If you are maintaining the same temperature in your house, you will always need the same heat energy for that - even the most advanced control system cannot work its way around the physics. If you need less heat than similar sized properties this means you have better insulation and/or air tightness and/or more favourable external climate conditions.
I would go so far to say that any gains in efficiency due to a control system like your thermostat should only be compared within the same house and not between different houses as the uncertainties will completely drown out the “signal”.
You could compare against properties with the same ‘Heat loss at design temperature’ or same annual ‘Assessed space heat demand’, though these are both estimated rather than measured.
Indeed, a well installed heat pump with correct controls will control its heat output very precisely using a combination of weather + load compensation. It’s the poor system design, oversizing and incorrect controls that cause some to under perform.
I still don’t follow what this calculation is. 0.15838£/kWh seems way too expensive. Can you explain it?
Here’s a paper a found [link below], which looks at thermostats with different precision levels.
According to the study:
Precision Impact on Savings:
A thermostat with 0.1°C precision achieves energy savings nearly identical to theoretical adaptive setpoints, enabling optimal use of heating and cooling strategies.
Thermostats with 1°C precision (common in older HVAC systems) result in significantly reduced energy savings. In some cases, energy consumption can even increase compared to static operational patterns due to the inability to finely tune the setpoints.
Energy Consumption Variations:
In heating, the energy consumption increases by 22.44% on average when using a thermostat with 1°C precision compared to 0.1°C precision in the current climate scenario
Quote from the paper
In the current scenario, an average increase in annual
heating energy consumption of 1.02, 10.47 and 22.44% was obtained
with AP-2, AP-3, and AP-4, respectively, and in the annual cooling
energy consumption, the increase percentages were greater: 5.09%
with AP-2, 41.40% with AP-3, and 76.44% with AP-4.
In simpler terms, here’s what this means:
Heating energy use:
If the thermostat is very precise (AP-2, accurate to 0.1°C), the extra energy used for heating is very small—just 1.02% more compared to an ideal thermostat.
If the thermostat is less precise (AP-3, accurate to 0.5°C), heating energy use increases by 10.47%.
With the least precise thermostat (AP-4, accurate to 1°C), heating energy use jumps by 22.44%.
So again, if your heating source (whether it’s boiler or heat pump), cannot work down to 0.1 precision, then your property might be missing on about 20% energy savings
Let’s not look at the paper and instead start with the basic physics and a quick worst-case estimate. Let’s say we have to heat for 220 days a year and during that heating period accumulate 2500 degree days (should be somewhat representative of London) with an internal setpoint of 20°C. My crappy thermostat however delivers 21°C all the time (very precise, but not accurate). This means that every day I accumulate one extra degree day, ending up with 2720. Thanks to linearity of the underling physics, this translates to an 8.8% increase heating energy use; with heat pumps the electrical energy use increase will be slightly higher because of the reduced COP at higher flow temperatures.
One thing to note here: The added degree days scale with the number of days where I use my heating per year and have to be put in relation to the total accumulated degree days. For colder climates, like in Germany, we get closer to 3800 degree days per year and assuming the same number of days for heating the increased energy use already drops to 5.8%. The article you cite is for Spain, which is much warmer and 1°C inaccuracy will have a higher impact. But they are also probably heating for less days during the year.
In reality things will be much different. If I feel too warm after setting to 20°C and getting 21°C, I will reduce the setpoint and the problem is solved. Real thermostats will however not simply systematically over- or underestimate but rather oscillate around the setpoint with varying amplitude - let’s use 1°C again. So the temperature will overshoot, but also undershoot. The nice thing about the linearity is that we just have to care about the average temperature. When it’s 21°C, we use slightly more energy, when it’s 19°C we use slighly less. Over a year (or even shorter timeframes like a day), it takes the same energy to keep the house at exactly 20°C compared to having it oscillate by ±1°C, provided the average temperature is 20°C.
Finally, let’s quickly look at the paper. They are analyzing an adaptive setpoint technique where (for the heating part) they reduce indoor temperature when it’s colder and increase it when it’s warmer. This of course leads to savings. If the adaptive model calls for 22.268 °C but my thermostat can only be set to 23°C, then of course I will use more energy. This is a discretization error, and if their adaptive energy savings model is built based on continuous temperature adjustability then it is not surprising that quantization/discretization errors can lead to large deviations from the desired result - this is a well-known issue in a number of fields. So what they are saying is that for their adaptive setpoint model, a thermostat adjustability of only ±1°C leads to increased energy use - probably also because they seem to always round towards the next highest integer and never down. This has exactly nothing to do with what most people here are doing: Keeping their house constantly at a temperature they find comfortable and affordable.
Thank you André! I respect your important points raised and I also politely disagree In general, I encourage that we promote for our heating controls to become more open source and more configurable, not less
Which points do you disagree with? You can disagree all you want with the physics, but it’s not going to change reality.
At no point did I advocate for less or closed source controls. I just stated that the benefits of these high accuracy controls in terms of energy savings was severely overstated by you.
With open source, we could swiftly test the algorithm on heat pumps and then celebrate our enhanced efficiency at a cozy pub with excellent German beer and Wurst
.
