Optimizing LG Therma V Controller settings

[Hogboon]

OBJECTIVE: to reach and maintain the target house temperature of 20º C as efficiently as possible.

Which begs the question ‘define efficient.’ As far as I am concerned, this is the shortest time taken to attain the target temp and the lowest energy consumption thereafter in maintaining it.

 

CONTROLS. As far as I can tell I have no way of controlling power [i.e. electricity] consumption other than through the LG controller.

 

MEASUREMENTS. There is a separate electricity meter for the heatpump. Half-hourly readings are recorded during the schedule as are the outdoor and indoor temperatures.

 

SCHEDULE. The heatpump is turned OFF between 22:00 and 07:00. From 07:00 to 22:00 the target house temperature of 20º C is controlled by a Hive thermostat. The DHW [hot water] is controlled by – you guessed it – the Controller. I occasionally turn it OFF in order not to interfere with collection of house heating data. I am fully aware that this schedule [especially being off between 22:00 and 07:00] is open to criticism but I am the customer here; there are reasons for adhering to this schedule which are IRRELEVANT for the purposes of this topic. I am/we are where I am/we are and I start from here.

 

The INSTALLATION. LG Therma V 14kw monoblok ASHP. Worldheat High Gain unvented tank with system buffer, low loss header and immersion heater. A roof array of 12 x 320 w PV panels a] to supply power to the heatpump b] to heating hot water and c] surplus energy to grid.

 

The HOUSE. 1½ storey 5 bed detached bungalow built c. 1975 described in EPS as 176 m². Double glazed throughout with insulated walls and roof [although could probably do with more in roof] Solid floors to ground floor assumed uninsulated.

 

Existing radiators upgraded at installation of heatpump to 15 in total with new pipework replacing 10mm micro-bore; 22mm flow and return with 15mm drops to radiators. No under-floor heating. No other source of heating other than one small log-burner lit regularly in the evening.

 

Space heating demand: 16,184 kWh per year.

Water heating demand: 3,537 kWh per year.

Total heat loss seems to have been 21,729

 

The story so far – or as they say on telly “Previously…”

 

I altered some of LG’s Controller Installer settings as follows:

 

Outdoor temp for auto mode: min -10º max 15º changed to -1º/15º

Indoor temp for auto mode: min 16º max 21º changed to 18º/23º

LWT for auto mode: min 20º max 50º changed to 30º/50º

 

All other settings were left at their original or ‘default’ settings. Whether these were from the manufacturer or installer I have no idea. [Someone new to this thread might wonder why I do not get the installer to see to any/all of this. Tried that and failed. Also their level of understanding seems even more basic than mine; at least I know what the password for the Installer menu is!]

 

Apart from the general thrust of this post I have a couple of other specific questions relating to the above.

1. Using ‘Auto’ in the ‘User’ menu on the LG Controller what exactly do +1, +2 etc pr -1, 2 etc do? Is this some sort of temperature override or instruction to increase/decrease the thermostat temperature setting?

2. In the LG Controller Installer settings, what should I set for ‘Temperature Sensor’ ? The options are air or water or air+water. It is currently set to water.

 

Here is a graph of the heatpump’s performance from yesterday. Missing data points are those where I forgot or was unable to make a record. The user settings were ‘Auto’ and set a +1

 

[graph not uploaded for some reason. will fix tomorrow]

 

It started off quite briskly but then ‘relaxed’ as if the aim was to arrive at the target temperature mid-afternoon rather than get there a.s.a.p and maintain it there till shut-down at 22:00. What I want is the graph to resemble the FICTITIOUS representation below.

 

[graph not uploaded for some reason. will fix tomorrow]

 

Today’s data was collected with the user’s ‘Auto’ set to ‘Heat’. Here’s the graph.

 

[graph not uploaded for some reason. will fix tomorrow]

 

There is a slight ‘improvement in performance on 28 Nov but apart from that there seems to me little difference between the two. Apart from starting from a slightly lower temperature on the 28 Nov, and reaching the target temperature a little quicker than under the ‘Auto’ setting the curves are pretty similar and display a tendency to delay getting to the target temperature until mid afternoon which is not what I want.

 

Finally 3 sets of data for comparison of kWh consumption for 26-28 Nov. a] to reach 20º C from the starting temps. at 07:00 and b] total power used from 07:00 to 22:00 hrs.

 

	Starting tempº C	kWh to 20º	KWh 7am-10pm	User setting
26 Nov	17.4º	20.4	25.4	Auto
27 Nov	17º	24.81	29.56	Auto
28 Nov	16.7º	24.91	32.37	Heat

 

Apologies for droning on at such length! and many thanks for any practical advice.

[J1mbo]

The + 1/2/3 are a user offset for the target room temperature.

 

The house is being allowed to cool too much. Set the hive, if you must use that, to 18 degrees between 10 and 7.

 

Set the Hive also to 25 degrees in the day.

 

It seems that the heat pump sensor should be water leaving temperature as you have third party thermostat. If the hive is replaced with manufacturer controls then it would be air plus water, air being the room temperature.

 

Spend a few quid on an infrared thermometer if you don’t already have one as the radiator surface temperature is one aspect that needs to be determined.

 

The overall energy demand of the structure in 24 hours at freezing outside is the total heat loss x time, ie 6kW x 24 hours = 144kWh per day. This is reduced somewhat by the night set back. Assume there are 9 hours at an average of 18 degrees then the fabric loss is reduced to 139kWh per day in my example. The heat pump has to supply all of this in whatever operating window it’s run. If it’s run for 15 hours it needs to produce an average of 9.2kW instead of 6kW. Both those numbers are within the capability of the unit fitted but the radiators fitted simply don’t have 9kW of output at 50*C. Therefore, the heat pump operating window has to be increased or the flow temperature increased, and the latter will increase the running cost per kWhr input to the house as the COP will likewise be lower.

 

There is an optimisation trade off between flat 20 degrees inside 24 hours a day (lowest flow temperatures and highest COP as a result) vs night set back amount (lower total fabric loss but higher flow temperatures and so reduced COP). Finding that takes time.

Edited November 29, 2021 by J1mbo

[Hogboon]

Good morning J1mbo! Nice to have you on board again. I have just 'popped in' to upload the graphs here in the order they should have appeared as I seem unable to add them to the original post.

 

 

 

 

I'm currently trying to stay awake r.t.f.m. On p. 20 it says: "Leaving water temperature...Set the desired temp higher than the water temp." What desired temp? The heating temp? The house temp? The water temp? and, come to that what 'water temp' - the leaving water temp? [and is that the same as LWT? re the installer settings?] Attached are all my current 'Installer settings'.

 

installer settings.pdf       

[J1mbo]

LWT = leaving water temperature = the temperature of the water leaving the heat pump, i.e. the radiator circuit flow temperature. It's calculated by the weather compensation curve configured based on outside temperature and adjusted also by the +-1/+-2/+-3 setting in the controller.

[Hogboon]

Got it - thanks! So when the manual says to set the heating temp higher than the water temp., is it the 'LWT' they mean? My current settings are in 'heat' mode for 'Operation' the temp = 50º but the "water heating temp" in Installer settings = 20º-50º Do I need to change anything? [Similarly when operating under 'Auto' the LWT for auto mode is currently set 30º - 50º. Anything to change there?]

 

2 hours ago, J1mbo said:

Finding that takes time.

 

... and everything else!

[ProDave]

Mine has clearly separate settings for "leaving water temperature in heating mode" and "leaving water temperature in hot water mode"  Exact wording may differ but it is clear there are separate settings for heating and hot water.

 

In hot water mode it is slightly different as you set the target hot water temperature, and the leaving water temperature is the maximum you will allow it to run at.  In my case I have set the max LWT in HW mode to 55 with the HW target temperature at 48 degrees.

 

In heating mode I have the LWT set to what I want the UFH temperature to be, so I am not relying on the blending valve in the manifold (which is set slightly higher)

 

 

[Hogboon]

Thanks ProDave. Presumably for the purposes of house heat I do not need to fiddle with any settings for 'hot water mode' ? 

[J1mbo]

@Hogboon, I've put together an hour-by-hour energy model based on 6.5kW heat loss as per the other thread and a hypothetical cold day with two operating scenarios:

 

• The first, on the left, assumes the heat pump is run between 7am and 10pm - 15 hours.
• The second, on the right, assumes the heat pump is run 24x7.

 

 

The house loses heat to the outside whether it's being heated or not. The model shows that that allowing the house to cool overnight (by not running the heat pump) reduces the overall daily energy requirement of the house as would be expected. In this example on this made-up day it would need 147kWh.

 

Substantially all of that energy - 147kWh - has to to come from the heat pump. Because the heat pump is being asked to provide 147kW with only 15 hours of run time, the radiator surface temperature must be increased from 50° to 60° to provide an extra 50% output to nearly 10kW (15 hours x 10kW = 150kWh), instead of the 6.5kW the system was designed and sized to provide. So, to have the house warm with only 15 hours run-time will need bigger radiators or a higher flow temperature.

 

Compare that to a steady-state 20°C target. Over the same 24 hours, very obviously the overall heat demand is higher, in this case 158kWh for the 24-hour period.

 

This is where things get interesting. If the heat supply was a gas boiler then of course it will be cheaper to heat it for 15 hours. It will use less energy and cost less.

 

However, the heat-pump muddies this because it will harvest less energy from outside when supplying higher temperatures. The ratio between electrical supply and heat provided to the heating system is the COP. A COP of 3 would show that 3kW of heat are delivered to the water in the radiator circuit with 1kW of electrical input. The other 2kW are recovered by cooling the air outside.

 

Because of the shortened operating time of 15 hours, the radiator temperature must be increased by 10°C. This reduces the COP, all else being equal. The datasheet for my particular product shows a reduction from 2.39 to 2.07 at -3°. Therefore, the amount of electricity required is actually more when operating the shorter heating period than just leaving it on 24 hours, in case about 7% more electricity to heat the house for 15 hours compared to heating it for 24.

 

TL;DR: 15 hours is not enough time to heat your home with the radiators you have. Mid-season this can be achieved by increasing the flow temperatures to compensate but this will ultimately cost you more to run than leaving the system running 24x7. Run the heat pump 24x7 or change the radiators to achieve 50% more output.

 

Edited November 29, 2021 by J1mbo

[SteamyTea]

2 hours ago, J1mbo said:

TL;DR: 15 hours is not enough time to heat your home with the radiators you have. Mid-season this can be achieved by increasing the flow temperatures to compensate but this will ultimately cost you more to run than leaving the system running 24x7. Run the heat pump 24x7 or change the radiators to achieve 50% more output

Or use supplientary heating for those extreme days.

[Hogboon]

How interesting ! Using less grid power 24/7 compared with 15 hrs on/off is impressive. My aversion to having the heat pump on 24/7 is not entirely on, as it turns out, fallacious grounds of economy, but also partly due to the noise waking me up but does your calculation take into account the fact air temperatures at night tend to be lower than those in day time? so to extract xº heat the heat pump will have to run for longer and/or work harder? The other admittedly small factor favouring my 15 hour on/off regime is that no benefit can be derived from the PV roof panels during approximately half the working hours of the heat pump in continuous operation since when most required, i.e. midwinter, it is dark for approx 12 hours in 24.

 

The other detail that strikes me is that even given my wayward requirements, the perverse behaviour of the system is not explained. There are occasions early in the day on 27 and 28 Nov [and again today] when the heat pump squirts out sufficient heat to raise the house temp by 0.4º If it did that consistently starting from say 16º C at 07:00 it would reach the target temp in about 5 hours rather than the 9 hours it is averaging at present. 

 

For the record, replacing the radiators is a no-no.     

[dpmiller]

can you factor in occasions of defrost cycling?

[SteamyTea]

9 minutes ago, dpmiller said:

can you factor in occasions of defrost cycling?

@joe90 and myself discussed this at length a few years back.

We wondered if there was a way to sense the conditions that would cause frosting i.e RH and temperatures, then stop the ASHP for a set time i.e. half an hour.

But as Joe90 found out  it took 4 years before his ASHP frosted up last winter.

Has it done it again in the last few days @joe90, I am still up country  were it snows and freezes.

[dpmiller]

I'm wondering if this might be some of the differences that @Hogboon has seen in rise time. I'm certainly looking very closely at setpoint vs defrosts currently- threre's lots of time with a misty ambient 2-4C locally...

[ProDave]

I wondered if you could sense when the ASHP was about to ice up and just turn it off and let it defrost on it's own.  But then I observed mine defrosting once, and concluded with the outside temperature barely above 0, it would probably take a couple of hours to defrost on it's own, which would mean it would have to work harder to make up for lost time and ice up again quickly.

[SteamyTea]

2 minutes ago, ProDave said:

wondered if you could sense when the ASHP was about to ice up and just turn it off and let it defrost on it's own

Yes, it needs to happen a proper amount before frosting, not as it happens, that is then too late.

Do you remember if we discussed it here, it on eBuild, I can't be bothered to look.

[joe90]

15 minutes ago, SteamyTea said:

Has it done it again in the last few days @joe90, I am still up country  were it snows and freezes.

No, but then again the room Stat has only (today) called for heat, we lit the woodstove a couple of evenings during the frosty nights and that (with shedloads of insulation) has kept the house temp up to 21’ (room stat set at 20.5’).

[dpmiller]

On mine, the fan speed ramps up noticeably as frosting occurs-  from 85% in "normal" full power operation to 100% when it's blocked up.  But the defrost actually happens on a time and temperature basis. Based on @J1mbo's calculations re 24h operation I think I'll play at reducing what setback I've currently got programmed...

[SteamyTea]

Just now, joe90 said:

No, but then again the room Stat has only (today) called for heat, we lit the woodstove a couple of evenings during the frosty nights and that (with shedloads of insulation) has kept the house temp up to 21’ (room stat set at 20.5’).

But you are still using DHW, which is possibly your largest, continuous draw on power.

It was an exceptionally cold and damp spell last year mind.  I suspect the air was dryer this time as it was a North East wind, not a North West one.

[SteamyTea]

1 minute ago, dpmiller said:

On mine, the fan speed ramps up noticeably as frosting occurs-  from 85% in "normal" full power operation to 100% when it's blocked up.  But the defrost actually happens on a time and temperature basis. Based on @J1mbo's calculations re 24h operation I think I'll play at reducing what setback I've currently got programmed...

I think the problem is, they do not frost up enough to have to worry about it.

It is only us that want to extract the last 2% efficiency out of them.

[dpmiller]

No they absolutely do frost up enough (in our microclimate for sure) and the difference between a light frosting and solid blocking ice is less than ten minutes of operation.

[SteamyTea]

2 minutes ago, dpmiller said:

No they absolutely do frost up enough (in our microclimate for sure) and the difference between a light frosting and solid blocking ice is less than ten minutes of operation.

Then maybe the idea Joe, Dave and myself had could be worth a revisit.

Should only need three Temp and RH sensors, external temp/RH, heat exchanger temp, and after heat exchanger temp/RH. When I was heat exchanger, I mean the radiator that the air gets blown though.  Then a dew point predication algorithm that has a built in offset i.e. 3% above dewpoint, or 1 K above dew point, that turns the refrigerant pump off, but keeps the fan going to blow out moisture as it would not be 100% perfect.  Then after a set time, it turns it all back on again.

[dpmiller]

complex to implement as you're already at the whim of the inverter with all it's sensors and lookup tables.

[SteamyTea]

3 minutes ago, dpmiller said:

complex to implement as you're already at the whim of the inverter with all it's sensors and lookup tables.

Yes, basically means junking all the existing control system.

May be easier to make it work on a non inverter one.

I have often wondered how hard it would be to put a speed controller on the fan, I am not sure how much difference it would actually make.

One of those 'retirement project'.

[dpmiller]

On mine (which is built on a standard Carel platform as used in retail fridges etc) the fan, compressor, expansion valves etc are all under constant control, and the characteristics of the compressor and gas are programmed into it. I've no doubt there's some map in it that you could alter to improve things at the frost-point (or maybe it already does it, dunno)

[J1mbo]

 

3 minutes ago, dpmiller said:

I've no doubt there's some map in it that you could alter to improve things at the frost-point (or maybe it already does it, dunno)

 

Maybe "SuperChips" will move into ASHP remaps since the ICE chip market is presumably evaporating

Edited November 29, 2021 by J1mbo