Posted by daestrom on July 7, 2004, 2:35 am
If you model with only an 'upper grey water' and a 'lower grey water', then
the stratification is limited to two layers. More time steps won't make
more layers. Why do you think it would? Try it.
The time between 'bursts' will be a factor here. Even with perfectly
stratified water 'layers', heat will be conducted between each 'layer'. If
the time between 'bursts' is really long compared to the time-constant for
the layers, the tank will reach a uniform temperature throughout before the
next 'burst'. You don't want that. You want the grey water in the upper
layers to remain hot and not conduct its heat to the cooler layers
underneath it. You want the bottom grey-water layer to be as cold as the
incoming freshwater. That way, when it is pushed out by the next 'burst', i
t won't be taking much energy with it. On the other hand, if the entire
tank equalizes, then the next 'burst' will push warmer water out the
grey-water outlet, wasting energy.
In the case of tube-within-a-tube, the same problem will occur if the
sucessive coils of the outer tube are in direct contact.
Well, you mentioned the price of a drum in your earlier calcs, so I
*assumed* you had to pay for the drum.
Of course, a more expensive setup might be made with several straight
sections of copper pipe. Two different sizes, and reducing 'Tee's on each
end with the smaller tube 'crossing the top of the tee' while the outer
tube's water comes out the base of the 'tee'. But that's a lot of joints to
'sweat' and you couldn't disassemble it for cleaning. But with copper, you
might get away with much shorter total length (owing its higher
conductivity). And put the grey water through the smaller central tube with
just the 90 degree elbows on the ends. Could probably flush it easier than
the outer tube, and with just a straight-thru shot and elbows, it might not
foul as much.
Oh? And emptying out a 55 gal drum with this 'gunk' in the bottom of it is
any better? Oh wait, you might just replace it with one of those free drums
you have. Sorry, we're not all that lucky.
Posted by nicksanspam on July 7, 2004, 6:50 am
We might simultaneously increase the number of layers and the number of
time steps. If half the tubing fills in 4.8 minutes, we might consider
at least 2 layers... with at least 4 layers for 2.4 minutes, and so on.
When x = 1 and erf(x/sqrt(4at)) = erf(0.5) = 0.5... from page 89 of the
1998 Shaum's Outline on Heat Transfer. Section 4.4 on "One-dimensional
systems: fixed surface temperature, semi-infinite body."
This could be close to ideal stratification. Bouyancy would help even more.
How do we combine the bouyancy and diffusivity effects?
It looks to me that very little heat will conduct downward between layers...
180 hours is very long compared to the time between bursts. Altho the drum
walls are steel, about 1/16" thick.
Which it isn't...
Sure. That's what we want.
We don't want that.
Rarely. You can probably find used drums where you live as well.
Bakeries tend to buy syrups and honey in these kinds of drums.
I'd hope to avoid that with natural flow, but it might have a bulkhead
boiler valve near the bottom for backflushing into the septic system.
Posted by Fred B. McGalliard on July 6, 2004, 3:51 pm
I have been thinking on a stratified storage method. There are two ways I
can think of to do this. One is to let the warm water enter near the tank
bottom and flow up an insulated column with downward pointing vanes, so the
hot water will rise till it reaches the level of water at that temperature
plus a small delta, then it will flow out, adding to the strata at that
temperature. The other method would use a temperature sensor and a moveable
drain tube/orifice to achieve the same stratification. This would work to
capture and store solar warm, solar hot, ambient (when you have a hot day),
and grey water heat, all in the same tank.
Posted by daestrom on July 6, 2004, 9:16 pm
Interesting idea. Just have to make the vanes large enough and/or keep the
flow low enough that the inflow rate doesn't disrupt the stratification too
much. Keep the Reynolds number down below the turbulent point and it could
The other thing you would need is that the bouyancy/rising force be able to
move the water up the column as fast as the incoming water. Otherwise, the
extra flow of incoming warm water would 'overflow' the downward pointing
vanes and spill out in a region colder than you want.
Active motion control of the tube may be difficult to attain. But perhaps
some sort of 'bulb' on the end that would change bouyancy with the flowing
water's temperature? Then it would rise/fall automatically with the
temperature changes (with some slight delay for the bulb to adjust).
Posted by Fred B. McGalliard on July 7, 2004, 1:49 pm
Nice. A smart thermal toilet bowl valve, sort of. What we want though has to
rise until the outside temperature is slightly cooler than the inside, or
the top, whichever comes first. A simple thermal riser should be easy enough
to design but I don't off hand see how we would do this relative to the tank
temperature? Well, I'll think on it some more.