|
|
re: most efficient refrigerator
31 jan 2004
daestrom wrote:
>> >...a 'full' unit has less free air to 'spill'. which is why a
>> > unit that is full or nearly so runs more economically than an empty.
>> >>
>> >> i'm questioning that...
>> http://enews.lbl.gov/science-articles/archive/energy-myths3.html and
>> http://enduse.lbl.gov/info/lbnl-45862.pdf
>
>these studies say the change in efficiency is not measurable. fine, but
>they don't detail how much dirt or any test conditions.
they were just summaries of studies. you might look at the studies cited.
>...i have serviced more than a few refrigerators that were running all
>the time and not keeping cold. probably 25% of them were just because
>the coils were caked with pet-hair, small trash paper and 'dust-bunnies'.
>i *still* maintain that if the unit is running continuously to try and
>keep cold it must be less efficient than one that cycles off/on as designed.
sounds plausible to me.
>maybe their idea of 'dirty coils' is not as severe as what some people think
>of as 'dirty coils'...
or the cleaning is not so important as thought to be. a lot of things seem
significant in principle, but pale with a few numbers. people sometimes
suggest running a building with micro-hydro electric power from rainwater
flowing through downspouts, or attaching a 4 watt timer to a water heater.
>> >i have pointed out that your natural convection flow formula calculation
>> >is poorly applied here.
>>
>> agreed. as you say, it seems to underestimate airflow.
>
>underestimate for the 'empty' unit and *overestimate for a full unit.
i don't think so. it seems to me that the full unit would have more cold
surface to cool air that flows down through it, so the air would emerge
cooler, and thus flow faster.
>don't you see that this would mean the flow for a 'empty' unit is higher
>than the flow for a 'full' unit...
no. i don't see that.
>> >have you compared the area of the entire inside surface of an 'empty'
>> >(three walls, floor and ceiling) with the 'face' (equivalent to the back
>> >wall in area) of a stocked fridge?
i wonder what we can agree on. suppose the 2'x3'x5' empty unit has 62 ft^2
of interior walls, and the full one has 5 wire shelves with 24 vertical
3" diameter x 10" tall 1-liter soda bottles on a 6" grid on each shelf.
surrounded by 3" of air, the bottles would introduce minimal flow resistance,
but they would add 90 ft^2 of cold surface with no insulation behind it,
giving the full fridge 152 vs 62 ft^2 of cold surface, about 2.5x more...
>> >the multiple surfaces of food stocks are close together, sometimes
>> >touching.
maybe. if you pack your fridge like a box full of bricks...
>> >those food items that aren't touching form channels with a high hydraulic
>> >diameter resulting in laminar conditions and much lower flow.
i'd estimate that a 1/2" gap would provide minimal airflow resistance.
>...to think the natural convection formula for a smooth vertical plate
>with free access flow from the bottom, applies to side access flow of
>several 'jagged' flow channels isn't science.
this chimney (vs plate) formula might be accurate to an order of magnitude.
maybe that's all we need to answer this question.
>shorter, narrower channels, with only one side open will have less
>'penetration depth' of the flow. the air only goes into the unit from the
>open door a few inches.
with what channel width?
>not the full depth of the compartment. with an 'empty' unit, it has a much
>wider channel and the flow stream can actually go further back into the unit
>from the door opening.
if the room is 70 f and the fridge is 36, how much free channel area is
needed to make the air velocity near the back of the fridge at least 90%
of the air velocity near the front, 2' away?
>but again, this steady-state flow doesn't tell the whole picture...
it seems to me that the larger initial flows point to full fridges
using more energy.
>> it seems to me that the "steady state" has less flow than the transient
>case.
>
>yes.
hey, we may have agreed on something :-)
>the 'slug' flow from a 'full' unit *might* be twice the steady state,
>but the 'slug' from an empty unit could be 10 times the flow of the 'empty'
>unit. different free volumes, different slug flows.
the empty unit above has a volume of 30 ft^3... 120 1-liter bottles occupy
about 4 ft^3, so the full fridge would contain about 26 ft^3 of air. but
these are both very small compared to the volume of air that may flow through
the fridge when the door is open, since 16.6x2x3/2sqrt(5(70-36)) = 650 cfm.
>> >the 'slug' falling out of an 'empty' happens differently than your
>> >formula suggests, and is much larger than the 'slug' from a well-
>> >stocked unit (say between 5 and 10 times larger).
30/26 = 10?
>> this reinforces my argument.
>
>no, it actually refutes it. not because the 'slug' transient flow is higher
>than your steady-state formula, but because it is very different between the
>two. it is *much* higher for a 'empty' unit because there is a much larger
>'slug'. the free volume of the full unit is likely 10% or less of the empty
>unit. so the 'slug' for an empty unit is 10 times the volume of the full
>unit. if the door is only open for a brief time, this 'slug' is a major
>contributor and the *size* of it is significant.
i disagree.
>>both of us seem to have some technical understanding of this situation, but
>>neither of us have definitive data. why on earth should anyone believe you?
>
>right back at you, why should anyone believe you?
i have data. this morning my kitchen was 59 f with 40% rh. i opened
the door of my ge tbx18tazaerwh fridge (about 5 years old) and used my
recently calibrated raytek raynger st ir thermometer to measure the temp
of a spot on the inside wall and the temp of a pickle jar on a shelf:
time wall temp pickle jar temp
(sec) (f) (f)
0 40.0 39.0
15 41.5 39.0
30 42.0 39.0
45 43.5 39.0
60 44.0 39.5
75 45.0 39.5
90 46.0 39.5
105 46.0 39.5
120 46.0 39.5
an hour before, i put a small mirror inside the fridge. when the door was
open, i saw no condensation on the mirror. the wall (vs pickle jar) temp
would increase a lot faster with condensation, no?
>i model heat-transfer and fluid flow systems with engineering calculations
>for a living, been doing it for 18 years. mixed gases, steam flows, a
>variety of boiling regimes from nucleate pool to film boiling and dryout.
>before that, actually worked commercial a/c&r as well as steam plant.
>before that, 11 years in submarine nuclear power program.
great. you are an expert :-) you can actually answer these questions...
>you quote a formula from ashrae (i assume that's where you got this one, you
>get much of your material there). how about some other, more detailed
>"first principles" texts, like.....
>
>"thermodynamics", kenneth wark
>"engineering thermodyamics with applications", david burghardt
>"principles of heat transfer", frank kreith
>"engineering thermodynamics", burghardt & harbach
>"modern refrigeration and air conditioning", goodheart & wilcox
>"fluid mechanics with engineering applications", daugherty & franzini
no thanks. but looking again at the ashrae hof, i see a formula (mcadams,
1954) for heat transfer by condensation from steam on the outside of vertical
cylinders: h = 4000/l^0.25dt^0.33, eg 8618 btu/h-ft^2-f for l = 1' and
dt = 10 f. the hof continues (somewhat cryptically):
condensation heat transfer rates reduce drastically if one or more
non-condensible gases are present in the condensing vapor/gas mixture.
the decrease in the heat transfer coefficient is approximately linear with
the weight fraction of the noncondensable gas present... in a steam chest
with 2.89% air by volume, othmer (1929) found the heat transfer coefficient
dropped from about 2000 to 600 btu/h-ft^2-f.
they go on to describe various methods to estimate this drop, including
a general method by colburn and hougen (1934), but, alas, they are all
beyond my expertise. i'd also like to know how air interferes with licl
water vapor absorption.
pages 14.13 and 14.14 of rhosenow, et al's incredibly expensive 1484 page
handbook of heat transfer (mcgraw hill, 3rd edition, 1998) talk about
non-condensibles even more cryptically:
minkowycz and sparrow [41] solved this problem under free convection
conditions using a similarity transformation... chin et al. [43] modeled
both free and forced convection and solved the complete two-phase boundary
layer equations using a finite control volume method with an adaptive grid.
figure 14.10 shows their results for a steam-air mixture... the serious
deterioration in heat transfer under quiescent conditions (using the
nusselt, pure vapor case) is evident... a very small concentration of air
of 0.1 percent decreases the heat transfer by about 32 percent...
they go on to suggest a general calculation method based on reynolds, prandtl,
nusselt, and local mass transfer and diffusion coefficients, as well as local
sherwood and schmidt numbers, and a mysterious parameter betax :-)
when eq. 14.52 is compared to the numerical results from sparrow et al.
[42] for sc = 0.55, the agreement is very good... equations 14.52 and 14.53
may be used to solve for the heat transfer coefficient iteratively. the
details are provided by v.p. carey [in liquid-vapor phase change phenomena,
hemisphere publishing corp., new york, pp 378-389, 1992.]
>i think my opinion is every bit as valid as yours.
equally useless? :-)
nick
|
|
