Earth Notes: 16WW Impulse Heating Test

I had thought that the day would be quiet and the test would be quick. The various minor crises that broke out as I was starting the run were a trifle distracting! And the room did not achieve equilibrium temperature even after 7 hours. Steady state would have allowed me to check heat loss and other parameters.

Also, I could not actually work in my study area with this test recreating the horribly hot conditions of summer 2019!

I burnt 13kWh of gas and created 2.5kg of CO2 emissions for this experiment. The heating has been off (other than the two exceptions described here) since before the end of March, saving rather more than these japes cost.

Observe that the maximum heating power that I was able to draw with the radiator was a little over 2kW. There are only six radiators in the house, though two of them are larger. Maybe a total maximum draw of ~(4x2kW+2*4kW) or ~16kW. The boiler is rated at 24kWh. Almost all boilers in the UK are massively oversized for space heating, and cannot modulate down far enough to work well at typical loads. That custom will not play nicely with heat pumps. (Note that this is gas consumption; nominal boiler efficiency is ~80%, so heat delivery is at 20% lower and there will be other system losses such as from pipework en route to the room.)

Aims

Things measured and calculable from this test include:

• Approximate boiler flow temperature.
• Maximum power that can be supplied by this radiator.
• In principle, heat loss for given δT to outside and rest of house, at equilibrium temperature.
• In principle, thermal capacitance (sometimes called "thermal mass") of room.
• The patience of the rest of the household...

Getting Ready

Main points of preparation for this run:

• Room curtains drawn to minimise solar gain.
• 4o outside temperature measurement in porch reinstated.
• Weather conditions: sunny (so some solar gain), cool.
• Second temperature measurement stream near geometric centre of room:
  • Eye level on corner of chest of drawers.
  • Using 1g for this temporarily. Have replaced 1g with 'off' Radbot.
• All Radbots turned down to * (frost).
• 5s Radbot removed from radiator and placed nearby to continue to measure temps.
• Heat-sensitive gadgets removed from room.
• Validate 4o, 5s and room-centre temperature data stream.
• Old house thermostat turned up to max (>30°C) to force-enable heating.

I was intending to collect the room-centre temperature readings another way. But that was not working so I borrowed 1g's Radbot for the duration. I put a 'shut' Radbot on that radiator to avoid drawing heat there.

Experimental Run

I was fighting work fires but nothing melted in 5s, thankfully.

• Gas reading before (m^3): XXX5.622
• Start time: 08:52Z
• Switch boiler to CH+DHW mode.
• Windows and doors of 5s to room closed.
• Measure radiator flow temperature at boiler as well as possible: ~47°C w/N84FR on boiler pipe.
• Ensure that gas not used for other purposes eg DHW.
• Gas reading intermediate (m^3): XXX6.127 @ 2021-04-29T11:27Z 5.6kWh
• Gas reading intermediate (m^3): XXX6.560 @ 2021-04-29T14:22Z 10.4kWh
• Gas reading intermediate (m^3): XXX6.664 @ 2021-04-29T15:18Z 11.6kWh
• Gas reading intermediate (m^3): XXX6.791 @ 2021-04-29T16:05Z 13.0kWh
• Temp reading from stand-alone digital thermometer at eye-height ~28°C.
• Continue until 5s room temperature stabilised if possible! (Did not happen.)
• Record end time: 16:05Z (433m, 7.2h)
• Gas reading end (m^3): XXX6.791 @ 2021-04-29T16:05Z 13.0kWh ~1.8kW mean.

I realised that at a bit after 4pm (15:00Z) that my usual winter-time peak-time avoidance of boiler off from 4pm to 7pm had kicked in. So the room stopped being heated. I manually overrode that when I realised, but the dip in temperatures is visible.

After

• Plot and publish implied thermal characteristics of 5s.
• Making good:
  • House thermostat back to 10°C (minimum).
  • Boiler back to DHW only, and on clock not always on.
  • Re-opened lockshield on 3l.
  • Put 1g back.
• Have left all Radbots on * (frost), or 1 for 1g because of dial issue.
• Charts and data captured.
• Windows opened wide to cool off!

Data

(This may get upgraded to a dataset...)

• Data folder.
• Experimental notes lightly edited.
• Temperature and other live data (JSON OpenTRV format, GZIPped) for relevant period from three Radbots, see IDs below. 1g was temporarily borrowed as the room centre(ish) sensor.
• Combi boiler is a ~80%-efficient [archive] "Potterton Performa 24" rated at 24kW, SEDBUK band D.
• Boiler flow temperature measured with Maplin N84FR probe on flow pipe to be ~47°C.
• 5s room dimensions (5s is in front bedroom).
• Temperature outside house ~9°C to ~11°C during test as sensed by 4o.
• Temperature of interior of remainder of house ~17.5°C.
• Radiator similar to Kudox Premium Type 22 Double-Panel Double Convector Radiator 500 x 500mm White 2571BTU (85836) rated at 753W output with δT to room 50K: note mean δt in this test ~25K thus mean heat output probably ~375W. (Implies <3kWh heat into room over 7.2h test, and ~20% overall efficiency from gas in to heat out, not the SEDBUK 80%!)

Notes

• May be minor DHW use from hand-washing. (May have been none in fact.)
• 3l radiator came on so had to turn off at lockshield.
• Error: heating went off a little before 4pm (normally off 4pm to 7pm) noticed at ~4:20pm (15:20Z) and put onto clock to always-on mode.

Thermal Capacitance

Looking at the shape of the graph suggests that the effective thermal capacitance of the room surfaces and contents is ~0.1kWh/K.

That nominally puts a limit on the speed of response of temperature to the radiator going on, very crudely 4K/h (4°C/h). This is in line with my expectations I think.

I may try to refine this crude estimate, eg with another testing run, and/or by estimating the capactance of key elements such as the plasterboard inboard of the aerogel IWI and the mass of materials in the room.

Conclusions

Maximum power consumption from gas mains ~2.2kW for 5s radiator. Given likely 80% efficiency of boiler, heat delivery maybe maximum ~1.8kW.

In fact, given the radiator characteristics and δT (flow-to-room), mean heat power to the room was likely a little under 400W, for ~2.7kWh total. That implies a ~20% overall efficiency from gas in to heat out, nowhere near even the SEDBUK ~80%!

My 24kW boiler is massively oversized for even whole-house peak demand, and so running modulated to minimum and then cycling is bound to be less good. This demonstates that correct system sizing is important!

2021-04-12: Other Impulse

Earlier in the month, after several days on my own and the rest of the family away... Mean outside temperatures were in the single digits and there was some snow! Mean internal temperatures got down to about 13.5°C. All the Radbots (except the hallway, on frost) were calling for heat. The boiler in DHW-only mode was roundly ignoring them, as intended.

To avoid (justified) complaints from returning family, I put the heating on briefly (ie switched the boiler from DHW only to DHW+CH) at 12:32Z with gas at XXX6.731m^3.

By 12:43Z the boiler flame was out for the first time (ie the heating was limited by power delivery via pipework and radiator, and circuit thermal capacitance.

At 12:57Z gas was XXX7.171m^3, ie ~5kWh consumed.

At 13:54Z gas was XXX7.442m^3, ie ~7.7kWh consumed, and I put the boiler back to DHW only. The estimated mean internal temperature was now ~17°C.

No complaints from returnees a little later, in fact at least one window was opened!

Historical Note

In both cases the rest of the family went to visit my partner's father, who is on his own, and whom we are in a coronavirus "bubble" with.