Back to Silas S. Brown's home page

Heating a flat: vegetable-oil candles or oil-filled radiators?

Jump to: Candles | Electrical heaters


Some people advocate using cheap paraffin "tealights" for heating electrical-only apartments when the cost of electricity is high. Besides the fact that this doesn't always `add up' (even if the tealights are good enough to give you 200 watt-hours each, you'll have to find a shop that charges less than your unit price of electricity per 5 tealights, but for lesser-quality ones call it 10---and if reaching that shop costs then you'd better factor this in), my main worry would be for safety---the American EPA noted significant pollution from paraffin candles (health risks + soot), not to mention the fire risks.

Slightly more promising is the use of "floating water candles" made by suspending a wick in a simple plastic float on a body of water topped by cooking oil. This pollutes less (but see below), and it might `add up', but the cost of buying (or making) the wicks might be a bigger factor than you realised!

Electrical heaters

All electrical heaters convert 100% of their electricity into heat (one way or another), but how much of a difference does it make when and where that heat goes?

[This section has graphs but they are explained in the text] In , an oil-filled radiator uses heated oil (top line) to dampen the oscillation from the thermostat's hysteresis, resulting in a more stable room temperature (bottom line). To achieve the same minimum temperature with a cheap fan heater, the room temperature must oscillate more (middle line), which (with these numbers) increases consumption by 20%. The fan heater does bring the room to temperature more quickly (left); it also cools down more quickly after switch-off (right).

If the room's insulation is improved so much that its total thermal conductance is halved, both heaters' consumption is roughly halved (the fan heater comes on for shorter periods and the radiator needs its internal temperature set lower), and interestingly the difference between them is also reduced (in this case the fan heater consumes only 10% more than the radiator, instead of 20%), although the radiator's consumption can be reduced a few more percentage points by using an element of lower wattage (while keeping the same size and internal temperature setting). This means (1) the benefits of dampening are more pronounced if you're not allowed decent insulation; (2) you should use the lowest wattage setting that reaches the internal temperature you need.

It could take some experimenting to find, for a given outside temperature, which internal temperature corresponds to your preferred room temperature. The more expensive oil-filled radiators tend to have embedded software to help with this, and let you set room temperature in Celsius rather than providing an unmarked dial. Some models also advertise "patented" chimneys etc (I haven't yet found the patent numbers)---the effect of these is to increase heat transfer, so the fixed maximum internal temperature can sustain a higher room temperature for a given size of radiator, but I can't see how they'd improve the overall energy consumption associated with your desired room temperature (although users might appreciate the more-obvious column of warm air when feeling above it). The main energy-consumption advantage of these units is the Celsius user interface.

Going back to the badly-insulated room, when we very roughly simulate the propagation of heat from the area near the heater (top lines in each pair) to cooler areas further away (bottom lines in each pair), it can be shown that if occupants stay around the heater they can adjust it to reduce consumption by 15-20% in both cases; the coolness in other parts of the room is more sustained with the oil radiator than it is with the fan heater.

Specialist heaters like the "Dyson Hot" range claim to improve efficiency by clever air distribution; if good, that means it's "as if" you're staying near the heater even when you're not, so now we have a rough idea of what savings might be possible: 15-20%. I'd expect less in practice, and if an expensive model has a pay-off period of many years then it might break down first.

Low-wattage "always on" elements can consume less than oil-filled radiators on similar comfort settings, but the wattage must be matched to the desired temperature rise, as there is no thermostat to compensate for over-provisioning. The savings arise because, once the wastage of oscillating the room temperature above your desired minimum is removed, the main difference is in the radiator's residual heat after the final switch-off. In this simulation (top two lines) the radiator consumed 8% more during a 1-hour heating session, but obviously the longer the session the less difference it would make. (The bottom two lines show the effect of under-provisioning the always-on elements. It's hard to get right.)

Conversely, if an oil-filled radiator is large enough, you might be tempted to deliberately set it too high, then switch it off at its peak and rely on its cool-down to keep the room above your desired minimum for a while. In this simulation, an oil-filled being set high then switched off is compared with an oil-filled set lower and run longer, and with a suitably-dimensioned always-on element run even longer, both with less than 5% savings in total consumption. This shows that, for a given total consumption level, you get shorter time above the desired minimum room temperature if you heat extra and rely on the cooldown. It might make sense if you're on Economy 7 or similar and plan to switch off just before the expensive tariff starts, but only to save the first 5 to 10 minutes of on-peak heating. In some circumstances that might be all the heating time you need, but to store hours of heat requires an actual storage heater.

These notes are provided in the hope that they are useful, but they do not constitute medical, financial or safety advice. I certainly hope they are better than following dangerous suggestions on Internet videos though.
All material © Silas S. Brown unless otherwise stated.