Wind chill in action. The OAT has warmed up a lot over the last 24 hours but my heat pump is still failing to get my house back up to temperature. The warmer OAT has caused the set LWT to drop but even so there is normally a bit in reserve (plus my auto-adaption script) which will bring the IAT back up to where it should be, 19 degrees, but not today: the IAT has stayed stubbornly at around 17.2 degrees. I blame Storm Bertie...it has been wet and windy all day, with significant wind chill. 

Fascinating.  Referring to your first post, is it a cycle or a defrost.  It looks like the latter but maybe Im misreading.  Can you confirm

Understanding this may be key to understanding your heat pump performance.  Is it flow rate (heat not taken away quickly enough) or something else?  I'm presuming the chart is energy at the heat pump.

It looks to me like it reaches and then overshoots the design LWT then switches off (presumably because it has exceeded the design LWT by more than the set hysteresis).  This suggests that, at the design LWT, insufficient heat is being carried away.  This could be flow rate at the heat pump or flow rate at the PHE/LLH/Buffer tank (I think you have one) or capacity of the emitters.  Would any of that make sense?

According to the engineering data I have the minimum output for the MHC-V14W/D2N8-B MHC-V14W/D2RN8-B at 0C and 50C FT is 6.6kW, which is suspiciously close the the ~7kW you are observing.  If so then its possible that it is cycling because it is operating at min output, at which value the target FT is exceeded because insufficient heat is being carried away, despite the fact your house isn't warm enough.  In this case you need to look to the rest of the system not the heat pump itself for the explanation (and cure)

Referring to the second chart, it took my house 2 days to adapt to the sudden chill (ie it took 2 days for my energy consumption to rise and level out) and so I wouldn't be surprised if it takes 2 days to react to the sudden thaw.  Wind chill is certainly also a factor, but what you are observing could equally be caused (at least in part) by the fabric heat capacity.   

We only hit -3C and 65kWh electricity in 24hrs, a couple of days ago.

We were using the wood burner too since we’ve got loads of scrap wood to burn, which costs less ££.

Since we’ve got low air leakage the wind can only blow our heat away from the outside surface chill (and that can be bad enough). With lots of leakage gaps the wind will blow your nice internal warm heat right outside.

Posted by: @jamespa

↑

Referring to your first post, is it a cycle or a defrost.  It looks like the latter but maybe Im misreading.  Can you confirm

Defrost cycles are ones in which the LWT falls below the RWT, ie the circulating fluid loses heat as it passes through the heat pump. Thus all the cycles in the first chart in the thread are defrost cycles except those between 1200 and 1500 on the 21st, when the OAT (AIT) rose a bit, and a couple of cycles on the right hand side of the chart. In the wind chill chart, the cycles on the left when it was cold outside are defrost cycles, those on the right, after it had warmed up a bit outside, are normal cycles. Interestingly (to me at least), the only time the heat pump manages something approaching steady state running is when the OAT is at the upper end or just above the defrost range, see the LWT trace between 1500 and about 1630 on the 21st, and the two periods after 0300 on the 22nd. The actual LWT reaches the set LWT, and stays there, with the RWT 5 degrees below the LWT, ie it is running by the book, as it is supposed to do. The mystery for me at least is why it doesn't do this all the time when it isn't defrosting, and instead uses 'normal' cycling as a way of modulating output. As I have noted before, the eyeballed average of the actual LWT is usually close to the set LWT, so it is putting out the right amount of heat, but by way of cycling, rather than running at a steady state. One possibility is that this may be more common than is generally supposed, because most data is collected hourly, and that (whether spot or trailing average) will very likely do a lot to smooth out cycling. The temperate variable in the plots I have posted are minute data, so we see minute by minute variations.     

Posted by: @jamespa

↑

I'm presuming the chart is energy at the heat pump.

Yes, all the variables except the MD02 IAT data come from the heat pump. The IAT comes from an independent sensor at chest height on an inside wall in the middle of the house. 

Posted by: @jamespa

↑

It looks to me like it reaches and then overshoots the design LWT then switches off (presumably because it has exceeded the design LWT by more than the set hysteresis).  This suggests that, at the design LWT, insufficient heat is being carried away.  This could be flow rate at the heat pump or flow rate at the PHE/LLH/Buffer tank (I think you have one) or capacity of the emitters.  Would any of that make sense?

Very possibly, or alternatively it is just how Midea (and possibly others, see note above about cycles getting lost in hourly data) heat pumps work, by using cycling as part of their modulation strategy. Yes I do have a PHE, and I have long wondered whether this acts as a choke on the system. I presume there is a way to work out how much energy a PHE can transmit, but I have't found it, and even if I had, I am somewhat stuck as the PHE is a true Headroom Heat Pumps black box, there is no way of knowing what it in the black box, nor is there any way of knowing for certain what the flow rate is in the secondary circuit. The only things I can measure with OK-ish accuracy is the pip temps at the PHE entry and exit points, but these are far from ideal as they have to be read manually and are not simultaneous, to do that I would have to get a six pack of temp sensors and attach four of them to the pipework, and then find a way to collect the data. Emitter capacity even at low rad/room delta t values should be adequate, they are generally slightly over rather than undersized, eg if there wasn't a rad size that matched a room's heat loss, I went for the next size up rad, though I do wonder what actually happens when the LWT (and so rad/room delta t) is low. Can they still emit enough heat? This is another view of our old friend the heating as a complex dynamic system, change one thing and everything changes.

Posted by: @jamespa

↑

According to the engineering data I have the minimum output for the MHC-V14W/D2N8-B MHC-V14W/D2RN8-B at 0C and 50C FT is 6.6kW, which is suspiciously close the the ~7kW you are observing.

I have the same data, and agree. here is the relevant table:

which also shows it can (supposedly) put out something over 11kW on 'max' but I have not found anything which details how the heat pumps decides to operate at 'min', 'norm' or 'max' output, let alone how it is implemented. By hunch it might be by increasing the flow rate, but I have not found any way to control that, instead it is set inside the black box that is the heat pump controller. One observed characteristic of the flow rate is that it remains mostly relatively fixed, though at times it does vary a bit. It does not appear to be used much to control output.

@judith - good to know I am not the only one watching ten pound notes being blown into the garden like confetti. I too have a wood burner, only it is an inglenook rather than a stove. It is an extremely attractive feature of the house which when it has a fire burning in it instantly transports me back to the 18th century but it does burn a lot of wood, and of course sucks a lot of warm air out of the house, so I don't routinely light it (and being a true open fire, it needs active management and a watchful eye kept on it). But more to the point, given my empirically determined heat loss, and Midea's claimed performance data, my heat pump should be more than able to heat the house even at low OATS, But in practice, it can't quite cope.

There is no doubt in my mind that a stiff breeze blowing across wet sandstone cools it, which must mean heat loss is greater, just as the sun on still days can warm it and reduce heat loss. I haven't actually done an empirical heat loss assessment during a spell of wet and windy weather, it might be interesting to do one and see how the loss compares with more standard conditions. Spreadsheet calculators of course make no allowances for such things, another reason to give less salience to their conclusions.

Posted by: @cathoderay

↑

which also shows it can (supposedly) put out something over 11kW on 'max' but I have not found anything which details how the heat pumps decides to operate at 'min', 'norm' or 'max' output, let alone how it is implemented

They decide is by measuring the flow/return temperature.  The control loop adjusts the output of the heat pump (within the limits it can achieve) to get the set FT.  If the heat pump cant go high enough then it will undershoot the set FT but probably keep going.  If the heat pump cant go low enough then it will overshoot the set FT and, once it reaches the max allowed overshoot, will start to cycle.  The exact algorithm of the control loop isn't disclosed, but that is the principle for sure. 

The behaviour you have illustrated is entirely consistent with the the heat pump not being able to go low enough for the load it is seeing.  Since it is in fact capable of going higher and since you need it to go higher this very strongly suggests that insufficient heat is being taken away from the pump, ie the limitation is downstream of the heat pump itself.  So that's where we need to look.

Posted by: @cathoderay

↑

Yes I do have a PHE, and I have long wondered whether this acts as a choke on the system. I presume there is a way to work out how much energy a PHE can transmit, but I have't found it,

There is something called a kW rating, but I dont know how to interpret.  Others might.  However the first thing I would try is to turn up the flow speed on (sequentially) the emitter side of the PHE and the HP side of the PHE.  That will take more heat away.  If you have the data do the calculation on flow rate x deltaT either side of the PHE.

If, for a moment, we believe the Midea data (why wouldn't we) then the broad brush explanation seems pretty obvious, and its just a case of chasing down where in the downstream system the limitation is occurring.  If it turns out to be wrong we will need to look elsewhere, but if/until that is shown to be the case there is little point in doing so IMHO.

This really was a road test opportunity - I learned a bit about my system.

First, I was stressing about 22kwh electricity usage - I know it's to be viewed as a part of a whole (year) but even so... 

I have no way of monitoring except via my smart meter and the Octopus app - thus all kwh numbers are for everything. Cooking is the big juice muncher apart from the heat pump. 

My house (similar to many modern houses) has no internal brick walls so only has external walls for a 'thermal store'. I can often feel a difference very rapidly when the pump turns off. (Plus although my loft is ok, my wall insulation is patchy). 

Defrost cycles in the night chewed up lots of kilowatt hours [😩] until I learned a trick (which may be Daikin specific) - make defrost settings 'Discontinuous' and limit to 60%.  That helped noticeably!

Curve tweaking - there was heavy snow nearby up on the moors and windchill off the sea. It snowed here too but disappeared in the rain. OAT was below freezing for a few days. Heavy frost when sleet stopped. 

Daikins famously always rev up at the beginning of a heat rise after set-backs so I tried cutting my set back and upping my curve to 40º so it didn't struggle. Even a 1º setback change makes it fire up when heat is called for in the morning. There was no change in consumption except up, slightly - it was still higher [22kwh] than my average 10-12kwh per day. 

I wasn't cold throughout this - house stayed at 21º -22º 

Finally, I dropped it back to 38º (@ -3º) and lowered the modulation to 3 - I have a log burner and I bulk season my own wood so decided if the HP system couldn't heat enough - I'd back it up with the burner. One smaller living radiator definitely needs changing from type 21 to type 22 but this is only apparent when it is windy. (Which, being near the sea and on a hill, is quite a regular occurrence.) 

It was fine and by Friday total house consumption was 16kwh - I lit the burner anyway. Meanwhile, it seems to be a sweet spot because as the temperature has risen it is now happily bimbling along at 29º and no cycling at all. 

My system, give or take a tweak or two, seems to be running quite well. 

I note lots of cheap(er) rate fixed deals came on offer when it was freezing and Cornwall Insight gave 'price cap rising' comments all over the media. The big boys are after the Octopus punters quite ferociously it seems. Tempting...

@cathoderay this looks similiar to a phenomenon I experienced while setting up my weather curve.

I fixed my issue by slightly increasing the flow temperature on the weather curve. The only explanation I can give is as the room temp increased it reduced the rads ability to transfer heat. My delta T went down (set to 5C) and the pump tried to increase the flow temp thus exceeding the weather curve limit which caused it to start cycling. The temp in the house then went down due to a lower average LWT.

Now what seems to happen is, when the room temp gets close to the desired level the heat pump has a little more headroom to increase the LWT to maintain delta T then simply backs off slightly due to the reduced rad output but stays in a steady state balancing heat in with heat loss.

Not sure if that maybe your problem but it maybe worth temporarily increasing the LWT a couple of degrees to see if the cycling settles down, if not return to original settings, worth a try?

Posted by: @technogeek

↑

Not sure if that maybe your problem but it maybe worth temporarily increasing the LWT a couple of degrees to see if the cycling settles down, if not return to original settings, worth a try?

Yes, I agree, and also agree that the the combination of low rad temps and warmer but not yet up to desired temp rooms reduces heat transfer. For example, this morning my set LWT was 31, against a room temp of 17.5 degrees, not a lot of delta t.

I was going to manually up the LWT bu raising the right hand end of the weather curve, and then got distracted by asking myself how to automate this. I need to find a signature for this behaviour, and it is probably a major rise in OAT after a cold spell, see this chart for the last week which shows the slow fall in IAT during the cold spell, and then the failure to recover when the OAT warms up. You can also see how fast the set LWT (orange line) fell from over 50 to 31 degrees this morning:

Posted by: @jamespa

↑

The behaviour you have illustrated is entirely consistent with the the heat pump not being able to go low enough for the load it is seeing.  Since it is in fact capable of going higher and since you need it to go higher this very strongly suggests that insufficient heat is being taken away from the pump, ie the limitation is downstream of the heat pump itself.  So that's where we need to look.

I agree - the choke effect. Putting aside the PHE as the possible villain for the moment, and looking at flow rates as your suggest, I have a very basic secondary circulation pump with 'traffic light' LEDs that show what mode it is in but nothing else. It is a Grundfos UPM3 Auto 25-70 130 currently operating in CC3 mode. This is 'constant control' mode (max), and constant control means 'the pump runs on a constant curve, which means that it runs at a constant speed or power'. An adjacent table shows this constant power as head in metres, which in the case of this pump size and setting is 7m, more than enough to blow the skin off a rice pudding, but it doesn't tell us the actual flow rate. Later on in the manual we have this, which to my eyes is as clear as mud, since it suggests for 'CC' mode (solid lines) the head falls (on the Y axis) as the flow rate increases (on the x axis - why this way round, surely the flow rate is the dependent variable): 

Presumably the upper solid line in the upper chart is the one that applies, but what it tells me is beyond me. Nor am I sure what metaphorical knobs to twiddle to increase the flow rate...

I do know what the primary flow rate is, it is collected as part of the minute data, and is typically around 1.4m^3/h, but I don't think there is any direct way of changing it, though there may be an indirect way. I'll have a look at the modbus register address tables and see if there are any likely candidates.    

Im not an expert on pump curves (or indeed very much else) however there is a good guide to reading pump curves here https://www.grundfos.com/uk/learn/ecademy/all-courses/the-basic-pump-curve-course/how-to-read-a-pump-curve-

I think the way to look at this is to start with the flow rate that is required.  That's simply power transfer requirement/(deltaT*specific heat capacity of the transfer medium).  So in the case of the primary, with a deltaT of about 4, and assuming no glycol, you need (for 11kW) 11000/(4*4200) l/s = 0.65 l/s = 2.3cu m /hr.  At a deltaT of 5 its 1.9cu m/hr  With glycol a bit more flow is needed, because the specific heat capacity of a glycol/water mix is a bit less.  If you are only getting 1.4cu m/hr its not enough to carry 11kW at a deltaT of 4 or 5.

The same calculation can be done for the secondary, but as I don't have the deltaT for the secondary to hand I cant do it.

Once you know the flow rate you can read the head of the pump from the pump curve, and so long as that exceeds the head required to push the fluid round the system, it will work.

You may want to check your flow rate estimates, but this is looking increasingly like the limiting factor in the system.

Posted by: @jamespa

↑

I think the way to look at this is to start with the flow rate that is required.  That's simply power transfer requirement/(deltaT*specific heat capacity of the transfer medium).  So in the case of the primary, with a deltaT of about 4, and assuming no glycol, you need (for 11kW) 11000/(4*4200) l/s = 0.65 l/s = 2.3cu m /hr.  At a deltaT of 5 its 1.9cu m/hr  With glycol a bit more flow is needed, because the specific heat capacity of a glycol/water mix is a bit less.  If you are only getting 1.4cu m/hr its not enough to carry 11kW at a deltaT of 4 or 5.

I think this is very pertinent. I believe the Midea reported flow rate (of around 1.4cu m/h) is about right because there is an anlague flow meter in the primary circuit and it shows similar values (in l/m, around 25, awkward to read as the gauge is buried in airing cupboard pipework). Elvis may have left the building, but he's not moving fast enough.

However - I don't think there is any way I can control the primary circulating pump flow rate. I can read a modbus register address to get its current value, but that register is read only, and there is no corresponding write register in the modbus register address tables. Even if there was, I think any value written by me would be very quickly overwritten by the Midea control logic. Or maybe not, and I might end up with an infinite positive feedback loop... 

Posted by: @cathoderay

↑

However - I don't think there is any way I can control the primary circulating pump flow rate. I can read a modbus register address to get its current value, but that register is read only, and there is no corresponding write register in the modbus register address tables. Even if there was, I think any value written by me would be very quickly overwritten by the Midea control logic. Or maybe not, and I might end up with an infinite positive feedback loop... 

If this is the case then its a bit weird its not enough particularly as you presumably have a primary that is pretty low resistance (or is it?).  So the questions to ask are 

• Does Midea attempt to control it
• If so are there any adjustable parameters
• Does the pump itself have an adjustment
• Does the pump have sufficient head for your primary circuit (almost inconceivable it doesnt but...)
• is there any flow restriction in the primary circuit 
• Is the primary circuit plumbed in 28mm pipe
• Have you any way to increase deltaT (perhaps by speeding up the secondary pump)
• Is the PHE of sufficient capacity (not sure how to judge)

Something is throttling the velocity and thus the output of the pump.  Until that is fixed it cant do what it should do.