Posted by Iain McClatchie on December 29, 2005, 6:03 pm
This isn't exactly related to solar thermal, but since this is where I
know people talk about thermal issues, I thought I'd post here.
I've recently bumped into a potential application for glauber's salts,
or really for any simple very dense thermal storage system: small
Modern computers now reduce the amount of power they dissipate
considerably when operating at reduced load. Our server room is rather
dramatic in this sense because we automatically turn off and turn on
computers in response to load. As a result, the power dissipation in
the room can vary from 61,000 BTU/hour to 12,000 BTU/hour. I don't
think this kind of swing is unusual.
The room is quite small and has negligable thermal storage. The air
conditioner for it, however, is pretty large (for a low-rise office
environment). When the servers are operating at part load, the air
conditioner can end up cycling on/off very frequently. This leads to
premature failure and inefficient operation.
Large server installations probably use multiple A/C units, and only
turn on a few when heat loads are low. Small installations with just
one A/C unit can't do this. There are units that can operate at half
load, but this does not address the entire problem.
What is needed is thermal storage: about 4,000 BTU/degree F. That's
4,000 pounds of water, which is inconvenient to place in such a small
room, especially near expensive computers. A better solution would be
200 pounds of Glauber's salt, about 4 ft^3, which can store about
20,000 BTUs. This could be sealed into pipes installed in the ducts.
The circulation fan stays on continuously, so there is very little film
insulation on the outside of the pipe.
Unfortunately, 90 F isn't the air temperature we want to maintain. 70
F would be better. I don't know of any compounds that melt at 70 F.
Posted by nicksanspam on December 29, 2005, 10:40 pm
How about 1000 4" diameter x 12" tall 2-liter Coke bottles on a 4'x4'x7'
tall shrink-wrapped pallet, or stacked in their plastic carton carriers?
Glauber's salts have a reputation for tiring of changing phase,
unless mechanically decohered.
Posted by daestrom on December 29, 2005, 11:02 pm
FWIW, some large installations and large A/C units also have staged cooling.
The full 8 tons of cooling might be comprised of four, two-ton units and
each cycles on with rising temperature. So when load is small, one unit
might right continuously and a second unit cycle. When load is high, three
units run continuously and the forth cycles.
One solution to your problem would be to use the same idea by using several
smaller units. But that would be a wholesale change of the system, and
quite expensive. You have, what you have.
On the face of it, your idea is a sound one though. Ideally you could
coordinate the phase change with the A/C thermostat so that when load is
light, the thermostat is set down low and both cools the room but also
solidifies the salt. Then once the salt is solidified, when the load
increases, you want the thermostat to 'set up' so that additional cooling is
gained from remelting the salt. You want to be able to automatically adjust
the cooling set point to match the conditions of load and salt.
Finding a specific solution that phase-changes at 70-75 might be a problem
though. Perhaps just changing the exact concentration would help though.
Be aware that many phase-change storage systems suffer from a sort of
'wearing out' though. Seems as the solution repeatedly changes phase, it
gets more and more prone to supercooling. That is, not 'freezing' at the
same temperature each time, but able to be cooled several degrees lower
without undergoing a phase change. Periodic agitation seems to help
alleviate this. Has something to do with the very stagnant conditions and
lack of nucleation sites.
Posted by SolarFlare on December 30, 2005, 1:26 am
I used to be involved with building pump control
panels. Dual pump control panels had a "first pump"
algorthym designed into them (in relay logic) that
would insure the constantly running pump would
alternate once the tank or vessel was below the
constant pump level. This ensured even wear on the
pumps and always a backup pump for breakdown or excess
flow requirements occasionally.
This algorythm could be put to use with an A/C control
panel with a little smarts. Many peak load controllers
do this with rotating initial zone for floor heat
This would mean the A/C units would get 1/4 as many
starts and yet have all the others for staged
BTW 4,000 pounds of water is only 400 gal (imp) or
aprox 60 cu ft.. 5 x 5 x 5 feet or 6 x 10 x 1 foot
thick floor tank.
Maybe a floating floor with some concrete or ceramics
and the A/C entering between sandwiches of thermal
mass. Most data rooms have a floating floor anyway.
Just make another sandwich layer.
since this is where I
I'd post here.
for glauber's salts,
server room is rather
turn off and turn on
power dissipation in
BTU/hour. I don't
storage. The air
part load, the air
frequently. This leads to
A/C units, and only
installations with just
can operate at half
BTU/degree F. That's
place in such a small
better solution would be
can store about
installed in the ducts.
is very little film
want to maintain. 70
that melt at 70 F.
also have staged cooling.
four, two-ton units and
is small, one unit
When load is high, three
idea by using several
of the system, and
Ideally you could
so that when load is
the room but also
solidified, when the load
that additional cooling is
to automatically adjust
70-75 might be a problem
concentration would help though.
suffer from a sort of
repeatedly changes phase, it
not 'freezing' at the
several degrees lower
agitation seems to help
stagnant conditions and
Posted by Jeff Thies on December 30, 2005, 12:49 am
I can't seem to find the site at present, but there is a company that
makes a wide variety of phase change materials (salts, waxes, oils) and
I'm sure some of those came in the temperature range you needed.
I found this in a link off of one the solar sites, perhaps someonelse
knows this source?