Jump to content

Modelling the "Chunk" Heating of a Passive Slab


TerryE

Recommended Posts

Real world data is what it is all about in the end.

I missed the opportunity to pick up some old storage heater bricks last week. They would have made a good testbed. I may have to pull mine apart and rig something up (just thought of that).

I am waiting for some more ds18s to turn up at the moment. Chinese new year has got in the way.

Link to comment
Share on other sites

I have about 10 day's worth of data now, but we haven't had any cold weather, if anything it's been pretty mild, so I think it's best if I keep the loggers going for at least another week or so, just to try and get as wide a range of outside temperature variations as I can.

 

Looking at the temperature displays on the loggers, the floor surface rarely seems to be more than 0.5 deg C  above the room temperature, and the room temperature seems to be fairly steady, between about 20.5 and 21 deg C.  I've not noticed the heat pump running for the past week, but it's possible that it may have come on earlier in the morning, before I get over here, for a short time.  I'll know for sure in a week or so, as the log will have the heat pump flow temperature, which will show when it's run and for how long.

 

If we get a spell of cold weather in the nest few days then I'll stop the loggers in a week or so and post the data here.

  • Like 1
Link to comment
Share on other sites

  • 2 weeks later...
On 11/02/2017 at 13:28, SteamyTea said:

Real world data is what it is all about in the end.

I missed the opportunity to pick up some old storage heater bricks last week. They would have made a good testbed. I may have to pull mine apart and rig something up (just thought of that).

I am waiting for some more ds18s to turn up at the moment. Chinese new year has got in the way.

 

 

I switched the loggers off yesterday, only really because the floor sensor was getting to be a bit of a nuisance (it was stuck to the floor surface, held down by a small bit of foam and sticky tape, and I kept nearly tripping over the thing).  I've concatenated the two data sets, and both clocks were still within a few seconds of each other when I turned them off, so it's reasonable to assume that the 6 minute sample times are synchronised to within a few seconds of each other.  The full data set is attached as a .txt file, but in reality it's a .xls Excel file, so just change the extension after downloading it.

 

What's clear is that I wasn't noticing the ASHP fire up, as it seems to come on every morning for a couple of hours, and I thought it only came on every one or two days.  What's also clear are my visits over to the new house in the afternoons; you can see the CO2 increase when I'm in the house quite clearly.  The relative humidity is on the low side, but that may well be just because we're not living there yet; I'm splitting my time between working on our old house and clearing out stuff and going over to make sure the new house is OK.

 

The ASHP flow temperature sensor is located very close to where the flow pipe enters the house, so is a pretty reliable indicator of when the ASHP is running.  It's set to deliver 40 deg C, and when it turns off the temperature can take a fair time to get back close to room temperature, but the sharp drop when it turns off is reasonably clear when the data are plotted out.  The slab temperature sensor is embedded on the North side of the slab, around 200 to 300mm away from the nearest UFH pipe and with the slab area from it's position to the edge of the slab unheated, as it's underneath units.  As a consequence this sensor under-reads the core slab temperature, I really should have fitted it nearer the centre of the slab, but it's too late to change that now.

 

I've been doing some playing with the data, as I think I can calculate the actual heat delivery into the house from the slab surface temperature to room temperature differential, using the standard UFH formula of P (W/m²) = 8.92 * (floor surface temp - room temp)^1.1 

 

Logged Data Feb - Mar 2017.txt

Edited by JSHarris
Link to comment
Share on other sites

Jeremy, I did a quick plot of this, for the benefit of others.  The temps are in °C, the RH in %, and the CO2 PPM are on the right hand scale.

 

Jeremy-data.png.db16a4f1ef7a006e2a53126e8ef5701d.pnge PPM is on the RH scale.

 

 

  • Like 1
Link to comment
Share on other sites

Thanks Terry, I've been crunching the numbers a bit to see what the heat power and energy input is, and have made this little table:

 

58bc002250728_Heatinput.jpg.0b9ee6b60f640c5bdd0469c7cad0e7cf.jpg

 

The cost column on the right isn't exact, because I've assumed a heat pump COP of 3, when in reality it may well be over 3 some of the time, or even under 3 occasionally, plus the cost assumes 100% import, which isn't that likely; the chances are that at least some of the heat pump power is coming from the PV array, even on dull days.  Even so, it looks like the house will only cost around £1 or so per week to maintain at a fairly steady 20 to 21 deg C, which is pretty good.  The room thermostat was set to 20.5 deg C, with a 0.1 deg C hysteresis.  The house is empty most of the time, with just me there for a few hours each weekday, and no cooking, showers etc to contribute to heating the house, so this is close to a worst case, in terms of heat requirement.

 

For the benefit of any statistics whizz reading this, the daily arithmetic means were calculated by summing the 6 minute sample values and dividing by the number of samples per day.  It's interesting to see the way that the relatively cold 24th February didn't really impact the heat input until the 25th February, due to the long thermal time constant.  For those reading this and thinking about "thermal mass", it's worth noting that this house is a lightweight construction, entirely made from timber, with the exception of the 100mm thick concrete slab and the ~15mm plasterboard and plaster skin inside; even the insulation is made from wood, in the form of blown-in cellulose.

Edited by JSHarris
Error in data descriptions - now corrected
Link to comment
Share on other sites

Right, have had a little play this morning and produced two charts.  One is the correlation between the invariant of external air temperature and the other against time.

I picked these two because we have no control over them and can therefore be used to both correlate and draw implications.

There is really not enough data to draw very precise data for the long term, but as the floor surface, slab internal and room air temperature are amazingly stable (basically within the accuracy of the sensors) the 11°C external temperature swings that have recently happened are a good indicator of overall performance.

As Jeremy a already said, the ASHP is coming on at around 3AM for about 3 hours.  This fits in with the coldest part of the day, though may not be the best time to start heating.  That depends on lifestyle and how you like your house heated (It would suit me as I like a warm evening).

The slab, on average, has the highest temperature at about 2PM until 1AM.

Room air temperature peaks at about 4PM.

On the Correlations, everything is pretty flat except the Mean ASHP Flow Temperatures.  This is to be expected because if the outside air is warmer, the ASHP will be working less often, rather than delivering a lower flow temperature.  This will force the mean flow temperature downwards.

 

I shall try and have a better look later, but life got in my way this morning and only just managed to get home again to pick this up.

Jeremy Hour of Day Temperatures.jpg

Jeremy Correlations.jpg

  • Like 2
Link to comment
Share on other sites

One advantage of this mini test is that you can fairly safely assume that slab temperature sensor is good enough to get a proxy for the slab surface temperature.

Not sure of it is good enough for the air temperature as we have not had any high air temperatures yet.

 

Am I right in thinking that you circulate the water though the slab constantly and then just inject from the ASHP/Buffer when heating is needed?

If that is the case, it may be worth looking at the thermal losses from the Buffer and the PHE as they have very long tails and significantly higher temperatures than the slab (7 - 10°C).

I shall try and have a look at what is happening at the daily level as I can just use the limited datasets of Slab Internal, Room Air, Outside Air and ASHP Flow temperatures to make it easier to look at.

Though in reality, one day off either a bit hotter or a bit colder does not make any real difference to the overall energy usage.

Edited by SteamyTea
Link to comment
Share on other sites

The UFH circulation pump turns off early evening and on again in the early hours, using a time switch.  The buffer and PHE are pretty well insulated, but there will be thermosyphon pipe losses, as the PHE is upstairs and the buffer downstairs, so it's likely that the pipe losses are the main heat loss path, even though the pipes are as well insulated as I could make them.

Link to comment
Share on other sites

So once you are living in the place full time, you could tinker with the pump timings and maybe shift the maximum temperature times to suit your lifestyle and the seasons.

It may also, once more data is gathered, allow for some simple weather compensation using a Met Office forecast for temperature and cloud cover (I think your house is more affected by sunlight than temperature).

Might be fun creating an infrared thermometer that points up into the sky to gauge the cloud cover, though wind direction is the best indicator that I can find for the UK.

Link to comment
Share on other sites

You're right, sunshine has the biggest impact on internal temperature, certainly seems to be a greater impact that just the outside air temperature.  I'm sure we'll need to tweak the time switch settings a bit, as they are just set to what I thought would be about right ages ago, when I didn't really know how the house responded to heat inputs at all well.

 

I tried weather compensation ages ago, with the slab temperature control system, but because it was just reactive it didn't work well.  If it had been able to use the predicted temperature/sunlight for the following day it may well have worked OK, but the long delay between heat input and maximum heat output, together with the relative insensitivity of the slab sensor (it needed to be better than the ~ 0.1 deg C resolution of the slab sensor, ideally)  meant that I couldn't come up with a compensation method that worked as well as the current system.  The simple room thermostat seems to do a pretty reasonable job of controlling the internal temperature, and is certainly a better long-term solution, as it's an off-the-shelf part.

Link to comment
Share on other sites

Right, a few more pictures.  I notice that the time between maximum ASHP Flow temperature and there it crosses the slab temp plot on the cooling cycle is pretty consistent at about 8 hours.

Not sure what these plots really show, other than you could shift the time that the ASHP is working to take advantage of the PV without affecting the house temperature unduly (less than a °C I would think).  This would take away for car charging time, which is probably a greater cash saving.

The charts, are for the hottest day and then including the two days either side, then the same for the coldest time.

 

Hottest Day Temperature.jpg

Hottest Day Hour.jpg

Hottest 3 Days Temperature.jpg

Hottest 3 Days Hour.jpg

Coldest Day Temperature.jpg

Coldest Day Hour.jpg

Coldest 3 Days Temperature.jpg

Coldest 3 Days Hour.jpg

  • Like 1
Link to comment
Share on other sites

Thanks again for this, ST.

 

Car charging can easily wait until the afternoon, so shifting the ASHP programming to get it to turn on later wouldn't be a problem at all.  Also, the ASHP only draws a few hundred watts, so doesn't need as much PV output as the car charger and could make better use of the relatively low early morning PV power.  I'll re-programme the ASHP times later and see what practical impact it has on the house temperature

Link to comment
Share on other sites

This is a good time of year to try things out as the sun is getting more powerful, and air temperatures are becoming reasonable.

It looks like anything below 10°C mean external temperatures you need some heat input.  This is probably caused by the MVHR hot being able to recover, at a useful temperature, as much energy.

Link to comment
Share on other sites

You're right about the MVHR, I haven't logged data from it, but the control unit displays the exhaust air temperatures before and after the heat exchanger, and the inlet air temperatures before and after it as well.  Whenever I've looked at it in winter the exhaust air before the heat exchanger is typically around 21 deg C or so and the fresh air feed to the rooms is usually around 18 deg C or so.  So the MVHR is always working to cool the house a bit in winter, but very much less than normal ventilation would.

Link to comment
Share on other sites

20 hours ago, JSHarris said:

You're right, sunshine has the biggest impact on internal temperature, certainly seems to be a greater impact that just the outside air temperature.

 

There are a few issues here.  As you both say, sunshine through the windows is a major source of heat input, and it isn't being measured directly.  Moreover whereas there is a lag of >24hrs with the external temperature, the input of incident radiant heat has no delay.

 

Picking meaning out of correlations especially in the 0.3 or less range is extremely difficult.  The underlying causal linkage isn't at all clear.

 

Tuning your system further is going to be difficult without some home automation system, even if it is just a chunk of Fortran that runs once a minute on RPi3B|  However, one obvious strategy to consider is to shift your ASHP cycle to around midday (or at least have a boost midday) when your expected PV output is maximum and therefore you have the highest expects PV input.

 

Another thing that I note is that having your services split over two floors in itself introduces losses.  I was originally going to have some in the "services" room in the loft and some adjacent to the GFL toilet.  However I was luck enough to be able to replan mine after reading your blogs.  Now we only have the MVHR in the loft.  All of our heating and water services is in a 0.65 × 1.4 × 2.4m services area off the GFL toilet.  (Though it has been a bit of a challenge arranging every thing to keep pipe runs clean and accessible, but that's another topic.)

 

My last thought is there's nothing like collecting hard data to understand how your system operates and how to optimise it.

Link to comment
Share on other sites

  • 8 months later...

Just out of interest for those that want to compare "real-world" results to the model shown on the previous page, here is a plot take from our commissioning runs where I had our Willis heater running for 6hr at Economy 7 low tariff (cost ~£1.40).  The house is still warming, hence the slab is still dumping heat into room space.   But it's already got to the point where I am stripping off jumpers when working.  The overall conclusions are:

  • The slab is reacting pretty much as I modelled it.
  • The output of the Willis is easily enough to heat our house.
  • Our actual flow rates through the slab on the pump's middle setting are about half what I estimated, hence the out:return delta temp is nearer 5° rather than 2° but this doesn't matter.
  • At this cost even if I use E7 electricity to heat my house the annual running costs will still be less than out BGas maintenance contract on our current boiler.
  • The Willis is simple cheap and effective.  Even if you have an ASHP or alternative, then it;'s still worth considering fitting one as a fallback heater

5a03132eb44d4_SlabPlot.png.fb98e61bab97d55848a7752e2770fc74.png

 

Trust the physics, but validate your assumptions.

  • Like 1
Link to comment
Share on other sites

Interesting stuff @TerryE, thanks for sharing.

 

Do you have handy a schematic and/or photos showing how and where you've installed the Willis heater? We're about to have some work done on our heating circuit and it would be a good opportunity to plumb in at least a connection point for a Willis heater for backup.

 

Many thanks.

Link to comment
Share on other sites

Guest Alphonsox

I've pretty much come to the conclusion that heating via E7 is the way to go for us. I don't see the ASHPs that are currently in the garage ever being installed. We are using a 90ltr UVC heated by a 3Kw immersion as the feed for the UFH and a fan assisted radiator. The system runs for the E7 period heating the slab. The hot water in the UVC gives an instant heat boost via the radiator if required during the day - all for a quid or so per day.

Link to comment
Share on other sites

Funny you should post this, Terry, as several times over the past few weeks I've been thinking of the simplicity and low capital cost of using a Willis Heater.  I've even been tempted to dismantle the base at the front of the airing cupboard to see if there is room in there to fit one, primarily as a back up at this stage, but with thoughts that I could could divert excess PV to it, once the Sunamp PV is charged.  Days like today, when it was pretty clear and cold, would be good days to just stick some extra heat into the buffer tank, even though the house was sitting at around 21.5 deg C and the heating hasn't been on for a few days now.  I did expect the heating to have been on this morning, as it was sub-zero last night, but the house still seems to have plenty of stored heat in reserve, and with the early morning sunshine there was enough solar gain to warm the house.

Link to comment
Share on other sites

I've included the upper and lower levels of my system.  This is the lot -- CW + Filter + DHW + CH all in a cupboard some 1.2m × 0.7m off the side of my ground-floor loo.  The hot-works are all lagged.  The system is stripped down to its bare essentials, and only electromechanical bit is the UFH pump.  If running your CH on this costs maybe ~ £200 a years and my current BGas boiler maintenance contract is over £300, then its hard to make any payback case for anything more complicated.

 

DHW_CH-upper.thumb.jpg.db70e13bf8e4f6d79d6b3a5400f3dd7c.jpgDHW_CH-lower.thumb.jpg.fc8666d7c5d4d500953d39c341ea3cfb.jpg

 

At the moment the heat exchanger preheats the DHW mix (which is the reason for the notch in my output temp above.   If I do decide to use an ASHP, then I'll pinch this HE and prime it from the ASHP.

Link to comment
Share on other sites

2 hours ago, Alphonsox said:

I've pretty much come to the conclusion that heating via E7 is the way to go for us. I don't see the ASHPs that are currently in the garage ever being installed. We are using a 90ltr UVC heated by a 3Kw immersion as the feed for the UFH and a fan assisted radiator. The system runs for the E7 period heating the slab. The hot water in the UVC gives an instant heat boost via the radiator if required during the day - all for a quid or so per day.

 

So if you don't use'em what'll you do with the ASHPs?

Link to comment
Share on other sites

Guest Alphonsox
2 minutes ago, dpmiller said:

 

So if you don't use'em what'll you do with the ASHPs?

 

Garden ornaments :)

 

....If however I decide not to use them they will be put up for sale (here first) at the price I paid for them. That decision will be a few years away, until i've been in the place for a winter or two.

Link to comment
Share on other sites

I usually think about space heating for my place this time of year.

I am rather limited in what I can do, and hate spending money.

As I have a particle board floor over polystyrene, I could router some semi-circular groves in it and fit UFH pipe and cover back over with the parquet flooring.  Then connect that to a cheap pump and hot water cylinder.

As the floor area that needs heating is probably about 20 m2 and only needs about 1 kW delivered to it to heat the whole house, I would only need to store 24 kWh.

I could store that in 2, 200 lt E7 cylinders (I have room under the stairs) and run them up to 85°C.

 

Not sure how realistic that is, but probably a fairly cheap to find out and I would be left with UFH pipework if I wanted to connect to a heat pump.

 

I may have to model it.

 

 

Edited by SteamyTea
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...