Sizing Circulation Pump

[MortarThePoint]

I'm going with a slightly different setup to many by having a single circulation pump in the plant room rather than a pump on each manifold. I'll need to have a bypass in case all the valves suddenly close, but I presume the bypass is a pressure relief, so is normally closed and only opens when the pressure is too high and allows time for the pump to switch off.

 

Flow Rate: We're installing an 11.2kW Ecodan. A major input to the pump sizing will be the dT flow to return which I'll assume is dT = 5C. Based on that I can work our a flow rate as F = P_Heat / (Heat_Capacity * dT) = 11.2kW / (4.2kJ/kg.K * 5K) = 0.47kg/s = 1.92m3/h. 

 

Pressure (Head): This comes down to the amount of pipe work you have and isn't a function of how high you house is (water goes up and back down). The flow rate dictates the resistance to flow which is the head. The main losses that came to mind were loop, at manifold and manifold feed:

• Loop: I've taken the longest loop and take that as a fraction of the overall loop length of the system and it works out as 6% (yes, I have a silly number of loops). That means it will have a flow rate of 6%*1.92m3/h = 0.12m3/h = 0.032kg/s. I don't have data for 16mm UFH pipe but 15mm Hep2O is probably a good proxy and they publish that. In other considerations I worked fitted a line to that data which gives Head_Loss_KPA_Per_M = 79.9*(FLOW_IN_KG_PER_S^1.76) = 0.187 kPa/m. My longest loop is 110m long so that works out as a pressure drop across the loop of 110m * 0.15kPa/m = 20.6kPa = 2.1m head loss.
• At Manifold: due to valving etc: Guestimated as 1m head loss. I expect/hope this is an overestimate but would depend on how we have the balancing valves set etc.
• Manifold feed: I've taken the manifold with the most loops which is also the furthest on it and take that as a fraction of the overall loop length of the system and it works out as 48%. That means it will have a flow rate of 48%*1.92m3/h = 0.92m3/h = 0.256kg/s. For 28mm Hep2O Head_Loss_KPA_Per_M = 2.89*(FLOW_IN_KG_PER_S^1.76) = 0.263 kPa/m. This manifold is 36m round trip from the plant room (difficult routing) so that works out as a pressure drop across manifold feed and return 36m * 0.21kPa/m = 9.5kPa = 0.95m head loss.

The sum of those three head losses is 2.1m + 1m + 0.95m = 4.05m. There will likely be other losses, but hopefully the 'at manifold' figure is pessimistic. I think conventional wisdom is to go with 6m of head loss capability.

 

A Grundfos UPS2 25/80 (180) should work. The 25 number represents the fitting diameter, 80 represents the maximum head (in kPa) so 8m, and the 180 is the distance between fittings. Operating at 4m head and ~2m3/h should give an efficiency of around 45% which whilst not peak is near. That pump is more powerful and expensive (twice) than a Lowara ECOCIRC L 25-8/180. That pump doesn't have efficiency data though.

 

Oversizing the pump (e.g. Grundfos Magna1 32/80 (180) not shown below) could push it lower of its efficiency curve.

 

Grundfos UPS2 25/80 (180):

 

Lowara ECOCIRC L 25-8/180:

 

[JohnMo]

Why not just use the circulation pump within the ASHP, or doesn't it come with one?

 

If you have a radiator or UFH heating loop without actuators do you need a bypass?

 

Your flow rate is basically double mine - at 6kW, so sounds right.

 

If you use an 8m pump, you have room to change speeds to get the flow rate you need, but heat pumps should have all you need and that should auto modulate to maintain delta T target.

[JohnMo]

What I forgot to add, if you are using a buffer, plumb as a 2 port in the return or flow line, then you won't need a secondary pump.

[MortarThePoint]

9 hours ago, JohnMo said:

Why not just use the circulation pump within the ASHP, or doesn't it come with one?

 

If you have a radiator or UFH heating loop without actuators do you need a bypass?

 

Your flow rate is basically double mine - at 6kW, so sounds right.

 

If you use an 8m pump, you have room to change speeds to get the flow rate you need, but heat pumps should have all you need and that should auto modulate to maintain delta T target.

 

ASHP kit has come with two LOWARA 25-8/130 pumps and the supplier's schematic shows these on the flow and return pipes to the actual ASHP itself (from the low loss header). The schematic then shows a pump on each manifold feed pipe after the low loss header. These pumps I am trying to common up into one pump and I'll need to balance the valves at the manifolds to achieve this.

 

I reminded myself of the no actuators thread earlier and it's a good one. I think I'll be trying to do something like that , keeping the overall flow the constant and varying the flow temperature

[Nickfromwales]

1 hour ago, MortarThePoint said:

ASHP kit has come with two LOWARA 25-8/130 pumps and the supplier's schematic shows these on the flow and return pipes to the actual ASHP itself (from the low loss header)

Crikey, is the ASHP in the next village?

 

1 hour ago, MortarThePoint said:

These pumps I am trying to common up into one pump and I'll need to balance the valves at the manifolds to achieve this.

Are they on the same floor of the house, or GF & FF?

[MortarThePoint]

13 minutes ago, Nickfromwales said:

Crikey, is the ASHP in the next village?

 

Are they on the same floor of the house, or GF & FF?

Different floors, but I am hoping that doesn't matter

[Nickfromwales]

7 minutes ago, MortarThePoint said:

Different floors, but I am hoping that doesn't matter

What are the emitters upstairs? 

[MortarThePoint]

9 hours ago, Nickfromwales said:

What are the emitters upstairs? 

It's all UFH. 12 loops upstairs and 13 downstairs.

[TonyT]

I would stick with the supplier ASHP pump

 

 

 

 

[Nickfromwales]

3 hours ago, TonyT said:

I would stick with the supplier ASHP pump

 

 

 

 

Which one(s)?

[MortarThePoint]

4 hours ago, TonyT said:

I would stick with the supplier ASHP pump

I'll use the pumps they have supplied for the ASHP side of the low loss header, but I need a pump for the UFH side of the low loss header.

[TonyT]

2 hours ago, Nickfromwales said:

Which one(s)?

Any erp pump, generally 7m head and on a new build such as this pipes should be designed to minimise loss.

 

not a fan of loads of pumps, primary and secondary etc unless it’s massive.

it all ends up over complicating the controls which for most are already complicated enough.

 

[MortarThePoint]

The Grunfos UPS2 does pretty much what I want in terms of pressure / flow control (see below). When an UHF loop shuts off with an old style pump, it will maintain the same flow rate by upping the pressure which isn't what you want as it will increase the flow in the other loops. I want the differential pressure at the UFH manifolds to remain constant. If there was no head loss in the manifold feed/return lines, that would mean I wanted a pump that has a constant pressure output. However, since there is head loss in the manifold feed/return pipes, the pump needs to make a slight reduction in output pressure. I guess that is what UPS2's proportional-pressure control is intended to achieve, but I think it may be a little bit too sloped.

 

In my example in the first post, there was 3.1m of head loss associated with the loop and its feed valves and 1m of head loss due to the manifold feed/return pipes at the furthest manifold. If half the loops shut off, manifold feed flow rate would half and the pressure drop in those pipes would go down to about 0.3m head loss. I have probably overestimated the 'loss at manifold' as it makes sense for that to be near zero for the longest loop and higher for shorter loops to try to balance. If it is zero, then I would need a proportional-pressure control curve which was linear from (0m3/h, 2.1m) to (2m3/h, 3.1m). PP1 looks to go from (0m3/h, 2.1m) to (2m3/h, 4m), so I guess that would increase the flow slightly in loops when valves shut off. It gets complicated by the different flows and lengths to different manifolds, but hopefully not too much. I have one manifold (3 ports connected) that is in the plant room, so has very short feed/return pipes. This feels better served by 22mm pipes and perhaps a manual orifice/valve.

 

Even if the manifold differential pressure changes by 50%, an unrestricted loops flow would change by 26% [1.5^(1/1.76) = 1.26] so power delivery would rise by around that same figure, perhaps 30%. Orifice pressure vs flow relation is similar to pipe, so OK modelled by the 1.76 power.

 

Of course, if I go the @JohnMo route of having no actuators, it will be balance and forget. Very tempting! I watched a good Heat Geek video about how shutting off a zone can mean that you up the heat demand in the still active zones (due to uninsulated internal walls), which ups the flow temperature, lowers the COP and actually costs more. Many variables, like house geometry, but it does make sense.

 

[TonyT]

I use both pumps on different installations from the original installer but I like the eco circ plus version that has an led display as it shows you head and flow data

[JohnMo]

26 minutes ago, TonyT said:

not a fan of loads of pumps,

Me neither. My last iteration of my heating system had 3 pump, boiler, buffer and UFH, all ran most of the time during the heating season, £128 to run at today's electric price for 180 days. With the heat pump I have one (although DHW still has three, but it only runs one hour per day) so heating season pumping costs are £45 including DHW.

[MortarThePoint]

The Lowara pumps actually have some additional goodies over the Grunfos:

• Constant pressure mode (modes 1, 2 3) as well as proportional-pressure (A, B, C) and constant speed modes like the Grunfos (I, II, III)
• Option to have a display and Bluetooth (adds about £30 ex VAT)
• Best-in-class energy efficiency (EEI ≤0.18) vs Grunfos EEI≤0.23
• Both state Sound level ≤ 43 dB(A)

Lowara are also about half the price of the Grunfos. So even if they don't last as long, they could be a better option.

 

[JohnMo]

21 hours ago, MortarThePoint said:

watched a good Heat Geek video

There is another one on balancing pumps across a low loss header and how if you get the sizing/flow rates wrong, it decreases CoP also.

 

If you go no actuators - really think if you need a low loss header? Or just plumb it in as a two port volumiser/buffer on the return leg of the heating system. Or if you need for warranty purposes install a 3 port valve so you can bypass around it. Set the logic so it also switches of the secondary pump also. A Simple switch power on, starts pump powers valve to route flow through LLH. Power off bypass LLH.

[TonyT]

17 minutes ago, MortarThePoint said:

The Lowara pumps actually have some additional goodies over the Grunfos:

• Constant pressure mode (modes 1, 2 3) as well as proportional-pressure (A, B, C) and constant speed modes like the Grunfos (I, II, III)
• Option to have a display and Bluetooth (adds about £30 ex VAT)
• Best-in-class energy efficiency (EEI ≤0.18) vs Grunfos EEI≤0.23
• Both state Sound level ≤ 43 dB(A)

Lowara are also about half the price of the Grunfos. So even if they don't last as long, they could be a better option.

 

Yes that the pump I recommended in my last post.  Value for money and it gives data for efficient running/commissioning 

[dpmiller]

why not Wilo?

[TonyT]

[Nickfromwales]

17 hours ago, TonyT said:

That's my afternoon gone  

[MortarThePoint]

I kin of expect my operating point to be at a head of 3m which is better suited to the Ecocirc M (3m constant pressure setting) than the Ecorcirc L (2m & 4m constant pressure settings]. However, 2m3/h would put it right near the maximum flow rate for the Ecocirc M at 3m head.

 

Running the pressure 33% higher (at 4m rather than 3m) would put up the flow rate by 18% [1.33^(1/1.76)=1.18] and the hydraulic power by 57%, probably the electricity consumption too.

 

 

Edited July 10, 2023 by MortarThePoint

[MortarThePoint]

Both pumps would be around 50% efficient. The Ecocoirc-L running at 4m and 2.31m3/h would be 50W of shaft power which I guess is the same as electrical input. The Ecocirc-M at 3m and 2m3/h would be 33W of shaft power. If that is the electrical powers too, then using the Ecocirc-M would save 149kWh/yr = ~£50/yr. It's running closer to its curve limit. Not sure what that means in terms of lifetime.

 

[MortarThePoint]

@JohnMo When you are running less than peak power, the flow temperature goes down, but does the flow rate go down too? I guess it can go down since the water dT between room and water is lower and so to get the same flow vs return temperature difference, the flow would need to slow down.

 

I expect to be optimal, you want the dT of at the ASHP to be the same as the dT at the UFH manifold. That wouldn't be the case though since I think the ASHP circulation pumps will run at constant speed.

Edited July 10, 2023 by MortarThePoint

[Nickfromwales]

49 minutes ago, MortarThePoint said:

That wouldn't be the case though since I think the ASHP circulation pumps will run at constant speed.

Check with the supplier?