Searching for an automation/blueprint for wifi thermostat temp adjust according to outside temperature and sun conditions?

I’m attempting to set an automation or find a blueprint that will take into account weather outside and if there is solar gain from the sun, ie cloudy weather would adjust the air conditioning differently than the extra heat through the windows when the sun is out. Primarily upstairs.

We have in our home a large central air heat pump that is connected to a wifi programable thermostat from Honeywell. It is either in heat mode or cool mode depending on the season, but I would like to make the unit “smart” via automations in HA. We keep our home at around 21C in the summer and winter.

The issue I’m trying to solve from the “set the schedule and forget” of the current thermostat is the variations in comfort upstairs when it’s summer and the sun is out. It can be 21C on the main level and 24C upstairs at night during sunny hot days outside. What I’d like to have set up or find in a blueprint, the state of the weather outside, (maybe with a combination of temp sensors on exterior cameras) temperature and variable cloud cover and apply that to the thermostat temperature setting. Making it a degree or two cooler starting at mid day to counter the solar gain effects probably coming in through the attic into the late evening. I keep the curtains and blinds closed which helps, but this is something I do manually in the summer, and would be great to automate it to keep the family discomfort of the heat at bedtime to a minimum. Or vice versa a particular damp, cool day, shift the Honeywell to heat mode.

The blueprints I’ve seen so far haven’t an adjustment for this variation of direct sun with weather conditions. At least not that I’ve explored yet.

Hello gm,

These search tools might help…

1 Like

Thanks, I was opening some posts in another tab while typing. I found a few situations similar but haven’t found one with sun effects, yet. Still searching though, but if someone out there has accomplished this, and maybe didn’t post details in the forum, I wouldn’t have to remake the wheel, if the solution has been addressed by someone, in their own home, already with more knowledge on the subject. I posted this question in the interim. Also, if I can set it up on my own and it works, I’ll post updated details of what I did or what didn’t work.

Are you perhaps over engineering it a bit? :grin:

If I’ve understood you correctly, the problem is not so much sensing and setting temperatures as predicting what the temperature will be later in the day and setting the Honeywell accordingly.

If this is the case, the simplest approach might be to look at your history and establish what typical temperatures are at 9am, 10am and 11am on days when the bedrooms are too hot in the evening. Those could be triggers/conditions in an automation to adjust the Honeywell at midday.

It’s the temperature in the bedrooms that is the direct issue, not the weather outside - and data from weather observations and cameras are always a bit fuzzy.

If you wanted to get more sophisticated, you could try using a Bayesian sensor to predict whether the bedrooms are likely to be too hot in the evening. That would allow you to add in any number of other factors - not only the weather, but the season, the day of the week, how many people are in the house, are doors/windows/blinds open… It can be surprisingly accurate.

1 Like

I could be overcomplicating it a bit, its happened in the past with my automations. ha That is a good idea to check the history. I’ll observe for a week and maybe go from there with an automation for example, if weather is fine and over 24C (or whatever the data shows) etc mid morning to afternoon, adjust the Honeywell down a degree or two until the late evening. Adding a temp sensor to HA in the bedroom itself could assist on when it is cool enough. Checking early is good advice, with our heat pump it’s better to adjust gradually while the house is still cool from the previous evening, instead of playing catch up I think. I’ll look into the Bayesian sensor. Some variables would be great for hours that I may have to work in the basement “office”, to pause a little of the cooling to not be too cold, compensating for the upper level.