Thermo-siphon

  • Active since 1995, Hearth.com is THE place on the internet for free information and advice about wood stoves, pellet stoves and other energy saving equipment.

    We strive to provide opinions, articles, discussions and history related to Hearth Products and in a more general sense, energy issues.

    We promote the EFFICIENT, RESPONSIBLE, CLEAN and SAFE use of all fuels, whether renewable or fossil.
Status
Not open for further replies.

Armaton

Member
Aug 22, 2011
147
Hastings, Michigan
Getting ready to put my system together, and was wondering about the thermo-siphon capabilities of my LK 810. I understand how it works, but.... Boiler loop pipe is all 1 1/4", and the tank is only 4' from the boiler. Top of storage input bung is 1 1/4", but the bottom of my tank where I was going to pull the water to the loading unit, has an existing 1" bung, (that I didn't want to replace, and the guru's on this site have said wouldn't hinder boiler flow), but, it is about 10" below the level of my boiler input, therefore 10" below the level of where I was going to place my LK 810. Will the thermo-siphon draw through the LK 810 and 1" bung, be enought to pull the cooler water "UP" from the bottom of my tank in a power outtage? Or do I need to draw from level? The input pipe to the storage tank is 9 1/2 feet above the output, so should thermo-siphon big time, I would think. Makes a big difference on how I plumb it. I'm hoping to utilize thermo-siphon as my overheat protection. As long as I use batch burns, and don't fill the burn chamber when storage is close to full it should work great, but not sure on the pipe levels and that 1" bung. I could also place the LK 810 at the same level as the bung, and use 45 degree elbows to get back up to the boiler input level, to help with the flow. What do you think?
 
Getting ready to put my system together, and was wondering about the thermo-siphon capabilities of my LK 810. I understand how it works, but.... Boiler loop pipe is all 1 1/4", and the tank is only 4' from the boiler. Top of storage input bung is 1 1/4", but the bottom of my tank where I was going to pull the water to the loading unit, has an existing 1" bung, (that I didn't want to replace, and the guru's on this site have said wouldn't hinder boiler flow), but, it is about 10" below the level of my boiler input, therefore 10" below the level of where I was going to place my LK 810. Will the thermo-siphon draw through the LK 810 and 1" bung, be enought to pull the cooler water "UP" from the bottom of my tank in a power outtage? Or do I need to draw from level? The input pipe to the storage tank is 9 1/2 feet above the output, so should thermo-siphon big time, I would think. Makes a big difference on how I plumb it. I'm hoping to utilize thermo-siphon as my overheat protection. As long as I use batch burns, and don't fill the burn chamber when storage is close to full it should work great, but not sure on the pipe levels and that 1" bung. I could also place the LK 810 at the same level as the bung, and use 45 degree elbows to get back up to the boiler input level, to help with the flow. What do you think?

LK should work the same in either position. The only way you could get in trouble would be to form a 'thermal trap', which it doesn't sound like you're doing.

A restriction down to 1" at the tank port won't amount to much.

As a point of reference, my boiler inlet is about 18" above bottom of storage, and top of storage is only about 40" above boiler outlet. (but it's all 1.5" pipe, four feet over to storage). I got storage up to 185 degF top to bottom with a full fire going on top of a 8 inch bed of coals and cut the power. Supply temperature got up above 220 degF for a while until lack of oxygen slowed combustion down, but it thermo-siphoned nicely with no kettling or pressure spikes.

--ewd
 
Hah, I read this post before there were any replies and I immediately thought of EWD. Good to have you back.

Noah

And don't forget, now is the time to shop! Pier 1 imports save up to 20%! Ebay buy now! Progressive makes it easy to compare rates and save!

I do welcome relevant (or irrelevant even) ads above or below the thread text, but don't see the point of inserting them into our conversations.
 
If I read right that your top of storage is 9-1/2 feet above your boiler outlet - you should thermosiphon like crazy even if your storage bottom outlet is only 1". I only have a couple of feet rise from my boiler outlet to top tank inlet, and my LK810 is also above the bottom storage outlet, by a foot or so. I haven't killed power while the boiler is up to speed to see what happens in the loading circuit, but I do know it does thermosiphon quite well at the end of burn after the Lk pump shuts off, to get the remaining heat out of the boiler.

9-1/2 feet - holy crap that's a lot. I'd like to see pics of that setup, and your tanks.
 
The tank is an upright Sylvan air receiver, 112" tall and 4' diameter, (found at the local scrap yard for $400) ASME plate is still on it, contacted Sylvan, and they emailed me the build sheet, should be around 850 gallons. What I meant was I have approximately 9 1/2' differential from the boiler inlet to the storage inlet, or 5' from boiler outlet to storage inlet, whichever you want to use. Have to put 7" leg extensions on the tank so I can get to the bottom bung a little more convienently. Sat the tank up in my polebarn this last Tuesday, and the top bung creased the drywall on the ceiling, but it is upright now. A little scary tilting a 1500 lb tank up, knowing that if it goes over too far it could/would go through the wall.
 
Hello Armaton,

What boiler did you end up with?

I'm guessing you will be in great shape with 5' of rise from boiler supply to the top of storage. I think maintaining good sratification will be important when using storage as overheat protection, especially towards the upper end of storage temps. Hopefully your boiler to storage flow rate will be low enough to encourage this but maybe its not really an issue.

Noah
 
After considering nearly all of the standard smaller units, I decided to get an EKO 40. $4700 delivered, savings actually payed for most of my other stuff, including the Rehau flex pex. Would have really liked to have had a Vedo 30, but the extra money and mandatory overheat loop, kinda pushed me to the EKO. Older technology, but tried and true, with a solid information base on hearth. Was a little disappointed that it was a 2008 model, but the fact that it's a Super kinda softened it.
 
Mandatory overheat loop, hmmm? Do you mean the cooling coil with the STS valve? I am using this and thermosiphon to/from storage as my overheat strategy and have discussed this with Dean and he is good with this Euro approach. And I don't have the thermosiphon potential that you have-I know I'll have some but don't know how much-We'll see. I plan on killing power mid burn at different storage temps to see what I can get away with. Start at the low end and work my way up, unlike EWD, who it seems can handle more excitement than me!!!!

Good, safe choice with the EKO. Lot's of Boiler Room knowledge of them, no doubt.

Noah
 
I plan on killing power mid burn at different storage temps to see what I can get away with. Start at the low end and work my way up, unlike EWD, who it seems can handle more excitement than me!!!!
The plan was to minimize excitement, but I must admit watching the temperature gauge climb was making me nervous. The boiler has a European style quench coil that I hooked up to a garden hose temporarily. The manual says the limits are 230 degF and 36 psig, so I killed the power and watched and listened. The system has adequate expansion capacity, so pressure didn't go up too much, and temperature 'only' got up to about 225 degF, so I rode it out and never had to open the quench valve. 'Worst case is the only case', I like to say.

It was supposed to work, according to some rough calculations I had done. And after all there are lots of old tractors that send upwards of 100k btu per hour up to their radiators with no water pump with only a foot and a half of rise to work with.
 
Mandatory overheat loop, hmmm? Do you mean the cooling coil with the STS valve?[/quote

No, while talking to Dean about the 30, I explained how I was going to use thermo-siphon for my overheat protection, he said that Varmebaronen would "probably" still require the overheat loop, as well as mandatory return temp valve, and storage, to keep warranty. I know the overheat loop isn't a big deal, just another couple hundred bucks, and I was going to use storage and an LK 810 anyway, but I don't feel that they should be mandated by the company to keep warranty. What if I wanted to use "Bang bang" to insure that it didn't condense? And I felt that my tank setup would provide equivelant, if not better, overheat protection, than their mandated loop. Anyway, felt the EKO was much less hassle as well as money.
 
It was supposed to work, according to some rough calculations I had done. And after all there are lots of old tractors that send upwards of 100k btu per hour up to their radiators with no water pump with only a foot and a half of rise to work with.

EW, would mind sharing your calculations, particularly how you figured the gpm you would be able to thermosiphon? I spent some time here http://www.engineeringtoolbox.com/gravity-heating-systems-d_189.html trying to figure things out but it's seems I wasn't up to the challenge of figuring out how this translates to flow rates.

Thanks,

Noah
 
EW, would mind sharing your calculations, particularly how you figured the gpm you would be able to thermosiphon? I spent some time here http://www.engineeringtoolbox.com/gravity-heating-systems-d_189.html trying to figure things out but it's seems I wasn't up to the challenge of figuring out how this translates to flow rates.
I just took the weight of two one inch square columns of water, the hot column and the not so hot column and calculated a pressure differential.

Then ignored elbows and took a length of pipe that was longer than actual to fudge in the right direction.

Then took a gpm-to-head Hazen-Williams formula and used Newton's method to find out what the gpm would be for any given head, seeing as how that seemed easier than solving for gpm at the time.

head = (1044 * gpm^1.85) / (hwc^1.85 * d^4.865)

where
head is ft H2O per 100 ft pipe
gpm is gpm
hwc is Hazen-Williams constant
d is diameter in inches
^ is exponentiation

ewdudley@dell-desktop:~/woodshed$ perl thermosiphon.pl

psi_t0=1.88922; psi_t1=1.85453; psi_diff=0.03469 psi; head=0.08001 ft_H2O
t0=185 degF; t1=230 degF; rise=54 inch; d=1.5 inch; length=16 ft; gpm=4.66294

psi_t0=1.90594; psi_t1=1.86281; psi_diff=0.04313 psi; head=0.09947 ft_H2O
t0=160 degF; t1=220 degF; rise=54 inch; d=1.25 inch; length=16 ft; gpm=3.24747



Code:
Code:
use strict;
package water_density;

#  t     ft^3/lb   lb/ft^3  lb/gallon
my @tbl = (qw(
   32    0.01602    62.41    8.344
   40    0.01602    62.43    8.345
   50    0.01602    62.41    8.343
   60    0.01603    62.37    8.338
   70    0.01605    62.31    8.329
   80    0.01607    62.22    8.318
   90    0.01610    62.12    8.304
  100    0.01613    62.00    8.288
  110    0.01617    61.86    8.270
  120    0.01620    61.71    8.250
  130    0.01625    61.55    8.228
  140    0.01629    61.38    8.205
  150    0.01634    61.19    8.180
  160    0.01640    60.99    8.154
  170    0.01645    60.79    8.126
  180    0.01651    60.57    8.097
  190    0.01657    60.34    8.067
  200    0.01664    60.11    8.035
  210    0.01670    59.86    8.002
  212    0.01672    59.81    7.996
  220    0.01678    59.61    7.969
  240    0.01693    59.08    7.898
));

sub new {
   my $class = shift;
   my $s = {};
   
   my (
    $j,
    $i,
    $cnt,
   );
   $cnt = scalar(@tbl);

   for ($j = 0; $j < $cnt; $j += 4) {
      my %r = ();
      @r{qw(t ft3_lb lb_ft3 lb_gal)} = @tbl[$j .. $j+3];
      push(@{$s->{density}}, \%r);
   }
   bless $s, $class;
}

sub lb_in3 {
   my $s = shift;
   my $t = shift;
   my $lb_ft3;

   for (;;) {
      my $d0;
      my $d1;
      ($t < ($d0 = $s->{density}->[0])->{t}) && 
       ($lb_ft3 = $d0->{lb_ft3}) && 
       (last);
      ($t >= ($d0 = $s->{density}->[$#{$s->{density}}])->{t}) && 
       ($lb_ft3 = $d0->{lb_ft3}) && 
       (last);
      for (my $i = 0; $i < $#{$s->{density}}; $i++) {
         ($t >= ($d0 = $s->{density}->[$i + 0])->{t}) && 
          ($t <  ($d1 = $s->{density}->[$i + 1])->{t}) && 
          (last);
      }
      $lb_ft3 = $d0->{lb_ft3} + (
       (($t - $d0->{t}) / ($d1->{t} - $d0->{t})) * 
       ($d1->{lb_ft3} - $d0->{lb_ft3})
      );
      last;
   }
   return($lb_ft3 / 12.0**3);
}
1;

package main;

# In order to calculate the friction loss of water, the Hazen-Williams
# formula is used:
# head =
# Where
# head = Friction loss in feet of water per 100 ft.
# gpm = Flow rate (gpm)
# hwc = Hazen-Williams Coefficient
# d = Inside Diameter (in.)
# The Hazen-Williams coefficient for Lamson Vylon HDPE pipe is
# 150 and doesn't change over time. With its superior corrosion
# resistance it will remain smooth and not corrode or tuberculate.
# head = (1044 * gpm**1.85) / (hwc**1.85 * d**4.865)
#
sub gpm_to_head {
   my $s = shift();
   my $ft_h20_per_100ft = 
    (1044.0 * $s->{gpm}**1.85) / 
    ($s->{hwc}**1.85 * $s->{d}**4.865)
   ;
   return($ft_h20_per_100ft * ($s->{l} / 100.0));
}

# Take d=diameter l=length head=head hwc=hazen-williams coefficient and calculate flow
#  using head calculator iteratively.
sub head_to_gpm {
   my $s = shift();
   my $gpm_was = 0.0;
   my $head_was = 0.0;
   my $head_goal = $s->{head};
   my $head;
   $s->{gpm} = 1.0;
   my $i;
   for ($i = 0; $i < 20; $i++) {
      $head = gpm_to_head($s);
      (abs($head - $head_was) < 0.000001) && (last);
      my $slope = ($head - $head_was) / ($s->{gpm} - $gpm_was);
      my $gpm_nxt = $s->{gpm} + ($head_goal - $head) / $slope;
      if (0) {
         printf("head=%.5f head=%.5f slope=%.5f gpm=%.5f gpm_nxt=%.5f\n"
          ,$head_goal
          ,$head
          ,$slope
          ,$s->{gpm}
          ,$gpm_nxt
         );
      }
      $gpm_was = $s->{gpm};
      $s->{gpm} = $gpm_nxt;
      $head_was = $head;
   }
   return($s->{gpm});
}

sub ft_h2o {
   my $psi = shift;
   return(2.3066587 * $psi);
} 

sub gpm {
   my $s = shift;

   my $wd = water_density->new();

   my $lb_in3_t0 = $wd->lb_in3($s->{t0});
   my $lb_in3_t1 = $wd->lb_in3($s->{t1});

   if (0) {
      printf(
       "t0=%s degF\n".  
       "t1=%s degF\n".  
       "rise=%s inch\n" .
       "lb per gal t0=%.5f\nlb per gal t1=%.5f\n"
       ,$s->{t0}
       ,$s->{t1}
       ,$s->{rise}
       ,$lb_in3_t0 * 231.0
       ,$lb_in3_t1 * 231.0
      );
   }

   my $lb_t0 = $s->{rise} * $lb_in3_t0;
   my $lb_t1 = $s->{rise} * $lb_in3_t1;
   my $lb_diff = $lb_t0 - $lb_t1;
   my $ft_h2o = ft_h2o($lb_diff);
   printf(
    "psi_t0=%.5f psi_t1=%.5f psi_diff=%.5f psi head=%.5f ft_H2O\n"
    ,$lb_t0
    ,$lb_t1
    ,$lb_diff
    ,$ft_h2o
   );

   $s->{head} = $ft_h2o;

   printf("t0=%s degF t1=%s degF rise=%s inch d=%s inch length=%s ft gpm=%.5f\n"
    ,$s->{t0}
    ,$s->{t1}
    ,$s->{rise}
    ,$s->{d}
    ,$s->{l}
    ,head_to_gpm($s)
   );
}

sub main {
   printf("\n");

   my %s;
   # Hazen-Williams coefficient=100 diameter=1.5 inch length = 16 ft.
   $s{hwc} = 100.0;
   $s{d} = 1.5;
   $s{l} = 16.0;
   $s{rise} = 54.0;
   $s{t0} = 185;
   $s{t1} = 230;
   
   gpm(\%s);
   printf("\n");

   $s{hwc} = 100.0;
   $s{d} = 1.25;
   $s{l} = 16.0;
   $s{rise} = 54.0;
   $s{t0} = 160;
   $s{t1} = 220;
   
   gpm(\%s);
   printf("\n");
}

main();
 
Now theres some "Rough calculations" for ya.
Hey look, here's an on-line calculator that would have saved some fuss if I had found it when I designed my system:

http://www.calctool.org/CALC/eng/civil/hazen-williams_g

You still have to calculate the pressure difference according to rise and temperature, but it calculates the gpm.

Looks like my algebra agrees with theirs:



Gravity-fed pipe flow calculator.png
 
EWD, you've given me the missing pieces! I've been wanting to figure this out for a while but couldn't really get my brain around till this evening. It is reassuring to know I have the potential to thermosiphon 3+ gpm in a overheat situation. More than I would of thought, honestly.

I appreciate the help,
Noah
 
EWD, when I put my numbers into the calculator, ie 1.25 pipe, roughness coefficient 100, the 10 foot of pipe, and only a 4 foot drop not 5, I get 30 gallons a minute. Is that possible? Or are my numbers off? If the numbers are right, maybe Zennon was correct when he told me that with my "proposed" setup I wouldn't need pump or danfoss, and I was sceptical. I still wouldn't do it that way but possible?
 
EWD, when I put my numbers into the calculator, ie 1.25 pipe, roughness coefficient 100, the 10 foot of pipe, and only a 4 foot drop not 5, I get 30 gallons a minute. Is that possible? Or are my numbers off? If the numbers are right, maybe Zennon was correct when he told me that with my "proposed" setup I wouldn't need pump or danfoss, and I was sceptical. I still wouldn't do it that way but possible?

To use the calculator for our purposes the 'drop' is the pressure differential between the hot column of water and the not so hot column of water. In my case I used a column height of 54 inches, which is 36 inches above the boiler plus half the boiler water jacket height, which is 18 inches, based on the assumption that there is a temperature gradient in the boiler jacket (where we see that it's not the calculations themselves that are rough, it's the model behind them: Garbage in, Gospel out to five decimal places!). So in my example the difference in weight per square inch between a one inch square column of water 54 inches high at 185 degF and another at 230 degF is 0.03469 psi, or 0.08001 feet of positive head.

So 0.08 feet of head driving flow through 16 ft of 1.5 inch of pipe with a Hazen-Williams roughness coefficient of 100:

thermo_siphon.001.png

And as a special bonus to those who have stuck it out this far, we can add about 10% to the gpm because Hazen-Williams is based empirically on 60 degF water and overheated boiler water is significantly less viscous, enough so to overcome the fact it is less dense. And another thing, some sources put the Hazen-Williams coefficient for steel pipe at closer to 120 instead of 100, so you can use that number if you're feeling optimistic.

However need to be sure to add three or four feet of pipe length equivalent for each elbow, several feet for the loading unit, and so forth for each resistance in the path. There's equivalent pipe length fitting tables out there you can google.
 
And a special bonus to those who have stuck it out this far, we can add about 10% to the gpm because Hazen-Williams is based empirically on 60 degF water and overheated boiler water is significantly less viscous.

I suppose we could enter 1.38" for 1-1/4 pipe and 1.61" for 1-1/2 pipe?(sch 40). Just to tweek the numbers even more in our favor, of course.
 
If the numbers are right, maybe Zennon was correct when he told me that with my "proposed" setup I wouldn't need pump or danfoss, and I was sceptical. I still wouldn't do it that way but possible?

I think the big snag would be return water protection. If it wasn't a problem.... Imagine a natural draft gasifier with large pipe/short runs to vertical storage and an old school CI radiator/gravity system. Non electric TRVs for some control. But alas, got to have a little electricity to keep the boiler happy.
 
I think the big snag would be return water protection. If it wasn't a problem.... Imagine a natural draft gasifier with large pipe/short runs to vertical storage and an old school CI radiator/gravity system. Non electric TRVs for some control. But alas, got to have a little electricity to keep the boiler happy.

I thought this through once and it seemed to me if you wanted to go off grid with a natural draft Varm you could do it with a mixing valve six or eight feet above the boiler and horizontal storage above that with living space even higher. This way the column of mixed water would be heavy enough to push the hot water out of the boiler and you'd still have return temperature protection.
 
I think a Varm could do off grid quite well - there are some circs out there that use very little electricity that I think could move water through a mixing valve quite well, and not take much juice from an off-grid power situation. The 15-58 three speed in my LK810 moves the heat away from my 40 very well on low speed, I've never switched it up. Those Alphas are sweet looking pumps that use a fraction of the electricity. I'm not sure if one would be a direct swap out for a 15-58 on an LK810 body, if not it could likely be utilized with a Danfoss valve. With my system, I can have full gassification going and full house heating with the LK 15-58 pump being the only thing using electricity. Then even at that, once everything is fully up to temps & system return water is above 140, it would function fine without even that running. So really if you could replace the 15-58 with an Alpha for loading purposes, it would only really need to run until system return water was up to the point it would stay above 140 - if everything else was setup for good convection thoughout, that is.
 
Status
Not open for further replies.