wet UFH - ashp fast cycling

[markocosic]

Low loss headers shouldn't be used with heat pumps. 4-pipe buffer vessels also shouldn't be used with heat pumps. They destroy efficiency in operation. (and on condensing boilers for that matter)

 

 

You want: heat pump > 2-pipe buffer > emitter (UFH / radiators) > heat pump.

 

If the emitters have a large enough output or thermal mass you can lose the 2-pipe buffer.

 

 

A quick and dirty bodge world be to introduce a 2-pipe buffer on they output of the heat pump before it goes to your low loss header/4-pipe buffer.

 

This will give enough (effective) system volume to avoid cycling.

 

 

The short cycle issue you have is that your (effective) system volume is just the heat pump and pipework up to the low loss header. The thermal mass in the UFH / radiators has been decoupled by the header and your data / observations show this nicely!

 

 

Awful system design from an OEM that ought to know better...but wanted to avoid getting sucked into explaining flowrates and design on the 'emitter' side by decoupling using the low loss header.

 

 

Bash in a 2-pipe buffer between ashp and header on the flow to add in some thermal mass quickly.

 

Work out how to design UFH and radiators that can deal with the full flowrate the heat pump requires later. Crummy units throw a hissy fit at anything other than constant flow. That might be why the header is there...

 

 

 

 

 

 

[SteamyTea]

3 hours ago, markocosic said:

thermal mass

Umm.

What are the SI derived units for this?

[markocosic]

Joules per kg per degree Kelvin.

 

E = m (CP) dT 

 

https://en.wikipedia.org/wiki/Specific_heat_capacity

 

Or:

 

 

100 litres gives ~2.5 kWh for a 20C temperature rise. 

 

Many heat pumps ask for a minimum (directly fed) system volume to use as thermal mass to minimise cycling. This can be substituted for screed to a degree. The thermal mass of radiators and pipework can be ignored here. (minimal compared with the water)

 

[markocosic]

When it comes to removing the header entirely. You say:

 

"The Ecodan documentation says that the flow rate for the heating primary for a 14kw Ecodan is between 17.9 and 40.1 L/min"

 

Can you provide the manual?

 

I suspect this means that if you want a space heating system to work with an ecodan without the header then the flowrate needs to be between 18 and 40 litres per minute at ALL times that the heating is on.

 

Your smallest zone needs to be 18 litres/min (1.08 M3/HR) and your system still needs to function (remain adequately balanced at 40 litres/min (2.4 M3/HR) with all zones open. And the pump(s) need to provide enough pressure head to drive this flowrate thorough not just the heat pump but also the pipework and emitters.

 

Your non short cycling problem (with both zones on) is a secondary flowrate exceeding the primary flowrate.

 

Your header is being "circulated" by the radiator circuit. It draws hot water from the heat pump from the top AND because the flowrate for the secondary exceeds the primary it is also drawing return water from the radiator circuit up through the header.

 

No amount of baffling etc will correct this.

 

You need to reduce the flowrate on the secondary circuit to be equal to or less than the flow from the ecodan. That's easier. Limit flowrate through the radiators and you're done. No getting wet.

 

I would set UFH to 18l/min, DHW to 18 l/min, and radiators to 22l/mum (max, with all TRVs full open)

 

Now you can run:

 

- DHW only

- UFH only 

- UFH and DHW

- UFH and any amount of radiator up to 100% open 

 

Whilst remaining within the (rather limited) operating range of this heat pump. (with poor installation instructions and a worse dealer/installer network)

 

No to a low loss header. Yes to 2-pipe buffer direct on heat pump output to increase system volume. (a 100-200L unvented cylinder with the heat pump feeding the cold water inlet and the hot water outlet feeding the heating will do)

 

 

 

 

 

 

 

[SteamyTea]

1 hour ago, markocosic said:

Joules per kg per degree Kelvin.

Which is, as you say, Specific Heat Capacity.

But that is not 'thermal mass'.

 

Try again.

[markocosic]

You pedant! ?

 

 

Thermal mass = joules per degree K difference 

 

Thermal mass = (the mass that you have) * (specific heat capacity of the material)

 

 

To avoid short cycling when the output of the emitters is less than the minimum output of the heat pump; the heat pump requires enough directly coupled thermal mass to have adequate cycle times.

 

The dumb ass header design decouples the thermal mass of the UFH from the heat pump, so to introduce mass you need to put a 2-pipe buffer on the output of the heat pump before the header. 100-200 litres.

 

If you eliminate the header and pump everything directly from the heat pump then you may not need the buffer, or if you do it can be smaller and between the heat pump and the UFH such that you're not heading it when you only prepare DHW.

 

The OEM is being very cynical in their installation recommendations (that design can only ever work on paper) and the application support from their dealers/distributors is rubbish in the UK. I'd go against their advice.

 

 

Graham has some decent direct connect schematics:

 

https://silo.tips/download/schematic-diagrams-of-systems-using-samsung-ehs

 

Don't use his district / multiple pump designs with hearts without specialist advice. Observe minimum flowrate and system volume requirements of heat pump. You may require additional circulation pump in series with the one in the heat pump to meet system head requirements.

 

(and I'd buy from them when the Mitsubishi dies of you need design support)

[SteamyTea]

12 minutes ago, markocosic said:

You pedant

Yes.

The trouble is with the term 'thermal mass'.  It has no official (or ofical if you are a card carrying Trump supporter) definition in the sciences, but is banded about as if it is something real, and good.  It gives the impression that the 'count of stuff' is the important part and all materials are equal if they have the similar characteristics. What people are really talking about is thermal inertia, which does have units attached to it because it takes in all top level properties.

 

[joth]

23 minutes ago, markocosic said:

 

If you eliminate the header and pump everything directly from the heat pump then you may not need the buffer

 

The OP is about a 2 zone system each with their own circulation pump. So the low loss header is there for  hydronic decoupling, so each zone can circulate at it's own rate (or not at all), not to increasing the system volume. 

If one zone is not big enough to consume the minimum ashp output the only answer (with this LLH design) is only call for heat when the other zone is open too.

 

Interesting you recommend 2 port buffer. To clarify this is Tee'ed off the primary flow and return? 

Edited August 7, 2021 by joth

[Olf]

This may help explain the problem: https://www.heatgeek.com/low-loss-headers/

 

[Dan F]

1 hour ago, Olf said:

This may help explain the problem: https://www.heatgeek.com/low-loss-headers/

 

 

Good "thermal mass" article on there too:

https://www.heatgeek.com/what-is-thermal-mass/

[SteamyTea]

15 hours ago, Dan F said:

Good "thermal mass" article on there too:

Or an article about heating controls.

Not one SI unit, or even a number in it apart from wall depth.

[BEJB]

On 06/08/2021 at 21:47, markocosic said:

When it comes to removing the header entirely. You say:

 

"The Ecodan documentation says that the flow rate for the heating primary for a 14kw Ecodan is between 17.9 and 40.1 L/min"

 

Can you provide the manual?

 

 

 

I've attached the installation manual. Page 7 gives the range of flow rates for the different models. Mine is 'PUHZ-HW140 17.9 - 40.1'.

 

I do have an option as to what to do when zone 2 calls for heat. The most obvious one is to fire up the heat pump and power up the pump for zone 2. This gives me the fast cycling and very inefficient heating of the UFH.

 

The other option is to power up both pumps - for both zone 1 and zone 2. This could work (by providing 'thermal mass'?) and sounds like it might absorb some of the flow, removing the fast cycling. However, the zone 1 flow pipe comes from the LLH above the zone 2 flow pipe. So the heat all (or virtually all) goes down the zone 1 flow, leaving nothing for the zone 2 flow.

 

I could rebalance the radiators in zone 1 closing them down, so some flow is left for zone 2 when both pumps are on (plus lowering the pump rate for the zone 1 pump). However, it feels like this would mean that, when just zone 1 is on and zone 2 is off, that remaining heat would continue down the LLH and would provoke cycling again (albeit probably slower). 

 

Really struggling here

 

Ecodan_FTC5_Pre-plumbed_Cylinder_Installation__Operation_Parts_Manual.pdf

 

[MikeSharp01]

40 minutes ago, BEJB said:

Really struggling here

Its a very odd in that looks to have been 'designed' at the extreme end of possible straight forward operation. I feel that the compromise of having both zones running is going too far. The header thinking from @markocosic is worth a working through on paper but I would have thought - not sure why, that the ufh and / or the radiators would be enough of a buffer. I feel sure @Jeremy Harris designed his system without a header or a buffer tank or perhaps this wine is befuddling me!

[SteamyTea]

4 minutes ago, MikeSharp01 said:

I feel sure @Jeremy Harris designed his system without a header or a buffer tank or perhaps this wine is befuddling me

He had a buffer.  He also used to to preheat water going into his original Sunamp and to recharge the Sunamp.

Edited August 15, 2021 by SteamyTea

[MikeSharp01]

2 hours ago, SteamyTea said:

He had a buffer

Is / are his post(s) out there somewhere I was sure he dumped the buffer because of the heat loss. 200 litres of water is a load of buffer when you think the slab must many tons although the water in the UFH will not be many litres. Ours is about 40 litres IIRC.

[SteamyTea]

17 minutes ago, MikeSharp01 said:

are his post(s) out there somewhere I was sure he dumped the buffer because of the heat loss.

Try way back machine.

He may have dumped the buffer, it was there when I last visited.

 

Not sure if the slab mass makes a lot of difference to the cycling once it is at operating temperature. More to do with the flow/return temperature differences. A kg of water needs 4.2  kj to raise it a kelvin, concrete only needs 0.8 kj. So a kg of water is worth 5 of slab. And all the water can be heated, but there will be a temperature gradient with a slab, which may effect performance.

[markocosic]

On 15/08/2021 at 17:07, BEJB said:

 

I've attached the installation manual. Page 7 gives the range of flow rates for the different models. Mine is 'PUHZ-HW140 17.9 - 40.1'.

 

Ecodan_FTC5_Pre-plumbed_Cylinder_Installation__Operation_Parts_Manual.pdf 5.52 MB · 4 downloads

 

 

This is the installation manual for the hot water cylinder; not the hat pump.

 

A quick Google for PUHZ-HW140 confirms your flowrate requirements though. Flowrate and return temperature needs to be in the range below in order for the unit to operate.

 

 

If the flowrate through your secondary system (the rads / UFH) and through the 4-pipe low loss header is HIGHER than the flowrate from the heat pump through the 4-pipe low loss header then you're going to have temperature dilution issues.

 

 

I would set UFH flowrate to 18l/min, DHW flowrate to 18 l/min, and radiator flowrate to 22l/mum (max, with all TRVs full open)

 

And the heat pump flowrate to the maximum.

 

Now you can run:

 

- DHW only

- UFH only 

- UFH and DHW

- UFH and any amount of radiator up to 100% open 

 

 

I would set UFH to 18l/min, DHW to 18 l/min, and radiators to 22l/mum (max, with all TRVs full open)

 

Now you can run:

 

- DHW only

- UFH only 

- UFH and DHW

- UFH and any amount of radiator up to 100% open 

 

That will resolve your dilution issue.

 

I would also fit a buffer (a volumiser if you prefer) on the heat pump side of the space heating circuit (flow or return; return probably better) to reduce cycling when the load being drawn off is lower than the heat pump is able to modulate to. 50-100 litres. A small direct unvented hot water cylinder will work.

 

I would also eliminate the 4-pipe header and drive everything directly form the heat pump; using zone valves for the different circuits instead of using pumps.

 

And that's your priority order for biggest bang per buck.