Turbulent flow

Discussion in 'General Discussions' started by docjenser, Feb 10, 2011.

  1. Looby

    Looby Member Forum Leader

    I must say it seems counterintuitive that EWT could remain unchanged
    when LWT is lowered -- but even if that is true, you have lowered the
    average loop temperature (approximately (EWT + LWT)/2 ) in order to
    maintain the same BTUH rate of heat extraction from the earth.
  2. docjenser

    docjenser Well-Known Member Industry Professional Forum Leader

    Could that mean that the lowering of the average loop temperature could compensate for the loss in heat transfer through lesser turbulent flow, given that the HE stays the same? That would mean much lesser pumping power is possible.
  3. Looby

    Looby Member Forum Leader

    Nope, the premise that "HE stays the same" is just plain wrong. Carrying
    that train of thought to its logical conclusion would lead one to expect that
    the "optimum" flow rate would be 0.00 GPM.

    We know that reducing flow increases delta-T across the coax, and "for a
    given EWT," it must also decrease average brine temperature inside the
    coax -- and thus, reduce both HE and heat pump efficiency. (Heat pump
    manufacturers' performance specs clearly show this to be the case.)

    However, for a given loop field, in the real world, the "for a given EWT"
    condition is thermodynamically impossible. Remember, what happens
    in the loop field is a mirror image of what happens in the coax, because
    at steady state, loop field HE and delta_T must be identical to coax HE
    and delta-T. So, if you decrease both HE and LWT of the coax, it's not
    possible for coax EWT (a/k/a loop LWT) to remain the same. It must fall.

    Bottom line: For a given heating load, and a given loop field, lowering
    the flow rate must lead to reduced EWT, reduced LWT, reduced HE,
    and increased run time. And notice, that's without any consideration
    of turbulence or Reynolds#. Of course, reducing turbulence can only
    make things worse by lowering average brine temperature even more
    and further extending run times (or up-staging the compressor).
  4. engineer

    engineer Well-Known Member Industry Professional Forum Leader

    I agree with all that, but it will be interesting to see if the loss of efficiency is balanced by decreased pump power. Circ pumps are relatively inefficient; losing one or using a smaller one could yield substantial net savings even if COP drops several tenths
  5. AMI Contracting

    AMI Contracting A nice Van Morrison song Industry Professional Forum Leader

    I think the opposite point would be......shorter loops with turbulent flow would reduce required pumping power or......the hidden penalty of "extra loops".
  6. Looby

    Looby Member Forum Leader

    I agree, interesting. At the moment, I have neither data nor an opinion.
  7. docjenser

    docjenser Well-Known Member Industry Professional Forum Leader

    Ok, theory guides, experiment decides.

    I went over to a customer, told him that we need violate all the guidelines from the manufacturers, disconnect one of his 26-99 (he has 2) in order to slow the flow, and measure everything in the name of science. The good part is that we put a fairly extensive WEL system in, so we have access to the data. I also must say that the temperature sensors are all calibrated and are accurate with the actual fluid temperature measured via a calibrated digital thermometer in pete's ports.

    First, I had to cut the power to disconnect the pump, so there is a few minutes break, which gave the loop some recovery time. Then I did run the HP all night to get it and the loop to steady state, unfortunately it did shut off for a few minutes, but it did not have that much impact on loop temp.
    So shutting off the pump took the flow down from 19 gpm to 14 gpm (pretty much head on the number, measured with a flowmeter in the flowcenter).
    So after 8 hours of runtime, the EWT changed from 30.9 before to 31.0 F. However, the LWT dropped from 26.7 to 25.2F. As I mentioned earlier here, the increase of the delta T came mainly from a drop in delta T.
    The first HE and COP is correct, 38.242 KBTU/H, the second one is off, since the WEL calculates it with the 19 GPM flow. The new numbers are 38.096 KBTU/H, COP is the same, 3.51 (actually it is 0.38% less now). However, I now save 245 watts of pumping power.
    So the EWT does not change much, the LWT drop is is mainly responsible for the delta T, and the HE and COP hardly changes. (I would say within the margin of measurement error).
    Of note, loopfield here was 10 ton on a 6 ton heatpump, 10 coils, 800 ft each.

    Attached Files:

    Last edited: Feb 17, 2011
  8. AMI Contracting

    AMI Contracting A nice Van Morrison song Industry Professional Forum Leader

    The problem I'm having with this Doc is a chicken or egg question.......
    As a strong advocate of "extra loops" how do we know you have not created the "extra" pumping requirement with your design habits.
    I do not say this to bust your st:eek:nes, it's an honest and (i think) fair question.
    What if you were to employ a 6 ton field designed by manufacturer's software for instance? would that not reduce pumping power as well?
  9. moondawg

    moondawg Member

    So, on average in the US, for every 4 hours of run time, I now save (roughly) 12 cents.
  10. docjenser

    docjenser Well-Known Member Industry Professional Forum Leader

    Or at 2800 hours annual runtime assuming 14 cents/KWh, that is about $100 annually. My one runs about 3300 hours each year....
  11. docjenser

    docjenser Well-Known Member Industry Professional Forum Leader

    The pumping power here are (2) 26-99, feeding the 6 ton heatpump with 19 gpm normally, but that brings the loopflow down to 1.9 gpm. Now, cutting that one pump off brought it down to 1.4 gpm/minute, changing all the Reynolds numbers, however, performance is the same. So a normally designed loopfield would also have the same pumping power, namely (2) 26-99. I have also observed this at "normally" designed loopfields. The rule of thumb is (2) 26-99 on everything between 4 and 6 tons, and I question that.
    I usually do not oversize by more than 1 ton, the 2 of us have discussed before why. So the point here is that the design habit of a larger loopfield might have reduced the pumping requirements, not increased them as you indicated.
    We are sticking 2 energy hogging circulation pumps in series because we are so fixated on this 3gpm/ton, when actually 1 might do the job just as good.

    When brought this up I had Looby commenting:

    So at the end of the day he told me this would not be possible, that my brain is falling out and that I should bring forward data. Here is the data, although I admit it is an n=1. We have enough people here with WEL and 2 pump setups, give it a try and report back!
    Last edited: Feb 17, 2011
  12. Mark Custis

    Mark Custis Not soon. Industry Professional Forum Leader

    don't cave in DOC

    I too think that a pair of 26-99s is not a one size fits all solution. I am a big fan of water to watts attention to detail. I put up a link to either a PM or PME piece by John Seigenthaler a few weeks back..

    My favorite hydronics go to guy, learned most of what he knows about pumping fluids, riding around in atomic subs for the USN, tells me the Amish folks are going crazy buying the new ECM pumps. If you need to know brand shoot me a PM. My dealings with the "plain Folks", shows me they do not spend anything they do not need to speed. If they like and use ECM drives that is a great referral in my book.

    The same folks that are asking you to rewrite the knowledge are the ones who do not include pumping the water in the COP. I know better. I get electricity and heat transfer, maybe I came hard wired that way.

    I agree in your theory and will do what I can to help clear the debate. Just remember Tesla, he and Buick died broke. That does not matter as I do not think we will need money when we join them.
    Last edited: Feb 17, 2011
  13. engineer

    engineer Well-Known Member Industry Professional Forum Leader

    I can't think of a reason why you can't post the brand of a pump that is making folks happy...what's not to like?
  14. Mark Custis

    Mark Custis Not soon. Industry Professional Forum Leader

    If could spell

    I would have type in the Swedich guys and their Alpha series. I like the looks of but have not tasted the RI guys and their delta T pump..
  15. Looby

    Looby Member Forum Leader

    Sorry, but sumpin' is very wrong with those HE numbers:

    With two pumps: 19 gpm x (30.9°F - 26.7°F) = 79.8 gpm-°F

    With one pump: 14 gpm x (31.0°F - 25.2°F) = 81.2 gpm-°F

    So, regardless of what brine factor you choose, the gpm and delta-T
    data indicate that shutting down one pump increased HE by 1.75%.

    I would say, within the margin of measurement error, those numbers
    make absolutely no sense -- unless the loop warmed up somewhat
    between the two runs.

    Without kW measurements, can't say anything about COPs.

    Without antifreeze info, can't say much about Reynolds#.
    (I'm assuming 3/4" SDR-11 HDPE loop coils ...correct?)
  16. docjenser

    docjenser Well-Known Member Industry Professional Forum Leader

    There is nothing very wrong with those numbers, I guess the limitation is that the thermometer only measures 1/10th of a degree, and the flowmeter gets down to 1/10th of a GPM. The 1.75% difference is exactly the 0.1 degree F difference. Antifreeze is 21% Methanol, I calculated HE with 485. Yes, 0.75 SDR-11 800ft slinkies x 10. Fact is that EWT does not go down, and one cannot see any measurable difference in HE. Calculating COP of Voltage, Amperage, power factor + HE extraction. HE, Amperage and Voltage did not change,thus COP did not change.
    You should not call it very wrong just because it does fit your thinking. I hear your rationale, but it simply does not fit the observation we are making here. So lets try to understand what we see, that way we learn.
    Last edited: Feb 18, 2011
  17. Looby

    Looby Member Forum Leader

    Fair enough, I concede that point. Any changes in EWT, HE,
    or COP were too small to be detected by your instruments.

    OK, you've demonstrated that for a 67% oversized loop,
    a 25% change in gpm has no appreciable effect on HE.

    Now, suppose you brought it down to a "6-ton loop" --
    by chopping 320' off the length of each of the ten coils.
    (Just as a 'thought experiment' -- it's easier to visualize
    the effect of shortening all coils than removing some.)

    What effect do you suspect that would have on EWT?
    If the EWT would still remain the same, why bother to
    pay for 67% more pipe, excavation, and loop P-drop?
    Last edited: Feb 18, 2011
  18. docjenser

    docjenser Well-Known Member Industry Professional Forum Leader

    I am not married to the 26-99, but they usually come standard on the non pressurized flowcenters. Like I said, I am playing a bit with some Wilo variable speed pumps, spendy, but I love the control you have. You can adopt them to the individual loop, dial in the number, and the amps they draw are unbelievable. Same for variable speed Grundfos. The holdup was always the fact that we were so fixated on the 3 gpm mark, and needed 30-40 ft of head at minimum, which limits the pump selection, most of the ultra efficient pumps are out or cost an arm and a leg. But the dynamics start to change now, if you are willing to think out of the box.
  19. docjenser

    docjenser Well-Known Member Industry Professional Forum Leader

    I guess we will find out what happens if the loop is not oversized. Give me a week or so. We are having a heat wave here, and I am leaving town for a few days. I have seen on regular loops too that EWT does not drop with lower flow, for example if you have one of two pumps fail. We monitor most of our systems, I have see pattern before. The LWT will drop, EWT will stay pretty much the the same.

    The loop length will determine the overall temperature of the loop, that why we cannot go the zero length or zero flow, we need area for heat exchange. However, within a certain window, and again, I am down to 1.2 gpm, HE will be not measurably change.

    The reason for the very large loop here was the fact that the house has a 108 KBTU/H load, but a relatively new boiler, so the design was to have one 6 ton HP do the base load, and use the boiler as aux heat. Down the road that boiler will be replaced with an other heatpump, so we already made the loop big enough, so we do not need to dig up the yard twice.
  20. AMI Contracting

    AMI Contracting A nice Van Morrison song Industry Professional Forum Leader

    I've never been champion of turbulent flow for unscientific reasons.
    My next question to you doc would be to have you re cover the math of your pump savings (in kwh) with actual pump amp draw (not rating plate). I'd like to know that energy consumed by second pump doesn't increase when first is disabled. I'd like to know if this is more than a $50/yr difference on a typical 3-4 ton system.


Share This Page