Hot water (DHW) circulation

[MortarThePoint]

It's nice to have hot water straightaway when you turn a hot tap on. To achieve this, I think some have a flow and return to the basin in question and hot water is circulated to it when the bathroom light is turned on or a PIR is tripped. Isn't this hugely wasteful?

 

Approximate calculation:

Water circulation flow rate, q = 12l/min = 0.2l/s (guess)

Hot water temperature, T_w = 55C

Room temperature, T_r = 25

Drop in water temperature flow vs return, dT = 5C (guess, insulated but still)

Pipe ID, d = 15mm

Pipe length to basin, L = 10m

Volume of pipe to basin, v = pi * 10 * (0.015/2)^2 = 1.77l --> mass of water m_w = 1.77kg

 

Model 0: system with only a flow. Water heated up and then allowed to cool down before next use some hours later

Energy wasted, E1 = (T_w - T_r) * m_w * c_w = (55C - 25C) * 1.77kg * 4.2kJ/kgC = 223kJ = 0.06kWhr --> cost 1.2p.

Potentially 1.77l of water wasted before hot water starts coming through, cost 0.18p.

 

Model 1a: flow and return water heated up and then allowed to cool down before next use some hours later

Energy wasted, E1 = (T_w - T_r) * (2 * m_w) * c_w = (55C - 25C) * (2 * 1.77kg) * 4.2kJ/kgC = 446kJ = 0.12kWhr --> cost 2.5p (0.04kWhr & 0.8p if ASHP)

No water wasted before hot water starts coming through

 

Model 1b: flow and return circulated for 5 minutes (=300s) each time loo occupied

Energy wasted, E2 = dT * q * t_occupied * c_w = 5C * 0.2kg/s * 300s * 4.2kJ/kgC = 1260kJ = 0.35kWhr --> cost 7p (0.12kWhr & 2.3p if ASHP) [equivalent to 4.2kW for the 5mins]

No water wasted before hot water starts coming through

 

With a hot flow and return system it would be approximately the sum of 1a and 1b, so about 10p for each time the loo is used. It would be less wasteful (of energy and money, but not water) to automatically run the hot tap until it is warm each time, ready for instant hot water. But that would be seen as an awful waste. Why isn't a flow and return system seen as the same.

 

Annual cost based on loos used 4 times a day by 4 people:

    365 * (4 * 4) * (0.12kWhr + 0.35kWhr) = 2,745kWhr (900kWhr)

    365 * (4 * 4) * 10p = £584/yr (£200/yr if ASHP)

 

Wouldn't be a waste in winter as it would heat the house, but that may be only half of the year. So best case you're still looking at wasting about 500kwhr and £100 per year.

 

Have I misunderstood how these systems work?

Edited February 14, 2022 by MortarThePoint

[MortarThePoint]

It would be less wasteful to have a sensor that shuts off the circulation when hot water is detected in the return pipe back at the tank. That would limit the damage to Model 1a, so around 200kWhr/year (elec).

Edited February 14, 2022 by MortarThePoint

[willbish]

Such a system is best combined with a building management system.

My circulation pump will be controlled by Loxone. It will only run for X seconds, the length of time it takes hot water to reach furthest point.

Activated by presence detectors and then pump will only run max Y times per hour. Thinking 2 or 3 times.

 

 

[Carrerahill]

Is it such a problem waiting a few seconds?

Edited February 14, 2022 by Carrerahill

[Nickfromwales]

How much do you spend per year on recreation, alcohol, hobbies etc?

It’s not about cost, it’s about value. If you really value instant hand washing etc, and I fit hot returns on nearly every clients new build, then just pay and move onto something else. This will use energy, either way it gets used, and you’re only arguing over the difference in the cost of different methods of execution, so less than £50 per year most likely   
The system costs are significant so focus on that, plus also ensure, as I do, that the HRC pipes are ridiculously well insulated to reduce losses. I put the 15mm flow and the 10mm return inside the same piece of 25mm wall x 22mm internal pipe insulation. Completely cold to the touch with 50oC water circulating inside. 

[willbish]

1 minute ago, Carrerahill said:

Is it such a problem waiting a few seconds?

Seems to me to be even more wasteful if you run the hot tap, wash hands in cold water and then by the time the hot reaches the outlet it is shut off and large dead leg left to go cold 

[MortarThePoint]

3 minutes ago, willbish said:

Seems to me to be even more wasteful if you run the hot tap, wash hands in cold water and then by the time the hot reaches the outlet it is shut off and large dead leg left to go cold 

 

It's half as wasteful as having flow and return cooling down.

[MortarThePoint]

5 minutes ago, Nickfromwales said:

How much do you spend per year on recreation, alcohol, hobbies etc?

It’s not about cost, it’s about value. If you really value instant hand washing etc, and I fit hot returns on nearly every clients new build, then just pay and move onto something else. This will use energy, either way it gets used, and you’re only arguing over the difference in the cost of different methods of execution, so less than £50 per year most likely   
The system costs are significant so focus on that, plus also ensure, as I do, that the HRC pipes are ridiculously well insulated to reduce losses. I put the 15mm flow and the 10mm return inside the same piece of 25mm wall x 22mm internal pipe insulation. Completely cold to the touch with 50oC water circulating inside. 

 

I've always been happy to wash hands with cold water (using the cold tap), but that's probably frowned on in the post Covid 20 seconds world.

 

As our house is laid out, we only have one basin at a significant distance. I had wondered about having a thermostated redring style water heater to heat the first bit coming through.

[ProDave]

If this is ONLY for a wash hand basin, would a small 3Kw instant water heater very close to the basin not make sense?  It only heats what is used.

 

For a larger use like a shower, a dead leg is not such an issue.  I turn the shower on and then in the time it takes to undress the hot water has arrived.

[JohnMo]

Not sure the maths is correct for model 1b

 

The model may be correct if the circulation water went in the system hot and was returned cold, but if well insulated the losses during circulation would be quite small.

 

For example I am circulating via a combi, so I can see when the boiler fires.  On start of the circulation pump the boiler fires, within a few seconds it turns off.

 

I am running a 30m loop for a couple of hours morning and evening, based on the maths I would be using around 300kWh a day.  I am not.

[MortarThePoint]

48 minutes ago, JohnMo said:

Not sure the maths is correct for model 1b

 

The model may be correct if the circulation water went in the system hot and was returned cold, but if well insulated the losses during circulation would be quite small.

 

For example I am circulating via a combi, so I can see when the boiler fires.  On start of the circulation pump the boiler fires, within a few seconds it turns off.

 

I am running a 30m loop for a couple of hours morning and evening, based on the maths I would be using around 300kWh a day.  I am not.

 

Thanks JohnMo! I was surprised by the numbers that came out, so hoped someone would shoot a hole. I have assumed a 5C temperature loss, but that may be too much. Thinking again, the water is only in the pipe for 2*1.77l / 12l/min = 18 seconds. An online calculator [1] suggests 8W/m based on 55C/25C/15mm pipe/15mm insulation which would make for 20m * 8W/m = 160W which is 4% of the previous figure so more like it. 8W/m from 0.177l/m for 18seconds works out as a dT of (8*18)/(4200*0.117) = 0.2C. So my guess was WAY out.

 

That makes the model 1b contribution much smaller:

Model 1b: flow and return circulated for 5 minutes (=300s) each time loo occupied

Energy wasted, E2 = dT * q * t_occupied * c_w = 0.2C * 0.2kg/s * 300s * 4.2kJ/kgC = 50kJ = 0.014kWhr --> cost 0.3p (0.005kWhr & 0.1p if ASHP) [equivalent to 160W for the 5mins]

No water wasted before hot water starts coming through

 

Checking Model 1a:

dT/dt = 0.2C / 18s = 0.67C/min. That means it would take about an hour to get back to 25C as it's not linear. That seems OK for model 1a.

 

So Model 1b doesn't account for much and it's the difference between 2.5p and 1.2p per loo usage of Model 1a and Model 0 that counts. That makes the annual difference:

    365 * (4 * 4) * ((2.5p + 0.3p) - (1.2p + 0.18p)) = £83/yr (£21/yr if ASHP)

Much more palatable.

 

[1] https://cheguide.com/heat_loss_insulation.html

Edited February 14, 2022 by MortarThePoint

[Nickfromwales]

Too much maths here…… Just enjoy the hot water and have a beer. If you have PV it’ll run practically for free all summer. Discount that and you’re spending very little for what I think is a super luxury addition.  Waiting for hot water annoys me, and mine is from a combi 3m away from each basin!  Counting to 5 is 5 seconds too long. 

[MortarThePoint]

A rather hacked solution which wastes about the same amount of energy is to fill the cistern from the hot water supply. Problem is it probably wouldn't fill before you reached the basin.?

Edited February 15, 2022 by MortarThePoint

[MortarThePoint]

16 hours ago, willbish said:

Such a system is best combined with a building management system.

My circulation pump will be controlled by Loxone. It will only run for X seconds, the length of time it takes hot water to reach furthest point.

Activated by presence detectors and then pump will only run max Y times per hour. Thinking 2 or 3 times.

 

If you were to harvest grey water for toilet flush, then an alternative simpler and potentially more energy efficient approach would be to have a valve near the basin to 'waste' that opens when presence detected and closes when hot water detected (or after set time). The water wouldn't be totally wasted and there's no need for a return pipe.

[Dan F]

16 hours ago, MortarThePoint said:

Water circulation flow rate, q = 12l/min = 0.2l/s (guess)

Most basin taps, ones I've seen anyway, have limits around 6 l/min.

 

Quote

Pipe ID, d = 15mm

15mm copper, which is max you'd use, is less than 15mm ID.  Why not run calcs with 10mm copper or MLP too though,  you don't really need 15mm for basins.  I used 12mm uponor MLP which has the same ID (8.8mm) as 10mm copper.

 

Quote

Model 1b: flow and return circulated for 5 minutes (=300s) each time loo occupied

Where does 300s come from? Surely you don't need that to flush 1.7l?

Edited February 15, 2022 by Dan F

[TonyT]

Serious overthinking of a problem  that’s not a problem.

[MortarThePoint]

49 minutes ago, Dan F said:

Most basin taps, ones I've seen anyway, have limits around 6 l/min.

 

15mm copper, which is max you'd use, is less than 15mm ID.  Why not run calcs with 10mm copper or MLP too though,  you don't really need 15mm for basins.  I used 12mm uponor MLP which has the same ID (8.8mm) as 10mm copper.

 

Where does 300s come from? Surely you don't need that to flush 1.7l?

 

Narrower pipe definitely helps so probably takes it to a trivial cost.

 

Volume of pipe to basin, v = pi * 10 * (0.0088/2)^2 = 0.6l --> mass of water m_w = 0.6kg

So about a third of the water volume, so all energy costs go down by 3.

    365 * (4 * 4) * ((2.5p/3 + 0.3p/3) - (1.2p/3 + 0.18p/3)) = £28/yr without ASHP

    365 * (4 * 4) * ((2.5p/3 + 0.3p/3) - (1.2p/3 + 0.18p/3)) = £22/yr without ASHP and just priming

Haven't included the pump power in any of this, but likely small if shuts off after only enough time to prime the pipes.

 

The 300s is 5mins average occupancy and was naively used for the duration of hot water circulation. It was a guess, but Model 1b doesn't account for much anyway after the correction prompted by JohnMo's comment. willbish's comment pretty much eliminates it and, if set up correctly, potentially the whole difference.

Edited February 15, 2022 by MortarThePoint
corrected that the volume of water otherwise wasted is smaller too

[MortarThePoint]

9 minutes ago, TonyT said:

Serious overthinking of a problem  that’s not a problem.

 

Valid, I wanted to qualify that it wasn't a problem. >£100/yr is possibly decadent for what it is, <£20/yr feels OK.