybe. As soon as things go complex
(and a nuclear power station is somewhat complex), programs will have errors.
As an error also include "unforseen conditions", that noone though about.
I know that the complete control system is double checked by both engineers
and authorities, and again when starting it up, and even there after, so
chances are very small, but it still can happen, as it did before.
Manual override functions are one other main source for faults, next to bad
maintenance. There have been enough reports about "minor incidents" in nuclear
power stations, when radioactive gas was released erronous, cooling pump
systems were leaking or out of operation. One day, this could lead to another
"Three Miles Island" incident. Though the containment resisted the core
meltdown, the radioactive gas was not completely contained and made it out.
Worse enough...
> Electronic sensors and circuits nowadays are so safe that the chances of
> them failing in such a way as to cause a dangerous situation are so slim
The problem with safety is that you have to multiply the possibility of
NOT-failing of each and every part in a signal processing chain.
When there are a hundred parts with 99.99%/day reliability the overall
reliability quickly drops to 99% - a failure each 100 days. Acceptable?
73, Patrick
Read previous mail | Read next mail
| |
