Optimum flow for in-slab heating

Santilal

Hi

I understand that when a given amount of heat energy is dissipated in a slab the flow rate determines the delta T that the fluid experiences from entering to leaving the slab. Does the flow rate also affect the amount of heat transferred to the slab. For instance, if I want to maximise the heat transfer to the slab, how do I determine the flow rate required. So by way of example if I want to dissipate 1000W of heat I could have 2l/min of water with a delta T of 7degrees C or 1.43l/min with a delta T of 1o degrees C. (Sorry about the metric units). But could I raise the heat dissipated to 1100W by increasing the flow to say 2.5l/min. How do you work out the optimum flow rate to acheive the highest heat transfer?

Rich_49

10* F (-12.2* C)

should be your Delta T in a slab floor application .  This will give you optimal heat transfer and keep the slab evenly heated across its' mass .   A 7* C Delta T would equal 45* F Delta T and a 1* C Delta T would be 35* F , are these equations correct or did you mean other ?

Zman

10

 10 degree delta is your target. The average temp is used to calculate output. The delta relates to the comfort or evenness of the heat. The actual flow needed really depends on the type of assembly and length of the loops.

Carl

Santilal

Thanks

Hi



I knew I hadn't worded my question very well. Let me try again. I have heard that the best heat transfer between two mediums in a heat exchanger occurs with turbulent flow. So I was wondering if this was the case for in-slab as well. So if you increase the flow rate so that it transitions from laminar flow to turbulent flow, maybe you increase the amount of heat being transferred.



My reason for this question is that I have ceramic tiles on my two most remote (8m distance) bathrooms and even with a Delta T of about 12*C, the slab barely gets warm. I thought that maybe I need to try to push more heat into the slab somehow. The house is two years old and has 40mm of polystyrene under the slab and is fully insulated to the building code. Other areas of in-slab appear to be working ok.



I hope I have clarified my question. Sometimes the words just don't flow.



Thanks again.



Cheers, Santilal

Rich_49

You

are moving the fluid too rapidly through the slab and not allowing for any transfer . forget turbulent flow with radiant heat , it does not matter , laminar is fine .  Turbulent flow is best practiced in other methods of heating . Slow the fluid down and your floors will heat up . You want a 10 Degree Fahrenheit Delta T . Possibly your supply water to the floor is not warm enough . All you will accomplish by putting more water through that panel is to get frustrated and make your equipment work real hard while accomplishing nothing .

    That being said , it appears that you are not moving enough water , but don't worry about turbulent . You want the Delta T to be negative 12*C (10*F) . This will allow the floor panel to heat evenly and heat up .   Where are you located ? What tyemperature water are you now pushing through the tubing in these panels ?

Jean-David Beyer

12*C (10*F)

Actually, 12C is 21.6F

Santilal

Inlet/Outlet Temperatures

Hi Rich



My installer told me that the inlet temperature to the In-Slab manifold was 30*C. I logged it the other day and I had the following range of inlet/outlet temperatures - all in *C: 25/21,32/20, 33/23, 40/23. This was measured over a two hour period. I am not sure why the variation, but there are other piping issues that could explain it.



So you guys recommend that the Delta T should be 10*F (5.6*C) - is that correct?



By the way I am live in Central Otago in New Zealand - the winter temperature varies from -7*C at night to say 8-10*C during the day - although this winter has been a little warmer.



Thanks for the help.



Santilal

hot_rod

a good design tool

for looking at how flow changes output was developed by Larry Drake and is available at the RPA website. The RadPad slide rule allows you to select tube size, flow rate, slab covering, etc and then shows output.



An example of a typical design of 300 feet of 1/2 pex at 12" on center



or a 91 meter loop of 12mm tube spaced 30cm, flowing at the RPA recommended .6 gpm (2.27 l) would output around 23 BTU per square foot



Increase the flow to 1 gpm (3.7 l) and the output would be 38 BTU/ sq ft.



Drop down to .2 gpm flow and you fall right off the chart, maybe a few btu/ sq at 12" spacing



There comes a point when heat transfer drops and drops quickly in a radiant loop, baseboard, panel rads any heat emitter really.



If the flow drops really low, say below .2 gpm (.75 l) the flow will tend to go laminar ( a boundry layer is formed in the tube that prevents heat transfer) and heat output will drop like a rock. When you plot that it is almost a straight line drop (hockey stick)



Keep an eye out for the August issue of PM mag, an excellent article on how flow relates to output in fin tube, and radiant loops, and what the delta T effect is on all that.



The design delta T for a system is a "target" that you design around. The actual delta T will move around. When a cold heat emitter first starts you will see a large delta, that will close up as the emitter warms or the room comes up to setpoint. It's not mandatory that you lock in that delta t for the system to perform efficiently and comfortably, and in some cases it is not desireable to "fix" the delta t.



Also there are also some limitations to how the universal BTU formula is applied which the PM mag article will address also.



Here is a link to a tech journal that helps explain the heat output pages 7-9

.

http://www.caleffi.us/en_US/caleffi/Details/Magazines/pdf/idronics_8_us.pdf

Zman

Speed and delta

First off. High mass slabs will have delta t's that are all over the map. From a cold start you can easily have 20-30 (f). That same slab might have a steady state in the 5-10 range. Don't worry about it.

As for flow rates, There is no such thing as moving the water too fast through a radiant tube. It is not as though the btu's are unable to jump off. Siggy has written some informative an funny stuff about this.

At the same time it is very difficult to move it too fast using common circs. The problem most people run into is that with long radiant loops the water will run too slowly  leaving you with cold areas and inadequate performance. If you are seeing steady state deltas of 5-15, you are probably fine.

The turbulent flow thing applies to heat exchanger design, not tubing.

Carl

hot_rod

why not tubing, Carl?

From what I understand the heat output of a radiant slab, which is transfered by the fluid flow in the tube, is indeed a function of the flow rate. More flow, more heat transfer.



This graph of a radiant in slab loop, shows when you approach those low, possibly liner flow velocities the output drops considerably and quickly.



I've read engineers will use the practice when running long pipe runs between buildings, underground for example. By designing for low flow laminar, the lose the least amount of heat thru the pipe, or tube wall. Page 206 of Modern Hydronic Heating Third Edition clearly shows this concept.

Zman

Diminishing Returns

Hot rod,

I guess I didn't word that very well.

My point is, once

you go beyond the 1 gpm mark on the chart you posted, is it really worth

the circulator energy required to generate more flow? Is the extra 500 btu transferred because of turbulence or is it because of the  warmer temps the return side of the loop sees? How much electricity does it take to get you there.

Carl

hot_rod

Agreed

the juice is not worth the squeeze, on the highest flow end.



I'm more concerned about the fractional GPM range and how slow, or low it too far. Turbulent flow is working from maybe .3 gpm on up, not just between 1 and 2 gpm, as the graph shows.



I built a large pumping demo with all clear plastic tubing, even built a couple clear plastic pump volutes. I can "watch" that turbulance across the range, and yes at excessive velocities it is a white cloud , more like a blender then a circulator.



It also clearly demonstrates that two directional flow in clear plastic closely spaced tees. also how velocity and air and dirt removal relate.

Zman

You Tube?

Hot Rod,

The transparent piping sounds interesting? Do you have a video?

One of the steam boiler companies did one to demonstrate piping designs. Very cool

Carl

Rich_49

Jean David

Let me ask a question of you !  If 0* C is 32* F how could you possibly deduce that 12* C (53.6*F) is equal to 10* ?  Hope you like your Ultra and 0 Delta Slab .

Rich_49

What other

piping issues exist that may account for the large difference in Deltas ?

Zman

Conversion

We are talking about temp differences not actual temp

A temperature difference of 1°F is the equivalent of a temperature difference 0.556°C

 A temperature difference of 1 deg C is the equivalent of a temperature difference 1.8°F.

Carl

Santilal

Rich

Hi Rich



Sorry I missed your post in the thread. You asked about what other piping issues I have. I have posted previously on this and received some helpful feedback and I am now in the process of refining my thoughts before going about fixing my system. It was installed by a plumber who was recommended to me as having good knowledge about heating systems. After researching for about two years, and close observation of how my system is performing, I am rapidly forming the opinion that my installers knowledge is wanting.



There are a number of issues: the boiler (conventional coal boiler) protection is not correctly piped, there are no closely spaced Tees anywhere, I have three pumps that are all pumping in series, the domestic water heat exchanger does not work if the radiators are going because it is starved of flow (because of the series pumps, I suspect), the in-slab also does not work with the radiators (for the same reason) and so the list goes on.



With the great help from this forum and others who have contacted me I am slowly building a plan to fix things. When I come to do the work, I will do it step by step as I can afford it. At present, I am just trying to get a good understanding of how it all works.



Thanks.



Santilal

Jean-David Beyer

If 0* C is 32* F how could you possibly deduce that 12* C (53.6*F) is equal to 10* ?

The post to which I was replying was about temperature differences. If there is a temperature difference of 10C, that is the same as a temperature difference of 18F. The 32F freezing point of water vs the 0C freezing point of water drops out of the difference calculations.



According to my calculations, a temperature of 12 C corresponds to a temperature of of 53.6F, as you say.



But if you measure the temperature at two points, it makes no difference what the freezing point of water is.



Say water supply is 150F and return is 130F. that difference is 20F,

If you do not say that a delta T of 20F is 11.111C, we can calculate it out by converting the temperatures first, and then subtracting them.



150F is 65.555C

130F is 54.444C

Gordy

metric delta t (difference in temps)

Jean and Rich,



Its pretty simple 5.55* C is 10* F.  The op is contemplating Delta Ts of 7* C (12.6F) vs 10* C (18*F).



Of course Carls numbers are more accurate. So lets not make the metric system complicated ( which its not). So he should be shooting for a 5*C (10* F) delta for more even slab temps. A floating target of course. So 2.5 L/Min seems appropriate for a flow rate.