Energy consumption

Over the past couple of weeks I’ve been trying to measure the heat input to the house from the heating system, and gain a better understanding of the way the house responds to changes in temperature.  I already knew that the air source heat pump (ASHP) only seemed to come on for a couple of hours or so each morning, and also knew (from power measurements on the supply to the ASHP) that it rarely seemed to use more than around 800 W to 900 W, settling down after about 10 minutes or so of running to around 400 W to 500 W.   What I wanted to find out was how much heat was actually being supplied to the house from the underfloor heating (UFH).

To do this, I used the data that the built-in house measurement and data collection system uses, together with a portable data logger than I adapted to measure the surface temperature of the heated floor.  After a couple of weeks there is a fair bit of data, as the house data logger records the outside air temperature, inside air temperature, inside relative humidity, inside CO2, floor slab internal temperature, buffer tank temperature, ASHP flow temperature and the hot water preheat plate heat exchanger temperature.  The portable data logger records the temperature of a remote temperature probe, relative humidity and CO2 concentration normally, but I adapted the temperature probe so that I could stick it to the heated floor surface, near the centre of the house, to measure floor surface temperature.

There’s a formula for calculating the amount of heat power, in Watts/ square metre (W/m²) that an UFH heating system is delivering:

P (W/m²) = 8.92 x (floor surface temperature – room temperature) ^1.1

The two data loggers both take and store ten measurements per hour, and so I ended up with over 4000 sets of data to process after a couple of weeks or so.  The data is challenging to present in a form that’s readable in this blog, but some things are quite clear.  The inside temperature is pretty stable, fluctuating by a bit over 1°C in total, around the room thermostat set point of 20.5°C, no matter what the outside temperature.

This plot shows the hourly changes in outside, inside and floor surface temperature over just a part of the measurement period, where the outside temperature varied the most, to give an idea of the temperature stability:

Unfortunately, we didn’t get any really cold weather, but even so, with a change of outside temperature of nearly 10°C the house air temperature barely changes, never getting lower than 20°C.  There is a little bit of room temperature over-shoot when the heating comes on, as the room thermostat turns the heating off at 20.6°C, but the heat stored in the floor slab continues to heat the house up by nearly 1°C above the turn-off point, without the UFH running.

I may be able to fine tune this, as it is directly related to the UFH flow temperature.  I have this set for around 24°C to 25°C at the moment, as if it is any higher the room temperature over-shoot increases.  The problem is finding a thermostatic mixing valve for the UFH that will reliably hold a relatively low flow temperature; most are designed for heating systems in houses with a far greater heating requirement, so don’t regulate well below around 30°C.

From all this data, I was able to calculate the arithmetic mean daily heat input power, as well as the actual daily heat energy input, together with the arithmetic mean daily temperature:

As we’re not yet living in the house, this is very much a worst case heating requirement, as each person in the house will add around 80 W to 100 W of heat, plus there will be additional incidental heat gains from cooking, running showers, appliances etc.  I suspect that the actual heating requirement would be non-existent for many of the days in the chart above, once we’re living in the house.

I also tried to estimate what the heating cost would be per day, using an assumed coefficient of performance (COP) for the ASHP of 3.  This is probably slightly pessimistic, as for much of the time the ASHP COP is likely to be well over 3, but nevertheless it gives an idea of heating cost.  It looks as if the average daily heating cost, during the winter, is unlikely to be higher than around 25p, and may well be a fair bit less than this once we’ve moved in.  Even in the winter months we usually generate more electricity than this from the photovoltaic (PV) panels, so, as hoped, it looks like we won’t have any energy bills at all.

4 thoughts on “Energy consumption”

  1. Hi Jeremy.

    Another great post. I wonder how solar gain plays a part – perhaps directly onto the slab in places, and what the time constant is for the house – IE how quickly the outside air temperature change takes to migrate across the insulation? There also seems to me to be a curious anomaly in the graph and I wondered if you had any thoughts on it – perhaps I have missed something. At about 3/4 of the way across the outside air temperature is generally rising but the slab is cooling while in all other parts the curve the reverse is generally happening – the slab temperature tracks down with the outside air temperature.

    1. As far as I can tell, solar gain dominates, and is significantly more powerful than the heating system in terms of increasing room temperature quickly. The reason, I think, is that the sun heats up surfaces, including the inner panes of any triple glazing that doesn’t have the external reflective film on (mainly the East and South facing kitchen windows), pretty quickly. It doesn’t seem to affect the slab temperature, but that’s only because I’ve taken measures to stop the sun shining on the floor. It does, for example, heat up the stone internal window cills in the kitchen, and these then radiate and convect heat into the room. The room air heating effect from this is rapid, because of the high differential temperature. It’s not unusual for the window cills, for example, to be at 30 deg C or more, even in winter sun, when it’s cold outside.

      The insulation has a decrement delay time that is sufficiently long that the effect of external temperature changes on internal wall and ceilings is negligible. When I’ve measured the internal surface temperatures with an IR thermometer, or the Seek Thermal camera, they always seem to be within a tenth of a degree or so of the room temperature.

      The reason for the apparent anomaly, where the slab is cooling but the outside air temperature is rising, is just because the house is very slow to respond to outside air temperature changes alone, the thermal time constant seems to be longer than a day. This means that the outside air temperature can be increasing, after the UFH has turned off from it’s short burst each morning, yet the slab will still be giving heat up to the room air, rather like a storage heater. As the slab cools, and the room temperature rises, the rate of heat transfer reduces.

      The room thermostat is set to 20.5 deg C, with a 0.1 deg C hysteresis, so it turns the UFH on at 20.4 deg C and off at 20.6 deg C. There is a bit of over-shoot, because the UFH runs at around 24 deg C, so even after it has turned off there is still heat in the UFH pipes that is slowly released into the slab over the next few hours. This over-shoot is very dependent on flow temperature, I’ve found. If I turn the thermostatic mixer up to around 28 deg C, then the fluctuation in room temperature gets a lot greater, with the peak being around 23 to 24 deg C, which I find a bit too warm.

      I still need to fine-tune the programming, as it seems that it might make more sense to have the UFH come on later in the morning, to time-shift the room temperature peak to later in the day. In practice the current level of fluctuation isn’t detectable when in the house, it feels to be at the same temperature all the time. The room temperature sensor is on the ground floor, and generally the first floor bedrooms seem to be about 1 deg C cooler, most probably because, with the bedroom doors shut, they are being fed with air from the MVHR (which is always a bit cooler than room temperature) and the heat transfer through the floors is limited, because there’s 200mm of rockwool acoustic insulation in the ceiling/floor void.

  2. Hi Jeremy.

    Re-read this today as I was working on arranging for sensors in various places and spotted something I did not see before which I find curious. On the 19th and 25th the power consumption has risen but so has the outside air temperature. I think I would have expected it to drop not rise. Wonder, has the previous colder day drawn more out of the fabric than the temperature indicates. Any thoughts?

  3. Hi Mike,

    Yes, I’m sure the effect is down to the long thermal time constant, and is as you suggest, I think. The internal air heating is a mix of the slab, plus the stored heat in the structure, giving up heat to the slightly cooler ventilation air that’s being fed in. Because the heating is turned off overnight, on the timer, the impact of a cool evening and night is really seen in the following day’s heating energy consumption, as the slab and structure are “recharged”.

    The house seems to behave a bit like a storage heater, which would show the same sort of pattern of energy use, with a cold day requiring more charge during the following off-peak period. Terry is hoping to use this heat storage effect as his heating buffer, I believe. I’ll be interested to see how well it works for him.

Leave a Reply

Your email address will not be published. Required fields are marked *