Has someone already come up with an app or rule that calculates "feels like" temperature, based on indoor metrics? This is to help settle some arguments with spouse about thermostat settings. Right now, I am driving the setPoints for cooling and heating based on dewpoint ranges. If I can convert these to "feels like" temps, it translates better for the spouse.
I wonder, though, if this is even worthwhile because some articles I have read indicate the heat index cannot be applied to temps < 80F and, similarly, temps > 50F for wind chill. Additionally, if it is indoors, what wind speed value would we be plugging into the formula? Does heat index even apply to indoors?
This might be useful, its a Virtual Dewpoint App from Guffman:
Periodically calculate a dewpoint in degrees Farenheit, given a humidity and a temperature. Allow for time series smoothing using an exponential smoothing filter algorithm.
Thirty five years of marriage experience says that if partner is cold, then partner is cold. Raise the temperature. If partner is hot, then partner is hot. Lower the temperature. Forget about science. Being right rarely wins a debate.
This may be informative, I stumbled upon it while on my quest for ultimate indoor comfort control...
Since I use Node Red with Hubitat, it was convenient to install this node, build a flow to calculate a thermal comfort index and sensation, and write the results into a Hubitat 'Virtual Thermal Comfort' user device driver. Sensation is a string mapped to ranges of the comfort index, defined by:
The Hubitat virtual device events are written to InfluxDB (time series database), then rendered in various Grafana dashboards. Here's an example of the Grafana output:
The objective was to understand (and then automate) how to manipulate the HVAC system throughout the day and night to keep the occupants happy.
Ultimately, I settled on dewpoint control, varying the setpoint on a time of day schedule. The ASHRAE Thermal Comfort exercise was interesting, but I didn't really learn anything that was actionable. I may revisit this after accruing several months of data, but probably just 'polishing the cannonball' at this point.
Any thoughts about how this might be useful are welcome
I think this is going to depend on where one lives. Because I think our bodies get conditioned to different ambient temperature and relative humidity. If I make my house reach a target dewpoint of 50F in sleep mode, I wake up with a very dry nose and throat. And I am guessing this is because my body has become used to the higher ambient relative humidity in New Orleans.
Anyway here's what I found works as a reasonable balance between comfort and equipment runtime.
Summer: 55F (Home), 57F (Away), 53-54F (Sleep)
In my home, in all three cases, the temperature stays between 70-73. The difference is largely by humidity levels. So it is dryer (and therefore cooler) at night.
Dewpoint control doesn't work well in the (brief) winter here. So I use temperature control when the HVAC is in "heating" mode.
@Guffman - I also would be interested in knowing what dewpoint comfort levels you set your equipment for. And also some general indication of where your home is. Thanks!
That's because in New Orleans it doesn't stay cold enough and long enough to require any humidification. Same here for San Antonio. If we lived up north, humidification would be required and the dewpoint control might be helpful.
You are exactly correct. The recent (brief) cold spell that lasted about 1 week around Christmas was the first time in years that the indoor relative humidity dropped into the 40-45% range. And I found that setting up my manual large evaporative humidifier made 70F feel much more comfortable than before I plugged the humidifier in.
It was also the first time I saw my heat-pump go into defrost mode (albeit briefly, and once a day). At least I know the frost sensor and defrost board work
Hello, all, from Georgia! @Sakman@marktheknife I have a heat pump, too. One window cracked during the cold blast the past December. Guess we kept it too hot in the house.
So far, great ideas and questions! Please keep them coming. If anything, I enjoy the banter.
@Guffman , I really like the approach of reviewing comfort levels over a period of time, in order to determine the dewpoint ranges throughout the day.
Would I be able to use your Hubitat virtual thermal comfort driver, if you are willing to share it? Is nodered a prerequisite, though? Sorry, not very tech-literate.
How well have the ASHRAE index-based sensations match your and fellow home occupants' subjective sensations, so far? As @aaiyar pointed out, it can be different for each person.
@aaiyar , I actually based some of my initial settings on what you posted in your topic. Thank you for sharing that.
I have gleaned 3 ideas so far:
add datapoints representing sensations/"feels like" (subjective or predicted/calculated) to my pre-existing hvac and sensor charts, to aid in dewpoint range determination
maintain dewpoint ranges by time of day
maintain also by season
In keeping with "polishing the cannonball" (I like this expression), I also want to balance out with energy cost/equipment runtime, as @aaiyar mentioned. So, I might input my energy bill/usage, at some point. I don't have a smart meter, so it will have to be manually input.
Home is in the greater New Orleans area, so same weather/climate conditions as you!
Dewpoint is set to 57F starting around wake-up time, then lowered to 55F a little before go-to-sleep time. This seems to keep everyone comfortable, and also keeps equipment runtimes more or less reasonable.
Happy to share the Thermal Comfort device driver, but it's really just an attribute placeholder, not a compute module. Node-Red is doing all the computing. I cracked open the javascript that powers the node-red-contrib-comfort node - given the complexity of the calculations, I decided there was little benefit in re-implementing as either a Hubitat app and/or a device driver.
As for the correlation between occupants' sensations and the ASHRAE index, they generally agree. There are several input parameters to the index computation that can't be measured (well, not without really going to extremes!) - airspeed, metabolic rate, clothing level. But these can be inferred 'close enough', given the situations that dominate in the time of day ranges I'm varying septpoints within.
So as I said, interesting exercise, but it didn't lead me to believe that manipulating thermostats and dehumidifiers using thermal comfort index as the controlled variable was better than dewpoint.
Strange, but the occupants aren't as impressed with the ASHRAE computations and fancy Grafana visualizations as I am. Nobody was demanding to know the current thermal comfort index and sensation. Does that sound familiar to the home automation enthusiasts out there?
I think dewpoint control gets you 95% of the way there, given my experiences here in my climate.
Does that sound familiar to the home automation enthusiasts out there?
