Posted by: @derek-m

↑

Posted by: @benseb

↑

Also, since this is ultimately cost Vs comfort, I 'ought' to build in the effect of time of use tarrifs.  I'm definitely not going to do that,

With the new Cosy tariff (and various EV tariffs ) available, it would be great to do this..... I understand your calculations are about the energy needed, but our off-peak rate is about 5x cheaper than on-peak so it's no longer about efficiency and energy use, for us it's about cost.

So would be great to get some modelling on whether we can take advantage of these tariffs. For example our off-peak is 23:30 - 5:30 which is pretty much when we'd have a set-back so does it make any sense to run hard between say 04:30 - 05:30? Probably, but be good to model...

I am looking to model the possible energy usage during different hours of the day under different operating conditions, you should then be able to calculate the cost.

 

Sorry, I wont be doing this because tariffs change all the time.

But... if night time set back saves 10-20%, but you can somehow store energy at 20% of the cost during the night and use it during the day, I think its pretty obvious that storing the energy is going to work out cheaper.  What will matter here is

a) how much energy can you store without the house (or at least the bits you are in at night) becoming uncomfortably hot?

b) how controllably can you release it?

This is more about the house than about the heat pump, and very akin to a storage heater problem.  In a house with UFH downstairs and rads up, living room downstairs and bedrooms up, and the UFH in a decent size insulated slab, letting the rads cool but heating the UFH at night is almost certain to be a winner if the night time tarrif is that low.  The exception to this might be on days where OAT varies dramatically from very cold at night to very warm during the day, but thats rare and still extremely unlikely to wipe out a 5:1 price difference.  If you want figures you will have to model it!

Probably the most important discussion in heatpump usage. Nearly all installers just put in stats and leave them turn off and on the heatpump  at will. A good few people i know who run ufh looked very confused when i suggested getting rid of zones and running all day at lowered temps. Last winter i ran off stats for my newly built bungalow. This winter with many helpful suggestions from derekm, i will trial adjusting the installer set wc curve and running it all day at much lower flow temp to achieve a constant 21deg throughout the house. Going to setback at 10pm nightly until 5am and see how that goes once i have achieved correct flowrates/temps to get all rooms to a stable21. Great discussion by 2 very knowledgeable posters. Any home owner with heatpump would do well to join this forum.

I am going to try to keep any posts of mine about this topic, on/off vs continuous running, in this thread. General recap of my thinking so far:

(a) observational data always trumps modelled predictions (aka as forecasts, whatiffery etc).

(b) that said, there may be a place for using models to do some whatiffery on extreme examples eg what happens if I run my heat pump with a LWT of 95 degrees. The problem with this sort of whatiffery is that it involves extrapolation beyond known data. Maybe heat pumps explode when the LWT gets to 90 degrees, but 'the model' doesn't know this, and confidently predicts what will happen at 95 degrees based on what happened at a LWT of 60 degrees. The heat pump owner then dials in a LWT of 95 degrees, and in short order ends up with a red face and cold ankles.

(c) in steady state conditions, OAT constant and IAT (indoor air temp) constant, the energy/heat loss from the building is the same as the energy/heat supplied to the building.

(d) the amount of energy/heat lost and supplied in steady state conditions depends on both the OAT and IAT, or more accurately the difference between the OAT and IAT, what we may call the dAT (delta air temp). A lower OAT at the same IAT, or a higher IAT at the same OAT will lose and need more energy.

(e) we generally calculate energy heat loss at a low OAT and a comfortable IAT. To keep the sums simple in this run through, I am going to use an OAT of -2 degrees, and IAT of 18 degrees (dAt 20), at which point the energy/heat loss is 10kWh.

(f) we assume there is a negative linear relationship between OAT (and dAT) and energy/heat loss. The above heat loss is 10kWh at -2 OAT, falling to 0kWh at 18 OAT; at 8 OAT the loss will be 5kWh

(g) however the energy required/used does not follow exactly the same line, because a heat pump becomes less efficient as the OAT falls (and the LWT, set by weather compensation, rises). This may be an  example - correct me if I am wrong - what @jamespa means by second order effects. However, it is relatively easy to accommodate this if we know the COP. Let us say our heat pump has a COP of 3.5 at 8 OAT, and 3 at -2 OAT. Thus at 8 OAT, we use 5/3.5 = 1.43kWh of energy, while at -2 OAT we use 10/3 = 3.33kWh. The energy out has gone up by a factor of 2, but the energy used has gone by a factor of 3.33/1.43 = 2.33. In absolute terms, this increase might seem quite small (0.33), but in percentage terms it is not trivial (16.5%). At lower OATs, we use more energy per se, and then a bit more because our heat pump is less efficient at a low OAT.

(h) the above pretty much describes continuous 24/7 running: given a set and maintained IAT, energy use will depend on OAT (or more strictly the dAT); and when the when the energy/heat loss doubles (as in the above example), the energy used will increase by doubling and then a bit for for the lower COP.

(i) let us now consider a setback. Because we call it a setback, it implies the setback is done by lowering (setting back) the thermostat. Let us say that from 9pm to 3am the thermostat is set back to a sAT of 15 degrees. What happens next depends on two things: the OAT and the thermal mass of the building. IF the OAT is also 15 degrees, nothing will happen. However, if it is less than the sAT, then the house will gradually cool down, the cooling rate depending on the OAT and the thermal mass of the building. In the case of my building, during the cold spell last December, during a trial setback, my house IAT fell by about 2 degrees over about 6 hours. Other properties will vary (this is one of the headaches of all this, and why at the end of the day, observational data trumps models data). Howver, for my building, even with a cold OAT, the effect of the setback is to turn of the heat pump for the duration of the setback. I therefore use less energy per 24 hours by the amount saved by not running the heat pump during the 6 hour setback. Sticking with the earlier numbers, and an OAT of -2, I might save 6 x 3.33kWh = 19.98kWh of energy use. This is not a trivial saving, but...

(j) ...cost savings are now complicated by time of use tariffs. I am not even going to try to incorporate the effect of the TOU tariffs here, primarily because i don't have a TOU tariff, but those who do will need to factor them in if the want to know about and ultimately control running costs). Here' I am just going to stick with I have saved 20kWh of energy consumption (or less when the OAT is higher). That much is pretty straightforward, at least in my eyes.

(k) we then come up against the heat pump's Achilles' heel: it is a steady Eddie, and left to its own devices it will takes ages, hours if not days, to recover the IAT, from whatever it has dropped to, to the desired IAT. In my case, it pretty much takes all day, ie it never really recovers before the next set back kicks in. The only fix for this is to boost the LWT to achieve a more rapid recovery period, and because higher LWT means more energy use per se, plus a bit more because COP falls as LWT rises, then i will use more energy per hour during the recovery period than had I been running in a steady state (continuous 24/7 running), when the LWT would not have been raised.

(l) and so we arrive at the 64 million dollar question: how much extra energy do I use per hour, and for how many hours, and then, is the sum of that extra energy used more or less than the energy saved during the setback period?

(m) in principle, I should be able to calculate this. If I know, from the manufacture's data sheets, how much extra energy is needed to get that higher LWT, I can then work back through the COP to how much extra energy I need to use to get that higher LWT. But it is painfully complicated. Not only will it depend on the actual higher LWT needed to achieve recovery (which I don't know at this stage), it will also depend on OAT, and then the other 1001 myriad factors that make up my particular heating circumstances, insofar as no two properties are identical. I could set up the mother of all spreadsheets to do this, but...

(n) there is an easier way: run the experiment. At times, run the heat pump continuously, and record the 24 hour energy use, at other times, run it with a setback and a recovery period, and record the 24 hour energy use. As I already record 24 hour energy use automatically, this is very easy to do, because, ahem, I am already doing it. OAT will vary from day to day, but over time I can build up a picture of what happens at various OATs, given either continuous or setback and recovery running. This I suggest will provide the definitive answer that applies to my heating circumstances, and may very well apply to others with similar heating circumstances (building characteristics, preferred IAT, seasonal OATs etc), though not to others with differing heating circumstances, who will have to run their own experiment to get their circumstance specific answer. Those using TOU tariffs will also need to consider the timings of setbacks and recovery boosts, and where they fall on the TOU tariff, because in their case energy use is not a proxy for cost (in GBP). I further  suggest that such answers, derived from natural  experiment and observation, will always be more accurate, and more convincing, than any answers provided by spreadsheets.

Footnote: I can do this, because my Midea heat pump allows me to monitor various parameters, and (I believe, though I have yet to prove this for routine live use in the field) control various parameters automatically. Other heat pumps vary. Samsung have modbus connectivity, but there may be problems with setting as opposed to retrieving parameters; Ecodans may already have a recovery option (auto adaption) but I am not sure how well they record actual energy use; Homely may provide a 3rd party way of achieving auto adaption, but does Homely record energy use accurately? - and so on. Maybe even Home Assistant can be set up to do it.         

@cathoderay

I think that you have covered the topic quite clearly, with the primary determining factors being the temperature difference between indoors and outside, the thermal mass, and the heat pump COP variations.

I am at the moment developing the formula to be able to predict the effect of setback under different operating conditions, which you may then wish to test against your system.

Posted by: @derek-m

↑

I am at the moment developing the formula to be able to predict the effect of setback under different operating conditions, which you may then wish to test against your system.

I think that is an excellent idea, we might even find out which definition of simulate applies! In any event, it will be interesting to see how the simulator output compares to real life data.

I am in limbo at the moment, having started a return for the dud temperature (IAT) sensor, and at the same time ordering a replacement from a different supplier. It is not entirely clear whether all these sensors are prone to be duds, ie bad design, or there are just dud batches, possibly made by rogue knock off manufacturers. Trouble is, they are the only sensor that I have found that can connect directly to my modbus set up.

Had another power cut this morning, only an hour or so, but enough to be a nuisance. They are a fact of life round here. 

Posted by: @cathoderay

↑

whatiffery

As previously mentioned, comfort.

For the whatiffery equation, indoor humidity, 30-40-50-60-70% and how much it affects the feel of being cold/warm vs cranking/reducing the indoor temp. 

Posted by: @fazel

↑

As previously mentioned, comfort.

I agree, that is the outcome variable that really matters (along with total running costs, not so good if we are very comfortable, but are bankrupt from paying the heating bill), but as I don't have a modbus socket in my brain, I have decided to use IAT as a proxy, ie if the house is at the IAT I am comfortable at, then I am likely to be comfortable. Others eg @toodles who prefers a higher IAT can use a higher desired IAT. But I absolutely accept there are other factors in play, notably humidity as you say. The sensor I have in mind also measures humidity, so I might be able to factor that in, though it will add extra complexity. 

Posted by: @cathoderay

↑

Posted by: @fazel

↑

As previously mentioned, comfort.

I agree, that is the outcome variable that really matters (along with total running costs, not so good if we are very comfortable, but are bankrupt from paying the heating bill), but as I don't have a modbus socket in my brain, I have decided to use IAT as a proxy, ie if the house is at the IAT I am comfortable at, then I am likely to be comfortable. Others eg @toodles who prefers a higher IAT can use a higher desired IAT. But I absolutely accept there are other factors in play, notably humidity as you say. The sensor I have in mind also measures humidity, so I might be able to factor that in, though it will add extra complexity. 

What's a 'bit' of complexity between friends?

Posted by: @cathoderay

↑

a modbus socket in my brain

Wasn't Elon working on that?

It's often the case that people crank up the thermostat as it gets colder outside.  I think the most significant reason for this is that the external house walls are getting colder, and this gives a radiative imbalance for the person indoors - which is totally uncaptured by only considering the internal air temperature as done by a thermostat or similar.  In fact, the wall temperature has almost as much impact on comfort as the air temperature - I made a simplification in excel assuming the case of a naked person of 1m^2 area at 36degC skin temperature convecting + radiating power in Watts.  Most people wear clothes, and are comfortable at around 100W.

Posted by: @fazel

↑

Wasn't Elon working on that?

He was, but he put it in the wrong place. Now, every time I touch a light bulb, it lights up for some reason. 

Posted by: @robl

↑

Most people wear clothes, and are comfortable at around 100W.

Speak for yourself! Seriously. though, you make a good point about another limitation of considering just IAT. I think that sort of perceived warmth is also the selling point behind far infrared heaters, which, like the sun, warm you, rather than the air around you. The problem is how to capture 'comfort'. When I have a chance, I'll have a look at the literature, and see if anyone has come up with anything. I suppose one way to do it might be to have five column grid and a pencil, and mark off on a scale of 1 to 5 how comfortable one felt every hour. I rather suspect if I told my friends I was doing this, they might suggest I need to get out more often.    

@cathoderay

Do you go out among the 'sick' people? 😋