@iantelescope

It is difficult for me to try to assess what may be happening within your system, since there are many of my previous questions left unanswered.

Is the buffer tank still in the circuit with primary water flowing through it?

Were the temperature measurements taken before or after any connection to the buffer tank?

How are you measuring the flow rate?

Was the secondary pump running during your test?

Were the radiators getting warm?

Are any TRV's restricting flow through the radiators?

A low DT will occur across the inlet to the PHE, if the thermal energy is not being taken away by sufficient flow through the secondary circuit, with the thermal energy being dissipated by the heat emitters. Is this happening?

@jamespa

Hi James,

I have eventually found the time to complete the modifications to the spreadsheet model that I created some time ago.

The first sheet gives details and instructions, but if you require clarification then please feel free to ask.

Whilst it was created for a 14kW Ecodan, the effect of changing parameters can still be useful when assessing other models or manufacturers equipment.

The cells highlighted in green can be adjusted, but I would caution against making other changes since many cells contain formula.

@derek-m Thanks for this, very nicely set out and logical.  I will have a play over the next days but it certainly looks like a great piece of work.

I have now, in my mind, constructed a rational argument for how to 'model' on/off operation (eg nightime/daytime setback).  Interestingly it is to quite a large extent about the behaviour of the emitters rather than the heat pump itself, and about what choices are made for control during the 'recovery' period, which is where inefficiency occurs.  The choice of control strategy may of course be fixed by the heat pump.  Fan radiators have an advantage (which I had not expected) here, because they can (and do) increase their output without increasing LWT, which means that the recovery period can be shorter without sacrificing efficiency.  

I have a crude excel model but its not yet sufficiently self explanatory to post.  Broadly speaking it suggests that switching off at night does save energy during mild periods, but is less likely to do so during very cold periods (qualitatively because the heat pump struggles too much during the recovery).  Also its more likely to save energy for a house with low thermal mass than one with high thermal mass.  I speculate that the optimum (economic) strategy for a house which is unoccupied during the day and where the residents want it a bit cooler at night. is to switch off just before bedtime only switch back on mid afternoon, using fan heaters for the breakfast hour.  I haven't proved that yet.  The model assumes radiators not UFH, I am guessing that most UFH systems have sufficient thermal mass in the slab that diurnal modulation of temperature isn't frequently considered.

I will post the rationale in the next days then try to get some time to tidy up the excel model.  I may draw on your tables as they are better set out than the ones I have in my weather compensation excel model.

@derek-m 

Hi Derek,

Again, many thanks for your patience.

I currently have the following fitted to my Heat Pump:

1) Four ( 4 ) , K Type Thermocouples connected across the input ports from the Heat Pump in my garden.

2) Three ( 3) DS 18B20 Semiconductor Sensors connected across the Heat Exchanger input ports.

3)  Two ( 2) , in Water Delta_T Thermal sensors connected to the Output and return in the Radiator circuit.

4) A "Sharky" In Water Power,Energy and Delta_T meter fitted into the Radiator circuit.

5) Two Flow meters , a SIKA Flow meter ( flow displayed on Samsung) and a "Sharky" Flow meter attached to the Radiator circuit.

6) A dual K Type Clamp portable meter is also available.

7) A switchable 50 l buffer connected in parallel with the Heat Exchanger.

How do I  practically measure, record  and describe the complete Static and dynamic status of this system?

An increasingly frustrated , and despondent

ian

@iantelescope

The first thing that you need to do is gather some data from the various sensors, at a sufficiently frequent time interval to be able to assess what is happening.

You need to record the indoor temperature and the outside temperature along with all the sensors that you listed. Put the readings into a suitable spreadsheet or table.

Take the readings over several hours without making any alterations to any of your system settings.

Once you have collected sufficient data under steady state conditions, then change one of the parameters, say a flow rate, and continue taking readings until the system stabilises. ONLY CHANGE ONE PARAMETER AT ANY ONE TIME AND WAIT UNTIL THE SYSTEM STABILISES.

Try the following:-

Vary the primary flow rate.

Vary the secondary flow rate.

Isolate the buffer tank, so that the primary flow only goes through the PHE.

Raise the heat pump LWT in say 5C increments.

Once completed post the results for analysis.

@jamespa

I fully agree with your assessment that FCU or A2A heat pumps are probably more efficient than A2W heat pumps. I remember reading somewhere that our European cousin's are often quite surprised that us Brits use A2W in preference to A2A.

I have the proof at home, since our A2A heat pump raised the indoor temperature by approximately 2C in 1 hour this morning, whilst using totally green and free energy from our solar PV system.

This is where the problem of 'folk' come into the equation. As a Yorkshireman may often say 'there's nowt so queer as folk'.

I suppose the starting point is to decide what one is trying to achieve. Is the objective to reduce energy consumption, or to reduce overall operating costs, which often can be at odds with one another.

Lifestyle needs to be taken into account since some homes are occupied throughout the day, whilst others can be unoccupied for large periods.

It should be possible to incorporate the effect of thermal mass into the 'playroom' calculations.

Possibly a good analogy would be a heat pump supplying a large thermal store, which in turn feeds the heat emitters. If the heat pump is switched off for a period of time, the thermal store will start to cool as energy is removed by the heat emitters. When the heat pump is switched back on it has to not only supply the heat emitters, but also replace the lost energy from the thermal store. The rate at which the thermal store is to be reheated will determine how much additional energy the heat pump will need to provide.

As you correctly suggested, running the heat pump during the mid afternoon period would probably be more efficient than the early hours of the morning.

It may be possible to try different settings of thermal mass within the equations to find one that more closely matches one's home.

@jamespa

Using FCU or a A2A heat pump should be more efficient during a morning temperature recovery, since they heat the indoor air directly, rather than trying to heat the whole building fabric.

A smaller thermal mass may be slightly more efficient than a larger one, since the building heat loss will reduce at a quicker rate as the indoor temperature falls faster.

By switching the heat pump off and monitoring how quickly the indoor temperature falls at different outside temperatures, it may be possible to obtain a reasonable estimate of the thermal mass.

If the desired indoor temperature setting is reduced from 21C to 19C, then with an outside temperature of say 0C, the building heat loss may be 10kW, which if it took 4 hours for the indoor temperature to fall 2C, the total heat loss would be approximately 40kWh. If at a warmer outside temperature, the building heat loss was now 5kW, then it would probably take approximately 8 hours to lose the 40kWh. So the heat pump may remain switched off for 8 hours rather than 4.

In the first example, when switched back on, the heat pump would need to produce at least 10kW, and would take a very long time to raise the indoor temperature from 19C to 21C. Even increasing the heat pump output to 11kW, it would still be very slow in raising the temperature and would mean operating the heat pump at lower efficiency.

Obviously in the second example, increasing the heat pump output from 5kW to say 6kW, would raise the temperature at a faster rate, again at reduced efficiency, but with the heat pump operating at a higher overall efficiency.

The other factor to consider is the input power saving in relation to the extra input energy required to replace the lost heat.

I therefore suspect that it is not quite straightforward to achieve temperature setback under different operating conditions.

There is also the personal comfort aspect to consider.

Posted by: @derek-m

↑

There is also the personal comfort aspect to consider.

1. The personal comfort can vary from person to person, 19C to 22C some can be fine with either and not feel a difference, yet measured it's quite a hefty %.

2. Have you considered how at a set air room temp, depending if the radiators/walls absorb or radiate heat, one can feel cold or not?

Hence the HP cold(under30C) radiators, despite that being enough to keep the room temp steady.  

@derek-m, @fazel thank you for your comments.  I think I should start another thread on 'on/off' working, it doesn't really fit here.

Posted by: @derek-m

↑

@jamespa

Hi James,

I have eventually found the time to complete the modifications to the spreadsheet model that I created some time ago.

The first sheet gives details and instructions, but if you require clarification then please feel free to ask.

Whilst it was created for a 14kW Ecodan, the effect of changing parameters can still be useful when assessing other models or manufacturers equipment.

The cells highlighted in green can be adjusted, but I would caution against making other changes since many cells contain formula.

 

-- Attachment is not available --

I have now had the time to have a better look at this model which I must say is very neatly laid out, congratulations.  I have a couple of questions

In 'Initial data' I presume I10, I14 etc are just flags signalling which regime its operating in (min, max etc)?

If duty cycle <100% do you make any mods to the calculations to account for start up losses (I think the answer is no, just checking)

I think you assume that heat pump deltaT remains constant, ie the HP modulates the pump speed.  Am I correct?  It seems some heat pumps at least are installed without the capability to modulate pump speed hence the question.

In the playroom I think Im correct in saying that energy (mis-) balance in any given hour is not carried forward to the next hour.  As the energy imbalances are small in the model this wont matter, but its crucial if you are modelling on/off operation (because the energy misbalance comes from/goes to the fabric).  This isn't a criticism just a matter for clarification in case I can use your spreadsheet to replace my current extremely crude and very ugly on/off model (which does carry forward energy imbalances as its the key factor in  modelling on/off).  

Do you have any idea of the physics (or more likely engineering) behind the often small, but sometimes large, variations in COP with load for a given LWT/OAT combo.  I presume its somehow down to compressor efficiency but beyond that I dont know.

Im thinking we should have a heat pump modelling thread where we deposit and discuss the various partial models that people have created from time to time.  What do you think?

Posted by: @jamespa

↑

I have now, in my mind, constructed a rational argument for how to 'model' on/off operation (eg nightime/daytime setback).  Interestingly it is to quite a large extent about the behaviour of the emitters rather than the heat pump itself, and about what choices are made for control during the 'recovery' period, which is where inefficiency occurs.

I note with interest that model is now in quotes! I still think the best way to 'model' the effects of on/off running is by using observational data, ie actually running the whole system and logging the important data, ie energy in, possibly energy out (but only for COP calculations), outside temp and room temp with the system running continuously and then in on/off mode. I think that is all that is necessary - all we really need to know is how much energy we use, and are we comfortable? The room temp is a proxy for the latter (and can be set to our personal comfort level, ie what room temp we want/like).

So far so good, but I am having great fun not setting up a simple modbus room temp sensor. My Midea wired controller is in the airing cupboard, so its room temp sensor is not going to reflect real room temps. I got hold of a SHT20/MD02 temp/humidity sensor (it is without the XY- prefix to MD02, although it was sold as an XY-MD02) and it is a dud all round, and uses the wrong modbus RTU settings (most likely a cheap knock off with the protocol wrongly added). I got it because it is the only sensor I could find that (a) had modbus RTU/RS495 output and (b) would run off 5V ie USB voltages, meaning I could power it using 5V from the RS485/USB converter over a spare wire in the modbus cable. It does power up, and I can get a temperature reading if I use the wrong settings, making it useless, and it is even more useless because I can't change the slave address, which is stuck at the same slave address as the heat pump, and you can't have two slaves with the same address. I will have to return it and see what else I can find.

Come the heating season, once I have a working modbus connected reliable room temp sensor, I am going to experiment with overnight setbacks with a post setback recovery boost. I have been giving some thought to the control logic for this. Initially I was going to do a simple if desired room temp - actual room temp > 2 then new LWT = current LWT + 1 every say 15 minutes but quickly realised this could lead to a Krakatoa number if the actual room temp stayed low. Instead, I either need to put a limit on the max LWT ie add an if current LWT < 60 then (for example) or alternatively, maybe I should set the LWT indirectly, by adjusting the ends of the weather comp curve. This would, I think, prevent the wired controller overriding my direct setting of the LWT, which may well happen, and would in effect mimic me manually changing the weather compensation curve. It would also factor in outside ambient temp, as the curve would be used to determine the LWT, depending outside ambient temp, with the ends of the curve being set a bit higher if the actual room temp was on the low side during the recovery period.

But first I need to get a modbus connected room temp sensor set up and working. I thought it would be so simple... Any suggestions very welcome, but please bear in mind my system constraints: needs to be (a) modbus RTU/RS485 wire connected (not TCP/IP) (b) runs off USB voltages (5V) and (c) doesn't requite a plethora of extra boards wires and software. My set up should simply be:

Oh, and by the way, I agree, this should really be in a thread of its own.

@jamespa

Thank you for the vote of confidence, but as I feel certain that you realise by now, it is still a 'job in progress'. Having discussed various aspects with you over the past weeks has led me to realise that there are further improvements that can be made.

You are indeed correct that the cells I10 to AG10 are used to select the probable operating range of the heat pump (min, mid, nom and max).

Without having empirical data from a working system, it is impossible to calculate what energy loss may occur during cycling, particularly since the frequency of cycling will probably vary dependent upon how the control system has been configured. If anyone has any actual data then please send it forward.

I added the DT calculation only recently, and you are correct that I should also consider a fixed flow system.

The energy balance calculation was again only added recently, and will hopefully lead to the calculation of thermal mass and actual changes in indoor temperature under different operating conditions.

My understanding of the internal workings of a heat are as follows, but if anyone knows better then please clarify.

The heat pump controller sends the required LWT setting to the compressor controller. The controller runs the compressor to raise the temperature and pressure of the refrigerant gas. The hot gas flows into the condenser, where it comes into contact with the cooler plates and hence starts to condense. Because a phase change is taking place, thermal energy is transferred through the plates to the central heating water. The speed of the compressor will therefore be varied to try to maintain the LWT at the desired setting.

The refrigerant gas, in liquid form, then flows through a pressure reduction valve into the evaporator. As the pressure falls, a phase change occurs once more, this time from a liquid to a gas. In doing so it absorbs thermal energy from the outside air.

To produce more thermal energy output from the heat pump, requires the water flow rate and/or DT to increase. To transfer more thermal energy from the refrigerant gas to the water therefore requires a higher gas flow and/or temperature, both of which require the compressor to work harder.

As the refrigerant gas flow increases, this may limit the overall quantity of thermal energy that can be absorbed in the evaporator, and hence the compressor needs to work harder to make up the loss.

I hope you find this useful.