Heatgeek website

[marshian]

19 minutes ago, ProDave said:

One thing never discussed in a "heat low and continuous" strategy is the ancillery items.

 

For instance, when my heating is on, I have four circulation pumps and 2 zone valves on.  At a rough estimate say 200W.  That is power not contributing to heating but used to push water around the system.  

 

I can't help feeling that is a significant proportion of the heating power a lot of the time so perhaps an argument for heating harder for a shorter time?

 

Just for comparison

 

When I have my heating on I have

 

1 circulation pump running at 15W

Boiler and Control system consuming 55W (when running)

No Zone valves powered up at all (NO zone valve for CH)

 

Only when I do HW do I have a zone valve energised for 21 mins a day.

[marshian]

5 hours ago, SteamyTea said:

Just struck me that one of the problems of getting a heating system more efficient, and especially true of gas, is that the usage monitoring, and the energy usage when the outside temperature varies is non-existent.

It will be impossible for most people see a difference in usage.

 

I'm clearly not most people

 

Below is my daily gas usage and below that my HDD data both by day for December

Pretty clear to me my heating consumption of gas is driven by the outside temp and there is a close relationship

 

[JohnMo]

39 minutes ago, DamonHD said:

went back to check whether the circulation pump counts within the SCoP value or not,

It does, so does the controller any thing in the system relating to heat pump. Each electric item should be metered. We take a big dent in SCoP because batch charge. The main circulation pump runs 24/7, as I use it to borrow heat from house to keep a summer house warmish. CoP of 5 doing heating and daily CoP about 3.0. but heated mostly via cheap electric.

 

Our heat pump has only one pump, and a diverter, diverter is powered only when doing DHW.

[DamonHD]

24 minutes ago, marshian said:

I'm clearly not most people

 

Below is my daily gas usage and below that my HDD data both by day for December

 Pretty clear to me my heating consumption of gas is driven by the outside temp and there is a close relationship

 

You are indeed not most people.

 

For gas heating, relationship is typically linear above a baseline temperature and baseline (non-heating) consumption, eg per https://www.degreedays.net/

 

For HP changing CoP with exterior temperature will complicate things, which I haven't entirely worked through in my head yet.

Edited December 31, 2024 by DamonHD

[JohnMo]

1 minute ago, DamonHD said:

HP changing CoP with exterior temperature will complicate things,

And defrost makes a big dent energy output per kWh input. One degree change in outside temp can be the difference between CoP of 5 and 3 due to defrosting.

[marshian]

4 minutes ago, DamonHD said:

You are indeed not most people.

I'll take that as a compliment 😉

 

Background - for my last 40 years in work I've been a data analyst - it's really bloody hard when you are good at working with data to ignore it in a home environment.

 

Basically if I can't validate something with data then it's a guess or a hunch and I don't do either of those

[JamesPa]

@DamonHD

 

I read your paper with interest and like the simple physics based approach, which makes it understandable.  I hope you and your colleagues are doing more of this, because we need it!

 

I do have an concern with the interpretation of the conclusion however.  In essence I think you confirm the 'bad setback' effect which can occur if a reasonably square/cubic house is only partly heated and which is described qualitatively on the heat geek website.  You also explored a range of layout and oat variations and I think find bad setback in most.

 

You then (I think) go on to show that if the occupants don't demand tight temperature control, the bad setback effect disappears and energy savings appear.  So far so good.

 

However not demanding 'tight temperature control' involves accepting that the occupied rooms reduce in temperature by up to just less than 2C and you assert that many will find that ok and thus part heating can deliver benefits.

 

That may well be true but if it's acceptable to reduce the temperature by up to just less than 2C, even greater savings are to be had by turning the WC curve down a notch.  Put another way if a 2C reduction is acceptable, the 'control' scenario is the scenario where that was the target temperature from the outset and with this as.a control the bad setback effect remains.

 

I suppose you could argue that,  if the WC curve can't be or isn't changed by the householder, then the original control is still the correct one, but this argues for a user friendly way to get WC right on the first place not a second rate substitute.

 

Perhaps I have misinterpreted (and if I have please accept my apologies) but if not I think your paper confirms for a fair range of situations that heating only part of a (reasonably square/cubic) house is likely to lead to higher consumption assuming that you have WC adjusted to keep those rooms you need at or around whatever your target is.  If so this is an extremely valuable confirmation of what heat geek describe somewhat qualitatively.

 

Is there by any chance a similar paper dealing with temporal (as opposed to spatial) setback.  If so can you post the reference?

Edited December 31, 2024 by JamesPa

[DamonHD]

Note that the bad setback effect applies if rads are only just large enough AND pure weather comp is not used.  That would not in general be allowed for an MCS install as I understand it, so I don't think it is a real problem if thought about.

 

(Maybe an installer noticing that a user is zoning and wants to carry on doing it should suggest upsizing rads a bit more than they would otherwise.)

 

Also note that the maximum temperature sag occurs when it's warm enough outside that you barely need heating - it's usually less.  See the end of https://github.com/DamonHD/TRVmodel/blob/main/SampleComputationsOutput.txt

 

I don't know of any paper dealing with your last para (I would be keen to read it too) but the research strand that I'm currently on may look at that and test it in a physical model too.  I really have to get up to speed on EnergyPlus and heat pump simulation in a hurry!

Edited December 31, 2024 by DamonHD

[SteamyTea]

4 hours ago, marshian said:

it's really bloody hard when you are good at working with data to ignore it in a home environment

Yes.

Wish I had known that there was a career in data analysis 50 years ago.

 

Almost the end of the year, shall be updating my spreadsheet in the morning.

[DamonHD]

Data Science is a thing (though has a newer more fashionable name this week I think) and all the data scientists that I know are amazing multitalented people...

 

I have a pile of EOM and EOY data capture / munging lined up for tomorrow too...

[SteamyTea]

1 minute ago, DamonHD said:

have a pile of EOM and EOY data capture

Something to do instead of going to work.

This will be the first full year that I have Smart Meter data to compare against.

Our old mate @Ed Davies recently compared his Current cost data against his Smart Meter, think he got to within 3% agreement, I did a quick analysis on my better CurrentCost with the optical sensor, rather than the current clamp, and got to 1.5%. but I know I lost a few days data (when I put my collapsible garden chair over the transmission path).

Hopefully I will know more tomorrow.

[JamesPa]

1 hour ago, DamonHD said:

Note that the bad setback effect applies if rads are only just large enough AND pure weather comp is not used.  That would not in general be allowed for an MCS install as I understand it, so I don't think it is a real problem if thought about.

 

.....

 

Also note that the maximum temperature sag occurs when it's warm enough outside that you barely need heating - it's usually less.  See the end of https://github.com/DamonHD/TRVmodel/blob/main/SampleComputationsOutput.txt

 

I don't think that changes my mind.  If rads are oversized then you can adjust the WC down until they aren't, thereby achieving greater savings than though zoning.  If puree WC isn't used then you are sacrificing efficiency a different way and it's a different scenario altogether.  

 

Clearly there is a difficulty here in defining what the correct 'control' is, we may disagree or we may just not be clear enough.  It's a really important thing to nail though and if there is any doubt (which there currently is) then it's important, I think, to state it as an assumption in any summary.

 

In other threads real experiments are discussed often with inferences that are quite extreme.  If we can't work out the control in a thought experiment, what chance does a real experiments have!

[DamonHD]

Yes, these things are debatable, but a core issue for me is that no one WC curve (with static balancing however good) is perfect for all combinations of variable occupancy, solar gain, wind, etc.  And comfort suggests wanting to maintain some differentials anyway where possible.  Note that I am primarily concerned with retrofits into relatively thermally leaky existing UK stock which may skew things.

 

Note also that I did reduce (~10x) some inter-room uncontrolled heat transfer in my house between the living room and a bedroom above while doing other works, so that option helps and is available too, if one knows about it and why it might help.

[Marvin]

9 minutes ago, JamesPa said:

I don't think that changes my mind.  If rads are oversized then you can adjust the WC down until they aren't, thereby achieving greater savings than though zoning.  If puree WC isn't used then you are sacrificing efficiency a different way and it's a different scenario altogether.  

We altered our weather compensation.  Outside temperature now 10C ASHP buffer tank set temp 35C. Living room 23.1 C

[Marvin]

1 minute ago, DamonHD said:

Yes, these things are debatable, but a core issue for me is that no one WC curve (with static balancing however good) is perfect for all combinations of variable occupancy, solar gain, wind, etc.  And comfort suggests wanting to maintain some differentials anyway where possible.  Note that I am primarily concerned with retrofits into relatively thermally leaky existing UK stock which may skew things.

I think its easier to calculate which horse would win the grand national. I think there are about 50 variables which effect heating a home. We calculated down to -8 and covered that because we were not sure (renovated bungalow) Most of the time we overheat too easily even though we only have one skylight window facing south.

[marshian]

13 minutes ago, JamesPa said:

I don't think that changes my mind.  If rads are oversized then you can adjust the WC down until they aren't, thereby achieving greater savings than though zoning.  If puree WC isn't used then you are sacrificing efficiency a different way and it's a different scenario altogether. 

 

If rads are also oversized then you can reduce the flow thro that rad at any given temp and basically make it behave like a smaller rad

 

Quite hard to go the other way mine except winding up the flow and sacrificing the return temps

 

I can definately save some gas by turning the heating off when I'm not in the house but the re-heat process means I have to start quite early before I get home and I'd rather have the house heated a little lower all the time than burst of high energy heating

 

13 minutes ago, JamesPa said:

Clearly there is a difficulty here in defining what the correct 'control' is, we may disagree or we may just not be clear enough.  It's a really important thing to nail though and if there is any doubt (which there currently is) then it's important, I think, to state it as an assumption in any summary.

 

I think there is a lack of clarity in the discussion but you only get clarity by discussion so lets keep it going in the good natured way it is

 

World is too damn polarised now

 

13 minutes ago, JamesPa said:

In other threads real experiments are discussed often with inferences that are quite extreme.  If we can't work out the control in a thought experiment, what chance does a real experiments have!

 

Exactly what I mean by polarised viewpoints - there is no right and wrong only different approaches (with the UK's housing stock - there is an awful lot of grey and I'm not talking about the fashion trend for windows and cladding)

 

[JamesPa]

11 minutes ago, DamonHD said:

Yes, these things are debatable, but a core issue for me is that no one WC curve (with static balancing however good) is perfect for all combinations of variable occupancy, solar gain, wind, etc.  And comfort suggests wanting to maintain some differentials anyway where possible.  Note that I am primarily concerned with retrofits into relatively thermally leaky existing UK stock which may skew things.

I too am principally interested in retrofits, it's a purely amateur interest though.

 

I also agree that comfort argues for some differentials, but maybe the differentials can be achieved without bad setback and without having a wc curve that is higher than it needs to be.  My thinking here is that maintaining a level of background heating in the room where setback is desired is, primarily facing, better than turning it off altogether because you are increasing the total active emitter area and can thus operate at a lower ft.

[marshian]

6 minutes ago, Marvin said:

I think its easier to calculate which horse would win the grand national

My other half and her mate bet on every single horse in the Grand National once.........  2009 was a lucky year for that process 100/1 winner

[JamesPa]

26 minutes ago, marshian said:

If rads are also oversized then you can reduce the flow thro that rad at any given temp and basically make it behave like a smaller rad

Quite so, but if all rads are oversized, which I think was the implied scenario, then better to turn the wc down.

 

26 minutes ago, marshian said:

48 minutes ago, JamesPa said:

Clearly there is a difficulty here in defining what the correct 'control' is, we may disagree or we may just not be clear enough.  It's a really important thing to nail though and if there is any doubt (which there currently is) then it's important, I think, to state it as an assumption in any summary.

 

I think there is a lack of clarity in the discussion but you only get clarity by discussion so lets keep it going in the good natured way it is

Agreed.

 

To my mind if we are looking only for basic physics effects (or only modelling basic physics effects) in the steady state then the control, unless stated otherwise, is a system which is adjusted to achieve a specified indoor temperature 24x7 across the whole house in the most efficient way possible.  With a heat pump this is a system where the balanced emitters are fed with water 24x7 at a temperature just sufficient to maintain the design IAT.

 

From this starting point one can I think tease out the effects individually (for exanple) of zoning, time based setback or varying the target temperature .  In the first two cases the control should be compared with a system which achieves the specified temperature in the non setback rooms or the non setback times (as applicable), not with one that achieves different temperatures in those rooms or times.  In third case the control should be compared with a system which achieves the new target temperature 24x7 across the whole house.  Each of these then teases out the effect of changing only one condition.

 

Of course it's also valid to look at cases where more than one condition is changed, but that's a different set of questions, and if two conditions are changed then they must both figure in the description of the test being done. 

 

If we are lucky the effect of changing two conditions is, for sufficiently small changes, the sum of changing either individually, so we don't need to model all the combinations.

 

That's my starter definition anyway.

Edited December 31, 2024 by JamesPa

[marshian]

4 minutes ago, JamesPa said:

Quite so, but if all rads are oversized, which I think was the implied scenario, then better to turn the wc down.

 

It's actually I think hard to match rads to exactly the room heat loss requirements and not get them oversized (OK I did in one room)

 

For 95% of the year they are oversized by virtue of the outside Air Temp being above the worst case outside temp.

 

4 minutes ago, JamesPa said:

To my mind if we are looking only for basic physics effects (or only modelling basic physics effects) in the steady state then the control, unless stated otherwise, is a system which is adjusted to achieve a specified indoor temperature 24x7 across the whole house in the most efficient way possible.  With a heat pump this is a system where the balanced emitters are fed with water 24x7 at a temperature just sufficient to maintain the design IAT.

 

From this starting point one can I think tease out the effects individually (for exanple) of zoning, time based setback or varying the target temperature .  In the first two cases the control should be compared with a system which achieves the specified temperature in the non setback rooms or the non setback times (as applicable), not with one that achieves different temperatures in those rooms or times.  In third case the control should be compared with a system which achieves the new target temperature 24x7 across the whole house.  Each of these then teases out the effect of changing only one condition.

 

Of course it's also valid to look at cases where more than one condition is changed, but that's a different set of questions, and if two conditions are changed then they must both figure in the description of the test being done. 

 

If we are lucky the effect of changing two conditions is, for sufficiently small changes, the sum of changing either individually, so we don't need to model all the combinations.

 

That's my starter definition anyway.

 

Pretty much agree with that

[JamesPa]

19 minutes ago, marshian said:

It's actually I think hard to match rads to exactly the room heat loss requirements and not get them oversized (OK I did in one room)

 

For 95% of the year they are oversized by virtue of the outside Air Temp being above the worst case outside temp.

Ok I didn't describe it well.  By 'right sized' I mean at design OAT, although a perfect WC curve will mean that they are also right sized at other oats (that's the point of WC after all). 

 

Of course rads will, in reality, be incorrectly sized to different degrees, that's where balancing comes in.  However if they are all oversized once balanced it's better to turn down the WC than further to restrict the flows through all the radiators!  If they are all undersized then you have to turn up the WC.  Either way (In this correctly adjusted 'control') you end up with right sized rads at the operating ft (which may or may not be the design ft).

 

As it's all a thought experiment you can just simplify this all down to correctly sized rads.

Edited December 31, 2024 by JamesPa

[marshian]

11 minutes ago, JamesPa said:

Ok I didn't describe it well.  By 'right sized' I mean at design OAT.  Of course rads will be incorrectly sized to different degrees, that's where balancing comes in.  However if they are all oversized once balanced it's better to turn down the WC than further to restrict the flows through all the radiators!  If they are all undersized then you have to turn up the WC.  Either way (In this correctly adjusted 'control') you end up with right sized rads at the operating ft (which may or may not be the design ft).

 

As it's all a thought experiment you can just simplify this all down to correctly sized rads.

Here's my math for my rad sizing

 

 

In a nutshell what I had was a calculated heat loss at -2 and a rad kW output at T50 - what I wanted to see was where as I stepped down flow temps was I likely to have a problem with rooms not meeting the target temp (ie the rad was borderline)

 

In truth I think my heat loss parameters for ACH and loft losses were a little pessimistic and so my heat losses for some rooms are a little higher than reality.

 

But it does show clearly that I am likely to have an issue with three or four rooms

[JamesPa]

13 minutes ago, marshian said:

Here's my math for my rad sizing

 

 

In a nutshell what I had was a calculated heat loss at -2 and a rad kW output at T50 - what I wanted to see was where as I stepped down flow temps was I likely to have a problem with rooms not meeting the target temp (ie the rad was borderline)

 

In truth I think my heat loss parameters for ACH and loft losses were a little pessimistic and so my heat losses for some rooms are a little higher than reality.

 

But it does show clearly that I am likely to have an issue with three or four rooms

For reasons to do with the history of my installation mine are similarly all over the place.  In reality the downstairs pretty much evens itself out, undersized rads being offset by oversized ones in adjacent rooms, and the bedrooms upstairs are on trvs because they are way oversized and anyway we want them cooker (I will sometimes get round to trying to adjust the bedroom lockshields and ditch the trvs).  The bathroom is, in principle, hopelessly undersized (we didn't change the towel rail), but I leave the door open to the hall and that, together with steam from the bath, deals with it.

 

It all seems to work, based on four weeks so far of operation, and with much less fiddling than I had expected.  It's a nice simple system, no hydronic separation nonsense, just heat pump connected to emitters and dhw via a diverter.  I'm operating a couple of degrees less than the design flow temperature so happy with that too.

Edited December 31, 2024 by JamesPa

[MikeSharp01]

On 31/12/2024 at 17:45, DamonHD said:

No WC curve (with static balancing however good) is perfect for all combinations of variable occupancy, solar gain, wind, etc. 

How hard can it be to derive a curve that takes account of all these factors. I know that here, because we are only exposed on one quarter really that when the wind is in that quarter the heating has to work harder.

[DamonHD]

Because one curve for the whole home is clearly insufficient at some level when the heat losses and gains of rooms relative to one another is changing.