Membership questions? Log in issues? Email info@brewersassociation.org

Author Topic: chiller performance: sep/oct 2013  (Read 9474 times)

Offline weithman5

  • Senior Brewmaster
  • ******
  • Posts: 1681
  • naperville, il
Re: chiller performance: sep/oct 2013
« Reply #15 on: August 23, 2013, 02:54:46 pm »
I was a chemical engineer, now a patent attorney.  I have a copy of Perry's where the data they use comes from.  The article is purely a paper study.  Their calculations are reasonably accurate for doing comparisons among common cooling options.  I would not assume that they are accurate for other purposes.    I do not consider what they are doing "BS," but I'm not surprised that one person's real world data does not coincide with an estimate from a paper study.

Theoretically, counterflow cooling should always be faster than batch immersion cooling for same coolant flow rate and same heat transfer area assuming proper design because in counter flow the coolant exiting the cooler is cooling wort at the initial temperature.  In immersion cooling, very little of the coolant is cooling wort at the initial temp.

yes, but you are assuming stagnant flow around the immersion cooler.  if you move the wort around it is similar to a typical bi flow heat exchanger. i would argue that the volumetric flow rate from a paddle moving the wort around the immersion coil is probably on the order of gallons per second as compared to a pump pumping it through a heat exchanger.  similar to how air moves over an engine be it air cooled or liquid cooled.  as long as something is moving air over the fins it will cool faster.
Don AHA member

Online hopfenundmalz

  • Global Moderator
  • I must live here
  • *****
  • Posts: 10676
  • Milford, MI
Re: chiller performance: sep/oct 2013
« Reply #16 on: August 23, 2013, 03:02:18 pm »
I was a chemical engineer, now a patent attorney.  I have a copy of Perry's where the data they use comes from.  The article is purely a paper study.  Their calculations are reasonably accurate for doing comparisons among common cooling options.  I would not assume that they are accurate for other purposes.    I do not consider what they are doing "BS," but I'm not surprised that one person's real world data does not coincide with an estimate from a paper study.

Theoretically, counterflow cooling should always be faster than batch immersion cooling for same coolant flow rate and same heat transfer area assuming proper design because in counter flow the coolant exiting the cooler is cooling wort at the initial temperature.  In immersion cooling, very little of the coolant is cooling wort at the initial temp.
Old MechE here. In Table 2, he has the counter flow with slightly higher Area for the study, the U is roughly a wash. My quandary is how does the immersion suddenly perform better than the counterflow with just a 5F change in cooling water? I think it has to do with the definition of the Delta T avg for an immersion. I might have to pull my dusty Heat Transfer book off of the shelf.
Jeff Rankert
AHA Lifetime Member
BJCP National
Ann Arbor Brewers Guild
Home-brewing, not just a hobby, it is a lifestyle!

Offline kramerog

  • Brewmaster General
  • *******
  • Posts: 2262
    • My LinkedIn page
Re: chiller performance: sep/oct 2013
« Reply #17 on: August 23, 2013, 03:03:55 pm »

Theoretically, counterflow cooling should always be faster than batch immersion cooling for same coolant flow rate and same heat transfer area assuming proper design because in counter flow the coolant exiting the cooler is cooling wort at the initial temperature.  In immersion cooling, very little of the coolant is cooling wort at the initial temp.

yes, but you are assuming stagnant flow around the immersion cooler.  if you move the wort around it is similar to a typical bi flow heat exchanger. i would argue that the volumetric flow rate from a paddle moving the wort around the immersion coil is probably on the order of gallons per second as compared to a pump pumping it through a heat exchanger.  similar to how air moves over an engine be it air cooled or liquid cooled.  as long as something is moving air over the fins it will cool faster.

I'm not assuming stagnant flow.  I'm assuming that the water in the pot cools over time due to the immersion cooler.

Offline weithman5

  • Senior Brewmaster
  • ******
  • Posts: 1681
  • naperville, il
Re: chiller performance: sep/oct 2013
« Reply #18 on: August 23, 2013, 03:13:29 pm »
so are you assuming that the water in the pot (being cooled) is moving only under natural circulation?
Don AHA member

Offline kramerog

  • Brewmaster General
  • *******
  • Posts: 2262
    • My LinkedIn page
Re: chiller performance: sep/oct 2013
« Reply #19 on: August 24, 2013, 09:12:37 am »
so are you assuming that the water in the pot (being cooled) is moving only under natural circulation?

The reason I say that countercurrent should theoretically always be better is because the flow of coolant through a countercurrent chiller is always exposed to wort at the initial temperature (~212 F), while an immersion chiller is only exposed to ~212 F for a very short moment because the immersion chiller cools the entire mass of wort.

Offline kramerog

  • Brewmaster General
  • *******
  • Posts: 2262
    • My LinkedIn page
Re: chiller performance: sep/oct 2013
« Reply #20 on: August 27, 2013, 02:50:20 pm »
I was a chemical engineer, now a patent attorney.  I have a copy of Perry's where the data they use comes from.  The article is purely a paper study.  Their calculations are reasonably accurate for doing comparisons among common cooling options.  I would not assume that they are accurate for other purposes.    I do not consider what they are doing "BS," but I'm not surprised that one person's real world data does not coincide with an estimate from a paper study.

Theoretically, counterflow cooling should always be faster than batch immersion cooling for same coolant flow rate and same heat transfer area assuming proper design because in counter flow the coolant exiting the cooler is cooling wort at the initial temperature.  In immersion cooling, very little of the coolant is cooling wort at the initial temp.
Old MechE here. In Table 2, he has the counter flow with slightly higher Area for the study, the U is roughly a wash. My quandary is how does the immersion suddenly perform better than the counterflow with just a 5F change in cooling water? I think it has to do with the definition of the Delta T avg for an immersion. I might have to pull my dusty Heat Transfer book off of the shelf.

I think you are onto something.  I would expect that the curve for the immersion cooler would curve upwards a lot more given a 3 degree F temperature approach.  I'm fairly sure that the immersion cooler equation is incorrect unless there is a very complicated equation for Delta T avg because the immersion cooler operates at non-steady state.

Offline albionwood

  • 1st Kit
  • *
  • Posts: 7
Re: chiller performance: sep/oct 2013
« Reply #21 on: August 28, 2013, 03:00:49 pm »
It's that "delta T avg" that points to the problem.  It is apparently an attempt to get around the problem of thermal gradients within the wort; the simplified linear equations assume the only gradients are between wort and cooling water.  For plate chillers this assumption is probably justified, as the distances between plates are small and the wort velocity probably results in complete mixing (within each channel).  For counterflow chillers it may be untrue but probably not far enough off to cause major error.  But for an immersion chiller, the assumption is not justified unless the wort is completely mixed at all times, such that the temperature at any point is very close to the average temperature.  Unless you stir really vigorously and continuously, this is far from true; there is a significant thermal gradient within the wort (from the center of the kettle to the chiller coils) and this greatly increases the time to chill the wort.  (The time required for heat to travel through the wort toward the chiller coils is nowhere accounted for in the equation; and the actual gradient at the coil/wort contact is much lower than the equation assumes.)

Failure to recognize, identify, and examine all the assumptions behind a mathematical model is the #1 cause of model failures.  That is the best and most lasting lesson my hydrogeology prof (Michael Campana) taught me many years ago.