@jamespa - thanks as ever for your analysis.

Posted by: @jamespa

↑

Interestingly the difference in energy supplied by the heat pump (ie heat to house) is large (29kWh, =27%).  

This is indeed interesting. Take the back of a fag packet assessment of the saving, 6/24 hours (off for 6 out of 24 hours) and the savings is 25%. This is for energy out ie delivered to the house. Yet as you correctly say the energy in saving is of the order 4%. What this tells us is that the COP must have changed, quite substantially. And indeed it has, the COP for the setback period is 3.9, for the no setback period it 5.1, a substantial difference, all the more so given the OAT was pretty much the same over both periods. If it isn't the OAT affecting the COP, then what else might it be? Let us look at various factors and numbers side by side:

                                  Setback                    No setback                     Comment

When                          Feb 24                      Feb 25

Mean OAT (°C)            10.2                          9.9                                Much the same

Mean IAT (°C)             18.7                          20.2                              As expected, should reduce energy in a bit over setback period

Energy in (kWh)          20.38                        21.22                             4% less in setback

Energy out (kWh)        78.73                       108.29                            27% less in setback 

COP                            3.9                           5.1                                 24% less in setback

Set LWT (°C)               39-41                       37-38                             Slight higher in setback, could lower efficiency (COP)

WCC                           56 @  -4, 34 @ 15     52 @ -2, 31 @ 15            WCC set ~ 3-4 degrees higher in setback, could lower efficiency (COP)

System state               'throttled'                  'free flow'                       Refers to lock shield valves restricting secondary flow

Taken all together, the bottom line is our old enemy, we are not comparing like with like. The setback period was characterised by a throttled system with a higher WCC and as a result Set LWT (and actual, though I haven't calculated it, only eyeballed it), all of which are going to lower efficiency (COP), and have the effect of increasing energy in, very possibly wiping out any saving. To make a valid comparison, we would need to control for those changed parameters, a difficult task to say the least. Overall, it will be easier to wait for further data to come in over a period when the key parameters remain the same, and then make a comparison.

In fairness I would expect the COP to change, because you are working the heat pump harder in setback (which is the whole basis of the argument that setback may cost money).  What is bizarre is that the continuous running case 'requires' 27% more energy to be delivered to the house.  The questions are

1. is this real?

2. if it is, where is it going?

As you can see from the spreadsheet I cant account for it.  It must, however (if it exists) be going somewhere!

The actual difference in electricity use was only 4% and as you say there are other moving parts and we aren't, it transpires, comparing like with like.  Given all of this we cant take much away from this at present and need to wait further data!

Posted by: @jamespa

↑

In fairness I would expect the COP to change

So would I, I am just underlining the differences between the two running states, and asking whether things are moving in the expected direction.

Posted by: @jamespa

↑

1. is this real?

A question I ask myself all the time! First of all, the continuous running period had a slightly lower (0.3°C) mean OAT, a and a definitely higher mean IAT, by 1.5°C. Both of these difference will need (consume) more energy to the house. If the 10% rule of thumb applies (10% extra energy up or down for each degree IAT up or down then that might suggest the continuous period will need perhaps 15% more energy per se. But that still leaves 10% going down a rabbit hole somewhere.

The energy out data comes from the minute readings, where I record Midea reported primary flow rate and LWT/RWT delta t and then multiply together with the circulating fluid specific heat to get the power and then over the hour the kWh of energy delivered. The flow rate seems to be reasonably accurate, judged against a separate analogue flow meter in the circuit. The LWT/RWT delta t also seems consistent with readings I have taken from the pipework entering and leaving the plate heat exchanger. There is nothing obviously out of whack.

Then we also have the Midea total lifetime energy out (the integer values, but rounding errors will lessen over longer periods), which can also be used to calculate energy out per hour (reading now - reading an hour ago). This is in column H in the spreadsheet, and it includes all energy out, ie DHW as well as space hearing. The numbers for the two periods are:

continuous period: 112kWh for both space and DHW heating (vs 108.3kWh calculated for space heating only)        

setback period: 85kWh for both space and DHW heating (vs 78.7kWh calculated for space heating only)

and they seem in pretty close agreement, given the continuous period used 3.6kWh for DHW giving a net space heating use of 108.4kWh, and the setback period used 7.2kWh for DHW (the DHW was a lot colder that day), giving a net space heating use of 77.8kWh. Both are within 1kWh of the calculated values. It is of course possible that two wrong values, ultimately derived from the same incorrect raw data, could appear to confirm each other, but... As things are, it seems they are plausible values, what is missing is the explanation of where the extra 10% (after making allowance for the lower OAT/higher IAT) went. Perhaps heat pumps have invisible flues, up which heat can mysteriously disappear!

Another possibility is the setback period, which happened during the 'throttled' ('pinched PHE') period, somehow throttled the energy out to make the total less than it would otherwise have been. @bobtskutter may have something to say on this.

As you say, and I said earlier, we need more data from a continuous period where fundamental system changes didn't happen.

Hello, sorry I've been quiet for a while.

In CathodeRays case the plate heat exchanger was acting as restriction in the flow of energy from the heat pump into the room.  The energy balance of the house would still work, but the temperatures would be different.

Posted by: @cathoderay

↑

Perhaps heat pumps have invisible flues, up which heat can mysteriously disappear!

You will loose energy through the outside pipework, so there is some "lost energy".  A 10% error doesn't sound too bad to me considering how much input data there is and how it's been measured.  Small errors on each measurement would compound up through the calculation.  You could do an uncertainty analysis of your calculation.

A lot of threads talk about Leaving Water Temperature because that's what makes the heat emitters hot.  However, it's the Return Water Temperature to the heat pump that determines the heat pump compressor power requirements. 

Remember the heat pump works by compressing gas, which makes it hotter.  That GAS then condenses in the heat pump plate exchanger and gives up it's energy to the flowing water.  You can only condense the gas if the return water temperature is cold enough, if it's not cold enough the compressor must increase discharge pressure.  The higher discharge pressure will increase the dew point of the gas and allow it to condense at a higher temperature, but the input electrical energy to the compressor motor will go up.

How could we account for RWT in the electrical power calculations?  I have a suspicion that set back saves energy because the RWT is colder for longer.  There was a thread on here with someone saying their UFH was most efficient if they cycled the heat pump.  I bet that was because the RWT was kept low.

Bob

Posted by: @bobtskutter

↑

sorry I've been quiet for a while

No need to apologise, we are all free to come and go as we please, but good to see you back!

The heat loss from the outside pipes is an interesting idea and the insulation isn't perfect but the runs are short but more to the point would any heat loss there be proportional to the conditions, and so unlikely to explain why similar conditions have similar energy in, but different energy out values?

The missing 10% may well just be random error, and possible outliers. We are after all comparing only two 24 hour periods when albeit the were similar. More comparisons when we have more data may help, by way of some some regression to the mean.

I think @jamespa may also have pointed out the real driver of the system is the RWT, which sort of makes sense to non-experts like me, since it is how the heat pump gets its feedback about what's going on the building. I do record the RWT so it is certainly available for use in any calculations. More generally, I think maybe we should talk a bit more about the RWT, so we don't forget its importance.

OK, so I'm just thinking out loud here:

You loose energy through the walls, roof, and draughts - anywhere else?

The 10% lost energy could be going out via draughts, do you have any weather data available?  There are websites that record weather station data which might tell you the wind speed.  I know for my house that wind speed makes a big difference to how fast the house cools down.

Bob

Posted by: @bobtskutter

↑

The 10% lost energy could be going out via draughts, do you have any weather data available?  There are websites that record weather station data which might tell you the wind speed.  I know for my house that wind speed makes a big difference to how fast the house cools down.

I too have observed that windy (which are often wet as well) days do have a wind chill effect on the house. I do have quite a lot of weather data (as another interest I am trying to look at how accurate the inshore waters forecast is, another 'interesting question' fraught with methodological problems) but it is coastal (English Channel) weather data but may be 'good enough' for our purposes. If its blowing a gale in the Channel it will probably be blowing pretty hard in British southern counties. The only problem which is not trivial for our current purpose is that it tends to be rather 'in arrears', years rather than months. I do have a nearby Met Office WOW station but it records only intermittently, and I suspect from its location it has a micro-climate. But one way or another I am sure I can find something, I know where to look!

But the real question about the n=2 study is how both periods had much the same energy in, yet very different energy out values, despite similar (but not identical) OATs and IATs. Certainly the first candidate explanation that has to be eliminated is measurement error, but I think I have done that to a tolerable degree of certainty. I'm inclined to think, as I think @jamespa also is inclined to think, that it is somehow down to system changes, in particular opening up the valves following your suggestions to get better flow. Basically, the Feb 24 setback period had the system 'pinched' and that restricted energy out, whereas the Feb 25 no setback period had full and free flow, and so was able to deliver more energy. Perhaps.

The real answer I suspect lies in finding periods to compare that have the same OAT and IAT and have the system running in the same state.      

Posted by: @bobtskutter

↑

A lot of threads talk about Leaving Water Temperature because that's what makes the heat emitters hot.  However, it's the Return Water Temperature to the heat pump that determines the heat pump compressor power requirements. 

Remember the heat pump works by compressing gas, which makes it hotter.  That GAS then condenses in the heat pump plate exchanger and gives up it's energy to the flowing water.  You can only condense the gas if the return water temperature is cold enough, if it's not cold enough the compressor must increase discharge pressure.  The higher discharge pressure will increase the dew point of the gas and allow it to condense at a higher temperature, but the input electrical energy to the compressor motor will go up.

Can you explain this please, the heat pump manufacturers always specify COP according to LWT.  The compressed gas must be hotter than the LWT otherwise heat cannot be transferred, and the energy required is (LWT-RWT)* flow rate.  If the compressor power were principally dependent on RWT not LWT, then we would run at a much greater radiator deltaT, whereas in fact we try to run at a low radiator deltaT.

I do recall a not dissimilar argument for boilers, but the ultimate energy source in this case is different so the arguments are different.

I have a related question about Rad deltaT, which is raging in another thread and which you may be able to help with.  We typically design for 5C, why? 

I thought it was because the thermodynamics would argue for a low deltaT because that gives highest average emitter temperature for any given LWT, however practical limitations (basically how fast you can pump water) mean that anything less than a few degrees becomes challenging, ie its basically an engineering compromise. 

However others have said that 5C (or somewhere in the region 2-8C) is somehow 'magic' for reasons 'to do with compressor dynamics' 9ie they dont have an explanation).  Do you have any idea?  Its notable that some heat pumps try to control rad deltaT as power output changes by modulating the water pump, and others (even if they have the ability to modulate the water pump and do so for other reasons) don't.  Why I wonder?

I think the theory should be that you save what the difference in heat loss is between normal and lowest temperature that is reached during the setback temperature. 
Two examples:
So if your house is very leaky and drops from 20 to 15 degree in an hour, you save about 1/4 of your heatloss at 0C during the setback period. The energy that you have to put back is really just the energy you borrowed to keep the house warm while it was cooling down, so I think it can be ignored. 
if your house is very well insulated, let’s say it only looses 1C during the setback period. Then you only save a small portion off a small heatloss. 
Of course if you have medium heat loss the math becomes complicated by all those real world factors. 

Posted by: @adrian

↑

I think the theory should be that you save what the difference in heat loss is between normal and lowest temperature that is reached during the setback temperature. 

You are almost right and have appreciated the key point ie that energy lost must be replaced (simple minded, and incorrect, thinking says that if you heat for three quarters of the day you use three quarters as much energy which, as you recognise, is not true.)

Posted by: @adrian

↑

So if your house is very leaky and drops from 20 to 15 degree in an hour, you save about 1/4 of your heatloss at 0C during the setback period.

Actually its about one eighth, because the house doesn't instantly cool from 20 to 15, it cools over a period of time.  The cooling is exponential, but for small drops fairly close to linear hence very roughly one eighth.  Obviously if it then stays at 15 for a period of time, its roughly one quarter during that period.  But houses don't normally behave like that.

Posted by: @adrian

↑

The energy that you have to put back is really just the energy you borrowed to keep the house warm while it was cooling down

Correct, plus of course the energy you continue to lose during the recovery period.

but....

What we are talking here is energy to the house which is supplied by the heat pump and other energy emitters, but primarily the heat pump.

What we pay for (and therefore care about) is energy to the heat pump.

Because the heat pump has to work harder during the recovery period it does so (all else being the same) at a lower COP.  To be more precise, to recover at all you will have to operate during the recovery period at a higher LWT than would be the case had you not performed a setback, because you have to supply a greater amount of energy from the same emitters.  This results in in a lower COP.  

This may offset completely, or even dominate, the reduced amount of energy you need to deliver to the house.  So you may lose, or you may gain.

Now add in the fact that it may (or may not) be warmer outside during the recovery period if, for example, the setback is at night and it begins to get complicated.

A very leaky house is a case where the saving in energy lost does not outweigh the COP penalty (a tent, which cools very quickly indeed, would be a perfect example).  However most heat pumps aren't (yet) fitted in very leaky houses and for them the situation is inconclusive (ie we don't have robust enough theory, robust enough experiments, or sufficient correlation between the two to draw definitive conclusions in many cases).

Hello @jamespa

Posted by: @jamespa

↑

Can you explain this please

I've attached a spreadsheet which attempts to show how a heat pump compressor system works.

The gas enters the compressor as a superheated vapour, i.e. it's hotter than it's boiling point.

The compressor then compresses the gas which makes it hotter, it's got more superheat.  The attached spreadsheet the compressor discharge gets to 72C.

That superheated gas is cooled in the plate exchanger untill it reaches is dew point, which is 34.5C in the spreadsheet.

The gas then condenses into a liquid at 34.5C.

Roughly 90% of the heat transferred into the heating water is by condensation of the gas at 34.5C <- this is an important number.

The heatingflow line on the chart represents the water getting hotter through the plate exchanger.  Notice how close the heatingflow line is to the condensing line.  The temperature difference between the condensing gas and heatingwater flow is what drives the energy into the water.  The heatingwater temperature must always be less than the condensing temperature otherwise energy will not flow from the refrigerant into the water.

If you change the RWT to a low number e.g. 10C (cell O8) the DT will increase, at which point you can lower the compressor discharge pressure (cell E7) until the condensing temperature matches the heatingwater flow (about 11bara).  This shows how lower RWT helps reduce ASHP input power which would help increase COP.

I'll get on to the DT question in a minute.

FYI, the charge is a diagram trying to show the energy transfer inside the plate exchanger.  The X axis represents the distance along the plate exchanger.  The numbers in the spreadsheet are close to what I get using commercial design and simulation software, but they're not perfect.

Bob

@jamespa

Posted by: @jamespa

↑

We typically design for 5C, why? 

I don't know where this comes from. I've heard a variety of numbers mentioned for heating design.  10C, 5C, 7C.

I think it comes from solid fuel systems and early cast iron boilers.  If the temperature difference across the cast iron boiler was too big there would be too much stress on the casting and the casting could crack.  So therefore "limit the DT to a small number and you'll be OK".

I've seen very large commercial heat exchangers with expansion bellows to account for excessive temperature difference along their shells.

My personal opinion is now "design for a DT of..." gives people a starting point for designing their heating system.

If you have a 10kW heat source and DT is fixed at say 10C, because "design for a DT of 10C" then the system flow rate is also fixed.  That allows you to select the proper pipe size and then choose a pump suitable for the system pressure drop.  Once you've done all that, if the heat source can modulate down and the flow is fixed (because you've already chosen the pipe size and pump) then the DT will go down.

So I think "design for a DT of..." is way to give people a starting point.

You also need to select the actual flow temperature.  e.g. design for flow = 35C and DT = 5C.  The DT specification allows you to select pipework and pump (for a given power output) then the flow temperature allows you to select the size of the heat emitter.  Then you need to check the selected heat emitter and flow temperature and flow rate gives you required DT.  And so the design loop goes round again.

In the previous post about needing a low RWT to get best efficiency you can see why running with a high DT might be a good thing.  The LWT must be hot enough to allow the heat emitter to work, and the RWT must be cold enough to allow the compressor to operate a minimum discharge pressure.

Maybe having a system where the flow leaving the ASHP goes through a radiator circuit, then recombines and goes through an underfloor heating circuit might be a way to design the best system. The rads need a relatively high flow temperature, and the underfloor does not.  It would need carefull pipework design to make sure the low flow (the system would run with high DT) would split correctly to heat radiator.

This post isn't much of an answer.  I don't have one.  But I do think "Design for a DT of..." should be used as a starting point for iterating around a design.

Bob