Posted by: @mookyfoo

The main disadvantage for me is all this glycol in the system meaning if we want to drain down we need to some how save it. 

There are also small but cumulative losses eg cleaning a magnetic filter (though I suppose one could catch and recycle the water) and bleeding rads. In a closed circuit filler with a fluid, it doesn't take much in the way of losses to depressurise the system, and there is no convenient way to top it up.

My primary circuit has gone from about 1 bar at commissioning to 0 bar now (ie it is not pressurised). I'm not sure whether this matters. If it does, there is no way I can see that I can re-pressurise it. The secondary circuit on the other hand is unvented and so can be re-pressurised from the mains. 

@derek-m (and others) - readings with (at 1200 noon yesterday) weather comp turned off (Preset Temperature > Weather Set Temp > Zine 1 H-Mode High T > On => Off) and leaving water set to 55 (Main display, LHS, number here can now be changed, and it has a water droplet icon to the left of the number. According to the user manual, this sets the 'water flow desired temperature' which I take to be the LWT, but it might not be, see results below) Room stat (in kitchen) is set to 21.0 degrees:

Yesterday evening at 2000h: LWT 59 RWT 52 Current 13A Capacity 8.95kW (COP 2.9) Ambient 8 Bedroom 20.0 Bathroom 22.5 Kitchen 20.2 

This morning at 0800h: LWT 57 RWT 50 Current 12A Capacity 8.1kW (COP 2.8) Ambient 5 Bedroom 20.5 Bathroom 23.0 Kitchen 20.3

The house feels warm, and most of the rads are warm/hot to touch.

Comments:

(a) the actual LWT is above the set LWT (assuming the number on the main display sets the LWT)

(b) the COPs are poor but not disastrous, despite LWTs over 55 degrees and approaching 60 degrees at times

(c) with this set up, LWT set to 55 (and reaching over 55 degrees, the rooms get to above design temps at 5 degrees ambient. Whether the system can maintain this when it is zero outside and defrost cycles kick in remains to be seen

(d) from @derek-m's Midea 14kW spreadsheet, at 5 degrees ambient, my heat loss (demand) should be around 8.1kW. It looks like the heat pump output is at or just above the demand, which makes sense, the rooms are just above design temps. It indirectly also suggests the heat loss calcs weren't a long way out, 8.1kW of predicted demand matched by 8.1 and a bit kW of output gets the rooms to design temps.

(e) 12A - 13A is give or take 3kW, The heat pump appears to be on pretty continuously, so that works out at about 72 units a day of electricity consumed. At a current average unit price of 25p, that amounts to £18 per day (£126 per week, over £500 per month. And that is in early April, not January). Ouch.

Taking everything together, it seems that at ambients of 5 and above, my 14kW heat pump can meet demand, but, with the current settings, very high costs. Running in weather comp mode appears to derange heating performance, but is probably a bit more 'efficient'.

For the time being, I have now set the LWT to 50 degrees (fixed, no curve). 

What next? Maybe try a custom weather comp curve? 60@-2/40@15? But with always high LWTs, aren't I condemning myself to very high electricity bills?          

Hi @cathoderay

It is now obvious that the weather compensation slope needs to be adjusted to accommodate the higher LWT that is necessary because of the PHE being installed. Optimising the weather compensation should help improve efficiency.

One probable way to improve efficiency and reduce running costs would be to have the PHE removed.

Another way is to improve insulation and draft proofing.

@derek-m - yes, an adjusted curve makes sense, maybe something along the lines I suggested in my last post, 60@-2/40@15? By the bye, the rads seem to be about 10 degrees below the LWT. With the recent LWTs in the high 50s, a good number of the rads were in the 40s, others in the 30s. The rads also heat up better, front panels as well as rear panels, and they are easier to balance.

I also wondered about a fixed LWT of say 45 degrees, where COP becomes tolerable, and the heat pump has better nominal outputs (more kW). It would run more like an on/off fossil fuel boiler, and I think it has been suggested that with modern ASHPs this may not be such a crazy idea. The whole thing about weather comp is to reduce LWT when possible to improve COP, but if I select a fixed LWT with a tolerable COP, then maybe that can work?

Posted by: @derek-m

One probable way to improve efficiency and reduce running costs would be to have the PHE removed.

I was thinking about the PHE yesterday. If we think in terms of % loss (of energy), rather than absolute loss of degrees (a proxy for energy), and say, for example, the losses are 10%, ie the PHE is 90% efficient, then that lost 10% has to go somewhere. If it is 'lost' inside the house then it isn't really lost, it ends up where it is meant to be, even if it doesn't reach the rads. I can't really figure out where, and how, the loss happens. Or perhaps the loss is more like a blockage that obstructs flow, or a resistance that resists flow, but resistors tend to heat up.

On balance, considering the pros and cons, I tend towards keeping the PHE because of its primary purpose, separating primary and secondary circuits. This protects the primary circuit from heating circuit crud, reduces the amount of lower specific heat capacity antifreeze in the system, and means that bleeding rads and cleaning filters on the secondary circuit doesn't drain the primary circuit, and any deficits in water can be topped up quickly and easily from the mains.

Posted by: @derek-m

Another way is to improve insulation and draft proofing.

Yes, necessary, and in hand, albeit at a slow pace. I'm toying with the idea of getting a thermal imaging camera to identify cold spots, but the price of a decent one has put me off. 

@cathoderay 

Often these days you can borrow a thermal camera, lots of organisations have them.  Eg. Cambridge Carbon Footprint will lend one out for free, as I'm sure other places will.  They're best used in cold weather, and not in sunshine or just after it's been sunny.  You could probably just use one now very late evening after a cold day when it's dark, but best results will be in winter.

Identifying, and eliminating draughts can be done any time of the year, and is usually easier than insulating.  While some ventilation is needed, "airtight and ventilate right" should be the motto - ie. deliberate ventilation of the level required, where it is needed, by opening windows or trickle vents.  It's likely more ventilation is needed when more people are home, and in colder weather there is more "stack effect" forcing air circulation due to warm internal air being lower density than external air, so most UK homes have plenty of ventilation as it gets colder!  A windy day is a good time to identify them - use the back of your hand near likely places to feel for them, or drop a feather, watch how it falls.  There's a great website dealing with them here:

https://readinguk.org/draughtbusters/   

Hi @cathoderay

Looking at the spreadsheet that I provided you will notice that some of the cells in the top row are highlighted in green, these are adjustable, and can be used to perform 'what if' assessments. For instance, if you were to improve insulation and reduce your heating demand from 12400W to say 8000W, you will see a lowering of the LWT with quite an improvement in efficiency.

It would appear that the major problem with your system is that the temperature of the water arriving at your radiators is possibly 10C lower than that leaving the heat pump, which of course means that to meet the heat demand, then the LWT needs to be much higher than normal, and in turn reduces overall efficiency. Running your heat pump at a fixed 45C or even 50C would mean that in the Winter months it would struggle to meet the heating demand. Running your system on weather compensation, with correctly optimised settings, would probably be more efficient and cost effective.

Any loss of energy from the PHE will be within the envelope of the building, so will not really be a loss. The reduction in efficiency is at the heat pump, because it has to increase the temperature of the LWT to compensate for the temperature gradient across the PHE.

I would check with your energy supplier, to see if they have a loan scheme for thermal imaging cameras. Failing that it may be cheaper to hire one.

Posted by: @derek-m

A possible solution, which should be quite easy to incorporate, would be to have timing functions built into the controller, which would prevent the heat pump from restarting until a suitable period of time has elapsed, but then also keep it running for a minimum period once it has started. Whilst operating in such a manner may cause the LWT to vary outside optimal limits, but I doubt that it would have that much effect upon actual indoor temperatures.

My add-on Raspberry Pi controller does implement just such a thing, with a next_permitted_change countdown timer, currently generally one hour minimum between starts and stops, with shorter periods allowed in some conditions. Of course, if the heat demand is too low, then the pump controller itself will start cycling sooner than that, but I've tried to set the curve to avoid that. The warmer-weather end of the curve is unnecessarily slightly higher than optimal: the ecodan FTC allows two gradients on its "curve" as well as the flats at the sides.

I forget what LWT is (except for London Weekend Television which doesn't seem relevant here) but I agree it doesn't have much effect upon actual indoor temperatures. I think that may be because the thermal inertia means temperatures vary slowly, often continuing to rise for long enough to bridge a short stop that maybe wasn't needed, ideally. According to my monitor, the last time it happened was 8 March, the 90 minute stop saved more energy than the restart used and the house was maybe half a degree colder than the target range for 40 minutes... so it doesn't make the wrong decision often, the penalty seems minimal and I feel the benefit outweighs that.

Maybe the main reason this sort of thing isn't implemented is because it would be another thing for installers to configure and potentially get calls about if not quite correct? Already, we seem to have to choose between compensation curves and thermostats or at least be very careful about compatibility. Adding countdown timers to the mix seems unlikely to make things simpler. I suspect the next addition to heat pump controllers will be "Smart Grid" features because the incentives to both user (save £s) and suppliers (maximise use of current grid) are too great not to do it.

Hi @mjr

Having read through quite a number of manufacturers manuals, it would appear that starting and stopping more than 6 times in a 1 hour period is denoted as short cycling. It would be quite a simple matter to have the controller program to have minimum operating periods set as standard.

During mild weather, once the heat pumps starts running, it could be kept running for a minimum period of say 6 minutes at minimum capacity, even if the Leaving Water Temperature (LWT) exceeds the calculated required value, then after the 6 minute running period it would be prevented from restarting for at least 6 minutes.

@derek-m 

For older heatpumps or fridges that don't have an inverter, starting up the compressor is quite stressful, the unit literally twists on the rubber vibration mounts as it starts.  The stress of startup fatigues the fixed copper pipe brazed joints, and eventually this will cause failure - so the rule of thumb was 6 starts per hour and the unit would have a long life.  

In addition, there is a separate wind on a small single phase compressor that is used to persuade it to start, often switched out with a PTC resistor (Positive Temperature Coefficient so as it gets hot the current drops).  If the compressor is stopped and quickly restarted, the PTC is still hot so the start wind does not engage - resulting in a locked rotor and blown fuse.

My small gshp is as described above.  To definitely prevent short cycling, I have a 5min delay on startup, and a 2 min pump over-run (but not compressor) afterwards.  In practice though, the start-stop-start period is around an hour.  I expect this time period is proportional to the volume of water, the delta hysteresis temperature of the thermostat used, and inversely proportional to the heatpump power. 

All the above is old hat though - perhaps relevent to many smaller gshp still, but I wouldn't expect a modern ashp with inverter drive to be difficult to startup.  I think they tend to have rotary scroll compressors with a 3 phase inverter drive, and would have a very gentle startup. 

I'd be interested to know what start-stop-start periods are found with modern ashp?  I assume now it's getting warmer, they will be clicking on/off now. 

Posted by: @robl

I'd be interested to know what start-stop-start periods are found with modern ashp? 

I think the worst I had was 15 minutes between starts. I expect it's no big stress for the system, but it really dents the efficiency, which is why I set the 60 minute minimum.

Thanks for the tips about detecting cold spots and drafts. On a windy day there are place near doors/windows where I can feel a draft, and during gales the curtains have been know to twitch. I've also used a lit candle. Not all of these drafts are easy fixes eg some round windows are badly fitting frames, even a bit of rot here and there. The necessary repairs are substantial and take time. I'm also slowly adding secondary glazing, but that can also be time consuming eg the window furniture projects beyond the frame, so the frame needs a spacer made up, added and painted, slow fiddly work.

@derek-m - according to the heat loss calcs, most of the heat loss is through the walls and metal single glazed windows. Adding secondary glazing does reduce U-values (measure of how much heat gets lost through a material/entity) quite well for the windows but they are small. The walls are relatively much larger, and ~80% of them are old solid stone walls. Being a listed building limits what insulation I can add (which I am OK with, external insulation would ruin the character of the building, and internal would encroach on what is already a small interior footprint. Both are also very expensive.

Possible reductions are not great, getting down to ~8kW total loss from12.5kW isn't really an option. Playing with the Freedom Heat Loss Calculator, going from all metal single glazed windows to all secondary glazed drops the U value from 5.8 to 3 but the total loss only drops by ~0.5kW because of the small windows. On the other hand changing the wall type from solid stone to modern cavity walls has a dramatic effect, the total heat loss almost halves to around 6.5kW. It is worth noting that almost all the homes in my local area are of similar construction, and a good number are listed. Old leaky buildings are a fact of life, with no reasonable and affordable modern solution. We have to go at the problem in other ways. Expecting them to behave like a modern build is a bit like expecting a pensioner to be as fit and able as a teenager. Even if you are a fit and able pensioner, you will still not be as fit and able as you were when you were a teenager!

Am working on the heat pump settings, will report back when I have more data.           

@cathoderay I've said it a few times on here. Insulated plasterboard might be your friend in this situation. It's very low profile and will significantly add to reducing your heat loss. However it only makes sense if you are redoing a room