phk at phk.freebsd.dk
Fri Jun 9 16:05:16 EDT 2017
In message <593A4677.5080509 at yandex.com>, Charles Steinmetz writes:
>The insides of mains alternators are almost entirely metal -- tons and
>tons of copper tubing, and the casings and rotor shaft are steel. And
>the alternators must operate at a relative humidity of absolute zero.
This is close to to being "not even wrong".
There is no requirement for "a relative humidity of absolute zero",
which is theoretically impossible if there is any iron around.
There are however stringent requirements for non-corrosion, dielectric
potential, cooling, corona behaviour etc. etc.
Cooling the stator is a no-brainer, it is almost always water-cooled.
If they could afford to use helium to cool the rotor, they would
jump on it instantly, but the cost is prohibitive at their leakage
Pure water would also be workable, but its high density means
it cannot be used to cool the rotor without a serious hit to
Next down the line is hydrogen, which comes with a shitload of issues.
Apart from all the obvious issues, lube-oil degradation, polymer
degradation, fire-risk, risk of explosion, health-risk etc. hydrogen
has "interesting" solubility in metals.
If you want to purify hydrogen, you press it trhough a filter
consisting of a solid slab of palladium, and that takes a lot less
pressure than you would expect.
For iron in particular, hydrogen means embrittlement, so a
major focus in rotor design is to keep the hydrogen away from
If you Google "generator hydrogen seal" and and you will find
little love for hydrogen cooling.
In 1993 Siemens put the first 170MVA air-cooled generator optimized
by computer simulations of the flow-fields and since then the
hydrogen cooling has been confined to an ever-decreasing top tier
of name-plate power.
Today fan cooling will take you to approx 400MVA and pressured air
cooling will take you to about 600MVA.
Above that, you are, almost by definition, in a nuclear power plant,
and all problems from hydrogen cooling of your single huge generator
pale in comparison to having a handful of smaller generators in
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
More information about the time-nuts