[TriLUG] OT: thermodynamics of A/C question

[Matt Flyer]

On 06/23/2012 11:35 AM, Joseph Mack NA3T wrote:
> still trying to get a handle on the connection between wet bulb temp
> and power required for cooling. Is it the evaporative cooling that
> does it?
>
The wet bulb temperature would definitely effect evaporative cooling
systems, because they can't go any lower than the wet bulb temperature. 
In practice, they usually go a few degrees above.  I've been told, but
haven't studied that with high wet bulb that air cooled system
performance goes down worse than evaporative.

This pdf (long) looks like it goes into an analysis of the subject:
http://www.google.com/url?sa=t&rct=j&q=condenser%20efficiency%20versus%20wet%20bulb&source=web&cd=10&ved=0CGMQFjAJ&url=http%3A%2F%2Fasset.sce.com%2FDocuments%2FBusiness%2520-%2520Services%2520for%2520Your%2520Business%2FEvaporativeandAirCooledReport.pdf&ei=h-3lT4DfDY2u8QTp24imAQ&usg=AFQjCNF9dJSuLgzSYel-R35rfKMfu-tUIw

> When I was a kid, I tried evaporative cooling on a towel, but the
> dissolved minerals crusted up the towel eventually. Is this a problem
> with your setup?
>
It would be, except for chemical treatment.  The system at UNC is a
closed loop on the chilled water side, with about 5 million gallons of
water in 25 miles of underground piping.  Surprisingly there isn't much
loss in the ground.  The water is kept at about 44F.  It is filtered,
treated with anti-corrosives as well as biocides.  One problem is that
contractors working with the exposed pipes like to drink the water
because it is cold and looks crystal clear.  On the closed loop system,
the pumping power is largely governed by differential pressure from the
supply and return.  As the building air handler valves open, it acts
like a short circuit reducing the pressure, which requires more pumping
for flow (with pressure being a resistance to the flow).

If you have a water system, you also need to consider the hydraulics and
whether or not your pumping water against pressure head (rise in
elevation), which costs power.  This is a factor in cooling towers,
where your pumping the water up a riser pipe to get it to the top of the
tower for it to fall through the tower fill.  This costs energy. 
>> The big electricity driver is being how low you can get the
>> refrigerant temperature on the hot side as this causes the pressure
>> to go up which is what the compressor needs to pump against, which
>> takes horsepower.
>
> This would also point to doing your cooling at night then?
>
Often times we do, but it depends on electric rates, which have to be
high enough to compensate for the losses.  In addition to the 5 million
gallons circulating through campus, there is a tank that holds another 5
million gallons.   It has so much thermal mass that it takes days to
weeks depending on the temperature to change it much.  So far this year,
Duke Power has been charging a "demand" date of about 3.6-4.2 cents per
kw (typical residential is 10-12 cents).  However, in peak demand which
occurs during the hot part of the day it can go as high as 25-30 cents
per kw.  During these times, we switch the chillers off and pump from
the tank.  This can save upwards of $70K (taxpayer dollars) a day.  Then
at night, the tank is recharged, with supercooled water (~40F). 
However, in order to do this during the hours available we have some
big, high horsepower machines that are more expensive to run.

> When you vary the speed, what are you varying? The speed of the
> compressor? If so, then you're dropping the pressure and hence the
> temperature in the condenser. If you drop the temperature in the
> condenser, then you aren't going to be able to dump as much heat either.
>
You can control both the speed of the compressor.  One factor involved
is that the machine outputs are not linear in terms of watts in, btu
out.  They have, more or less, a wide parabolic curve with a "sweet
spot" where you get the greatest efficiency.  Your home units would be
the same way but with only one machine there isn't a way to control this.

In addition, in larger systems, especially water systems, you can
control the speed of the pumps as well as the speed of the fans.  While
it may seem counter intuitive, a lot of times you can run two fans or
two pumps at a lower speed to produce the same flow at lower power.

One other cost with evaporative cooling is the need for makeup water as
you are dissipating water into the air.  (This is part of your
coefficient of performance or SEER)  How much do you pay per gallon of
water?  At UNC, we use 'reclaim' water which is effectively grey water
that would be dumped into the stream beds after treatment.  It costs
about 1/5th as much.

> I understand that A/C works most efficiently when it runs
> continuously, that it takes about 10 mins for everything to come to
> thermal equilibrium. On the ramp up (for some reason I don't
> understand), efficiency is low. In that case you are supposed to get
> an A/C, that on a typical hot day, has just enough capacity to handle
> the heat, when running continuously. ie don't get a bigger unit than
> you need.
>
> I imagine then if you had two smaller units, that for cooler days you
> would have one of the units run continuously. When it got hot, a 2nd
> small unit would kick in. I wonder how economies of scale work here. I
> can imagine that a 2x sized unit doesn't cost 2x, but maybe 1.5x. In
> that case you'd have to make up the difference in running costs. If a
> unit costs 5k$ and your annual bill is 1k$, then your savings might be
> 200$ for having two units. This savings would take a while to pay back.
>
> Joe
>

One way you could save money on a home unit is by using two speed fans,
or variable speed fans.  This would allow you to produce just as much
cooling as required, but no more.  Going variable speed (through
semiconductors as opposed to two speed via switching) has losses too,
though and these are a continuous overhead, typically about 3% of the
full load but with a wide operating span can save a lot of money.