Improving heat transfer in domestic hot water tank?

[sharpener]

TL;DR Summary

Will circulating the water in a hot water tank improve the rate of heat transfer from the heating coil?

A typical UK domestic hot water tank holds c. 200 litres of water. Hitherto the common method of heating it to say 60C has been by means of an internal coil of copper pipe through which is pumped hot water from a gas boiler at ~ 70C. Heat transfer on the outside of the coil is by natural convection in the tank.

The advent of heat pumps has resulted in lower available temperatures and so a widespread recommendation is to use special "heat pump" hot water tanks which have a coil with 3 to 4 x the surface area. This results in many heat pump installers insisting on replacing tanks that are otherwise perfectly sound.

It occurs to me that the heat transfer might be improved by circulating the body of water in the tank to replace the heated water in contact with the coil with fresh colder water more quickly than can be achieved by convection alone. This could be done by adding a small external pump which would cost only a fraction of replacing the whole tank.

Can someone tell me if this idea would give a worthwhile improvement and say what factor it might be, or point me to a calculation method or empirical heat transfer coefficient I can use for this?

TIA

 

[jrmichler]

What is needed here is a heat transfer coefficient for free convection, then another heat transfer coefficient for forced convection. My first attempt used search terms heat transfer coefficient free convection. One of the hits had the following:

• Free Convection - water and liquids: 50 - 3000 (W/(m2K))
• Forced Convection - water and liquids: 50 - 10000 (W/(m2K))

This is typical of what you find when you search for heat transfer coefficients. There is so much variation in heat transfer that usable numbers do not exist. It gets worse. The total heat transfer includes the heat transfer coefficient from the thermal fluid to the heating coil, the heat transfer resistance of the heating coil, plus the heat transfer coefficient from the heating coil to the water in the tank.

In order to make a coil work that is 3 to 4 times smaller, you need 3 to 4 times higher overall heat transfer. Since there are three resistances to heat flow in series, and you only have control over one of them, that heat transfer coefficient needs to be increased by about 5 to 10 times. There are equations for estimating heat transfer coefficients. Those are found in textbooks with Heat Transfer in the title.

Or you could run a simple test. Get a piece of copper tube about the diameter of the tubing in the heat transfer coil, and solder a cap on one end. Drop the sensor from an indoor/outdoor thermometer into the tube and fill it with water. Fill a bucket with hot water. Insert the tube into the water, and measure the rate of temperature rise in the tube. Do this while holding the tube still, then while swishing it back and forth. This test measures the rate of heat transfer of the total system. It also gives you a good idea of how fast you would need to circulate the water in order to get the water filled tube to heat up 3 to 4 times faster.

 

[Lnewqban]

It is possible that the new heat pump requires greater surfaces of condensation for its refrigerant, in order to keep the efficiency value that the manufacturer publishes and/or is required by local building-energy codes.
The pump that you have thought of uses additional energy, which reduces the mentioned efficiency of the heat pump whole system.
I recommend consulting the manufacturer's installation instructions and local energy code.

 

[russ_watters]

While the prior two posts do directly answer the question in the OP, there is perhaps a more relevant answer: The heat transfer from the coil to the water is 100% efficient and there is nothing you can do to improve it.

That isn't to say you can't increase the heat transfer, but any increase in heat transfer will exactly coincide with an increase in the heat input required by whatever heat source you proved. So ultimately the heat transfer effectiveness of the coil doesn't matter.
Edit: incorrect, please disregard.

 

[Rive]

It occurs to me that the heat transfer might be improved by circulating the body of water in the tank to replace the heated water in contact with the coil with fresh colder water more quickly than can be achieved by convection alone. This could be done by adding a small external pump which would cost only a fraction of replacing the whole tank.

The water in the tank is not homogenous in temperature. Hot water is on top: the outgoing pipe is connected here, so the whole capacity of hot water could be extracted. The cold water remains on the bottom, were the incoming pipe connects.

The heating element is at the bottom, so it is always the cold water which gets the heat.

If you stir this up, then two things will happen. First, the half full will mean 200l half-heated water, instead of 100l of hot water.
Second: the heating element will transfer the heat to warmer water, resulting of higher return temperature on the coil: thus degrading the efficiency of the connected heat pump.

So while it is true that in general moving water would allow the use of somewhat smaller coil, in the specific case of water tanks this does not works well.

 

[jrmichler]

That isn't to say you can't increase the heat transfer, but any increase in heat transfer will exactly coincide with an increase in the heat input required by whatever heat source you proved.

Well, no. An increase in heat transfer coefficient allows the heating source to heat the water with a smaller temperature difference. The total amount of heat transfer stays the same because the amount of water to be heated stays the same. Heat pump efficiency requires delivering heat at the lowest practical temperature.

I made use of this when I designed the heating system for my house. I wanted one gas burning boiler to heat the house, the shop, and the domestic hot water. I wanted that boiler to operate at low temperature for efficiency. The system has a boiler to heat water, then water to air heat exchangers for the house and shop. Those heat exchangers are sized for 120 deg F water, so are larger than heat exchangers sized for a more typical 180 deg F hydronic system.

 

[erobz]

So while it is true that in general moving water would allow the use of somewhat smaller coil, in the specific case of water tanks this does not works well.

To what extent you could use a smaller coil is not really clear as the factors $h$ and $( T_{coil} - T_{tank} )$ are competing in this system.

 

[russ_watters]

Well, no. An increase in heat transfer coefficient allows the heating source to heat the water with a smaller temperature difference.

You're right, I'm not sure what I was thinking there (maybe skimmed and missed that this was about heat pumps?). Apologies, OP please disregard.

 

[hutchphd]

The water in the tank is not homogenous in temperature. Hot water is on top: the outgoing pipe is connected here, so the whole capacity of hot water could be extracted. The cold water remains on the bottom, were the incoming pipe connects.

Yes. I only recently discovered that this is one reason why a standing 40 gallon resistive electric water heater behaves "larger" than the same volume natural gas one. Neither system can produce hot water "on demand" sufficient for most uses. Separate resistive upper and lower heating elements can facilitate maintenance of the vertical thermal gradient so that the "electric shower" receives warmer water longer into the use cycle. Duh!!

 

[sharpener]

Or you could run a simple test. Get a piece of copper tube about the diameter of the tubing in the heat transfer coil, and solder a cap on one end. Drop the sensor from an indoor/outdoor thermometer into the tube and fill it with water. Fill a bucket with hot water. Insert the tube into the water, and measure the rate of temperature rise in the tube. Do this while holding the tube still, then while swishing it back and forth. This test measures the rate of heat transfer of the total system. It also gives you a good idea of how fast you would need to circulate the water in order to get the water filled tube to heat up 3 to 4 times faster.

OK so the result is useful. Without any agitation the initial rate of rise inside the tube is about 6C/min and with agitation it is about 10. Without agitation the temp is still rising slowly after 10 mins and with agitation it reaches a plateau at 7. While the swishing motion is effective at mixing the water in the bucket it is not clear to me if it is effective at mixing the water in the comparatively narrow tube.

Conclusion: the heat transfer is improved by a factor of at least 1.5 and it might be more if the regime on the inside of the tube were continuously flowing as in the real-life situation.

Many thanks for all the help.

 

[sharpener]

The water in the tank is not homogenous in temperature. Hot water is on top: the outgoing pipe is connected here, so the whole capacity of hot water could be extracted. The cold water remains on the bottom, were the incoming pipe connects.

The heating element is at the bottom, so it is always the cold water which gets the heat.

If you stir this up, then two things will happen. First, the half full will mean 200l half-heated water, instead of 100l of hot water.
Second: the heating element will transfer the heat to warmer water, resulting of higher return temperature on the coil: thus degrading the efficiency of the connected heat pump.

So while it is true that in general moving water would allow the use of somewhat smaller coil, in the specific case of water tanks this does not works well.

Maybe the answer is to connect the pump between two ports on opposites sides of the tank but both of them at or near the bottom. The aim being to stir the cold water in the region of the coil without disturbing the stratification too much. Most of the tanks I have owned have in addition to the main tappings a drain cock at the base which could be used for this.

Obviously you could have a more complicated flow scheme, but the idea is to do it as a retrofit modification for existing systems to avoid the cost and disrupton of having to change the tank altogether.

 

[Rive]

...the idea is to do it as a retrofit modification for existing systems...

Igot that, but in my book that means that the modification of the insulation and tank wall is not really welcomed => only with the original outlets.

 

[sharpener]

only with the original outlets

Yes. Perhaps the cold inlet port and the drain cock tapping would be sufficient though?

 

[Rive]

Depends on the placement, I think. The coil is usually near the wall, as far as I know: if those pipes are in the center, then circulating through them won't have much use, unless you pull some tricks inside too...

 

[TonyStewart]

Circulation will reduce the hysteresis in the heat cycle while increasing the reheat ramp rate as well as accelerate the cooling rate of temperature.

If the heater is on 100% during steady tank consumption rate, I do not like the effect of circulation during ramping down, but there may be benefit ramping up. or maybe not. The heat will rise to the top where the tank exit is, so convection circulation may heat up fresh water in a cold tank faster with convection if the consumption rate is slow.

 

[sharpener]

The circulation pump would only run when the heat pump is heating the tank. I don't fully understand the other points you are trying to make, sorry, it reads to me like a google translation from another language (perhaps it is?).

As several posters have mentioned, there are drawbacks to circulating the water in this way, notably that the water at the top of the tank might become colder because of de-stratification before it is heated again.

In this respect using the pump is no worse than the alternatives that have been proposed which are
i) fitting a Mixergy tank. These are expensive and the "heat pump" add-on kit has an external plate heat exchanger and pump which will also stir the contents of the tank (thereby eliminating the claimed advantage of only heating the water you want to use), and
ii) fitting the same arrangement of PHE and pump to an existing conventional tank which will be a lot cheaper and easier to do, but still more complicated and more expensive than the pump on its own.

In my situation I intend to heat the hot water tank with the heat pump at two specific times of day:

i) in the afternoon when there is most likely to be free surplus PV electricity and no demand for h.w., and
ii) just before the end of the Economy 7 cheap rate at 0700, when there is no demand for h.w. either.

So any cooling of the water available for immediate draw-off is not likely to be noticed.

Hopefully the point about running the heat pump at the lowest possible temperature will be addressed by restricting the circulating volume to the bottom part of the tank as proposed upthread. Or arranging the pump so it draws from the bottom and feeds back into the top, so that the bulk movement in the cylinder is downwards and brings the coolest water into contact with the coil. This would require the pump flow to be matched to the heating rate; a small enough pump for potable hot water might not be readily available as it would only require a flow in the region of 180 litres/h or 3 l/min.