Add new metric - comfort score

Hi All,

Context
Whatever your heating source is, and assuming that the systems runs as efficiently as it can, what next? Comfort! Comfort (perceived temperature) is perhaps one of the ultimate goals for a system that’s already efficient. You could also say that ultimate efficiency is achieved if the comfort is achieved, because it means you’re only producing as much energy as you need, no less and no more.

New emoncms metric - comfort score
Please correct me if I’m wrong, however, I couldn’t see this metric yet mentioned or offered as a way to filter data on HPM website or directly on emoncms dashboards.

Comfort score
This is perhaps something to discuss, but I thought that the comfort score could be a % of total time where the current indoor temperature is within, let’s say 1 C degrees of the target temperature.

For example, a comfort score (over the selected period of time) of 90% indicates that the house is 90% of time within the range of e.g. 20 C degrees room setpoint.

Going even slightly further, the comfort score could be absolute (whether it’s bellow or above the specified range) or relative, where excessive heat is ignored. This distinction could work well where the house maybe has active cooling system or similar.

Thermostats
Perhaps the reason why we don’t have the comfort score yet, is because none of the existing thermostats is able to provide precise management of heat demand (or cooling). A good thermostat might be able to work within 1 C degree deviation from the target room setpoint on average?

Heat pumps
One of the reasons why heat pumps are considered better over gas boilers is that they provide constant temperature throughout the day, i.e. they are an always-on appliance. However, I am yet to see one heat pump which can work below 1 C degree deviation from the target room setpoint. Of course, they are still much better by default, compared to gas boilers so that is fantastic.

Summary
I believe comfort score is one of the ultimate goals for a heating system from the user perspective. Lack of good thermostat that can precisely manage heat / cool demand might be one of may reasons why that is the case. There perhaps is a gap to start putting more efforts on precise heat demand management from the thermostat perspective.

Additional information
Please see the following example where the thermostat is able to provide more precise heat management, i.e. below 0.1 C degree deviation. The heating source is a gas boiler, however, I’m a big fan of heat pumps, so please don’t see this as some kind of attack on heat pumps. Heat pumps are the future 100% but it just happens that this particular project developed a thermostat that uses PID and PWM (boiler is oversized) and is able to achieve astonishing levels of comfort. If your heat pump supports OpenTherm, the project behind this thermostat will work for you too! Likewise, please let me know if there is a heat pump with a great thermostat :slight_smile:

Emoncms dashbaord
https://emoncms.org/app/view?name=MyBoilerIdealLogicH24OpenthermSAT&readkey=1d29c637a4817acdf6e6e271850c9026

Grafana dashboard
https://gasboiler.grafana.net/public-dashboards/8d44381aafa94fe9bdc87839d8845ada

Thank you
I’m grateful that I can share my observations on this platform, I have learned so much from all of you and I hope you find this interesting. I’m also looking forward to getting a heat pump, and it’s great that I will be well-informed when the time comes, thanks to you. Please let me know if you have any questions.

1 Like

The point of Opentherm is that it allows you to modulate the Boiler Flow Temperature. You would normally PWM a boiler if it only had an ON/OFF input to approximate Opentherm modulation. I assume you are PWM-ing it as the Boiler cannot turn down enough to provide steady modulation at the low flow temperatures in the example, and would revert to modulation at higher flow temperatures?

I can’t think of a Heatpump that offers Opentherm control, it is a rather old Honeywell developed interface. Most Heatpumps use their own proprietary or semi-proprietary control interfaces, eBUS for example, or Modbus.

This is an interesting thought that got me Googling. There is a similar measure, the JND (Just Noticeable Difference), that quantifies the minimum temperature change that a human can perceive. This article from nature …

https://www.nature.com/articles/s41598-023-47880-5

where they measure this to be +/- 0.92degC, so close to your suggestion. So what is the point of controlling to an accuracy of +/- 0.1degC when it is well below the human threshold of sensitivity? Well, if you can control to that accuracy then you can sit very close to the lower bound (for heating) or upper bound (for cooling) and save some energy without compromising the perception of comfort.

2 Likes

Interesting suggestion @critictidier

I’ve have my own custom controller for my Ecodan and have the ability to adjust the room target temperature in 0.1C increments. I’m pretty sure I can feel the difference between 0.2-0.5C. What temperature feels comfortable however seems to change all the time… and is experienced differently by different family members :sweat_smile:

1 Like

I think we notice draughts, often due to different room temperatures, if not external air leakage and a good reason to avoid zoning too much!

1 Like

Thanks for your interest, questions and thoughts!

Yes, the boiler is oversized. The minimum output is 5.2kW, but my heat loss is around 4.5kW when the outside temperature is -3.5°C. However, the boiler is operating at a very low return temperature and therefore, is very efficient (for gas boilers), even at temperatures around 10-15°C outside.

I can only try to explain what it feels like to have a very precise thermostat.

Scenario A: Average thermostat

  • I wake up at 7am, feel hot, look at the indoor temperature and see 20.5°C.
  • By 11am, I start feeling cooler, despite the temperature still being 20.5°C.
  • Sleeping might be too uncomfortable.

Scenario B: More precise thermostat

  • I wake up and don’t feel “any” temperature. Same at 11am (of course, solar gains can make it feel warmer).
  • Sleeping feels comfortable.

Logic: The logic suggests that the more precise the thermostat, the less improvement you get with each increase in precision.

Perception from my own experience: With each order of magnitude increase in precision, the less noticeable the changes in temperature perception become. Of course, setting it very high would still make you feel hot, or the other way around… But, it’s something you experience rather than something you can easily explain.

I can’t even comprehend how this thermostat is able to work with 0.018°C precision (at night in most cases), but it does seem to make a noticeable difference when it comes to comfort.

I believe most people who have already moved to a heat pump from a gas boiler feel an overall improvement in comfort. What I’m suggesting is that with a more precise thermostat, there are even more comfort gains to be had.

So maybe, we should think about two factors that make the comfort score!?
Comfort Score Factors:

  1. Percentage of Time Within Target Temperature (%):
  • This factor measures how effectively the heating system maintains the desired temperature over time.
  • A higher percentage indicates that the indoor environment is consistently at the target temperature, which generally correlates with increased comfort.
  1. Thermostat Precision (Order of Magnitude):
  • This represents how precisely the thermostat can control temperature.
  • Assign a score based on precision:
    • 2 degrees precision = 1 point
    • 1 degree precision = 2 points
    • 0.5 degree precision = 4 points
    • 0.1 degree precision = 10 points
    • 0.015 degree precision = 20 points

Calculating the Comfort Score:

Let’s say:

  • Your indoor temperature is within the target range 95 percent of the time (so Percentage = 0.95).
  • Your thermostat has a precision of 0.1 degrees, which equals 10 points.
  • You assign a weighting of 70 percent to the percentage factor and 30 percent to the precision factor.

Calculation:

  1. Multiply Weight1 (0.7) by the Percentage (0.95):0.7 times 0.95 = 0.665
  2. Divide the Thermostat Precision (10) by the Maximum Precision (20), and then multiply by Weight2 (0.3):(10 divided by 20) times 0.3 = 0.5 times 0.3 = 0.15
  3. Add the two results:0.665 + 0.15 = 0.815

So, the Comfort Score is 0.815 or 81.5 out of 100.

2 Likes

If you’re still in bed at 11am maybe you need to go to bed a little earlier and get up bright and early for nice run in the fresh air - no stats in my house, stays a fairly constant 20-21C 24/7 apart from cooler bedrooms!

Sorry, this is actually all very interesting :grin:

1 Like

:joy:

Example performance of thermostat precision capability

Mine:

Heat pump (or any other heat pump)

Look at the indoor temperature (black line) to visualise the precision

I wonder if there is another term: Thermostat temperature match score? Feels like comfort is something more subjective. E.g you could imagine a user rated comfort score 0-5.

I imagine most on HeatpumpMonitor.org might put a 5 for comfort, but it might be useful to know when a system is not meeting the user desired target temperature due to heat pump system design fault etc.

Unfortunately lack of room target temperature data will probably scupper the feasibility of implementing this on more than a handful of systems.

2 Likes

Here’s mine, pure weather compensation, all rooms generally within 0.5-1C difference - kitchen sometimes a little warmer, no thermostats just some TRVs in upstairs bedrooms.

2 Likes

And a little more variation over the month

1 Like

My NIBE heat pump lets me specify the target temperature in 0.5 C increments. I have it set to 21 C. In my opinion, the standard NIBE controller does a very good job of hitting this target - generally it’s within 0.2 C of what I specify for days at a time.

My “problem” (yes, I know; not really…) is Solar Gain which can easily take the house over 22 C (and south-facing rooms briefly up to 25 C). I have passive and active shading to control the Solar Gain in the Summer, but in the Winter I want as much of that as I can get. The NIBE controller responds to the increased internal temperature by stopping any more heating until the house has cooled to just above the target - but it can’t prevent the initial temperature ‘overshoot’.

Even so, I still see similar temperature control performance to your system (Difference of 1.6 C) - the ‘blips’ in the black Room Temperature line are the sunny days:

From what I’ve read, there’s as much Physiology (and Psychology) involved with the science of “Thermal Comfort” as there is Physics. Trystan captured that nicely earlier in the thread:

Humidity has an impact, as do draughts, as does the radiant effect of cold surfaces like windows. I don’t think it’s as simple as controlling the temperature close to a set-point - that will still feel ‘uncomfortable’ to some people, under some circumstances.

I reckon the Wikipedia page on Thermal Comfort is pretty good: Thermal comfort - Wikipedia

The Passivhaus folks have also done a lot of work on this topic: Thermal comfort parameters [Passipedia EN]

1 Like

Thanks for all the interesting examples of graphs, thermostats and materials to read more about comfort perception! :smiling_face:

We are more likely to express dissatisfaction when something doesn’t meet our expectations. :upside_down_face:Similarly, as thermostat precision improves, we can expect fewer complaints about comfort.

I believe that with every significant (order of magnitude) improvement in thermostat precision, these ‘complaints’ from family members will be at least reduced :blush:.

@TrystanLea Thermostat temperature match score fits the context better, comfort might be too broad given all the factors mentioned so far—maybe it can be shortened somehow! :sweat_smile:

I thought thermal comfort was a factor of temperature and humidity? Ah I see that’s been mentioned

1 Like

It might be worth reading some of what Sue Roaf has written over the years (full disclosure: I disagree with a lot of what she says, but there are some nuggets in there). There’s no such thing as a single comfortable temperature.

Comfort (and even temperature perception) depends an awful lot on what you’re doing and what you’re wearing. When I come in from working in the garden, the house feels warm. After I’ve been sitting around for while it feels cooler. And it depends on whether I’m wearing a fleece or not.

1 Like

Thanks for the andditional information. I think that this only confirms that comfort score would be too broad. There are too many factors to consider.

Thermostat precision capability - indicates precision level the thermostat can work to
Thermostat temperature match score - how long is the indoor temperature within the target temperature as a % (i.e. the margin error depends on thermostat precision capability)

For example
My thermostat precision capability is 0.018 C degrees from the target temperature. And the question is, how long has the indoor temperature stayed within that range throughout the selected period of time?

31st Oct / 1 Nov - night time and so stable environment

A number of objective measures come to mind. We could have some sort of Min/Max/RMS deviation from target. Or maybe a histogram of the deviation from target?

1 Like