Home heating prediction for Home Assistant
**** OLD --- NOW INTEGRATED INTO PREDBAT https://github.com/springfall2008/batpred ****
Predheat attemps to model water based central heating systems based on a boiler or a heat pump.
Copyright (c) Trefor Southwell October 2023 - All rights reserved
This software maybe used at not cost for personal use only
No warranty is given, either expressed or implied
For support please raise a Github ticket
If you want to buy me a beer then please use Paypal - tdlj@tdlj.net
The app runs every 5 minutes and it will automatically update its prediction for the heating system for the next period, up to a maximum of 48 hours.
The inputs are as follows
- An external temperature sensor, can be a real one or one created by an Internet service
- An internal temperature sensor, ideally from your home thermostat.
- The target temperature sensor, this is what your home thermostat is set to.
- A heating energy sensor in kWh (not stricty required but needed to plot historical usage and calibrate)
- The flow temperature setting of your heating, can be static or a sensor
- Your current energy rates, either from the Octopus Energy plugin or hand typed into the configuration
- Some data about your home that you have to figure out for yourself and calibrate
The outputs are:
- Prediction of the internal house temperature going forward, including times when the heating will be active.
- Your predicted energy usage and costs. The energy usage, if electric, can also be connected into Predbat to help you project your home battery usage.
Future versions will also offer Predbat to run in master mode, controlling your homes heating in the same way as a smart thermostat (e.g. Nest)
- Install AppDaemon if you haven't already - AppDaemon install
- Install HACS if you haven't already - HACS install
- Install Apex Charts - Apex Charts install
- Install Openweather - Openweather install
- Install Predheat using HACS - Predheat install
- You must have GivTCP installed and running first (https://github.com/britkat1980/giv_tcp/tree/main)
- You will need at least 24 hours history in HA for this to work correctly, the default is 7 days (but you configure this back 1 day if you need to)
- Install AppDaemon add-on https://github.com/hassio-addons/addon-appdaemon
- Set the time_zone correctly in appdaemon.yaml (e.g. Europe/London)
- Add thread_duration_warning_threshold: 30 to the appdaemon.yml file in the appdaemon section
- Install HACS if you haven't already (https://hacs.xyz/docs/setup/download)
- Enable AppDaemon in HACS: https://hacs.xyz/docs/categories/appdaemon_apps/
See: https://www.home-assistant.io/integrations/openweathermap
First create an OpenWeather account and then register for a "One Call by Call" subscription plan. This does need a credit/debit card but won't cost anything. You get 1000 API calls a day for free, so edit your limit in the account to 1000 to avoid ever being charged.
Then add in the Home Assistant service and connect up your API key to obtain hourly weather data.
Use HACS to install Apex Charts (Lovelace frontend add-on) - https://github.com/RomRider/apexcharts-card
-
Once installed you will get automatic updates from each release!
-
Add https://github.com/springfall2008/predheat as a custom repository of type 'AppDaemon'
-
Click on the Repo and Download the app
After an update with HACS you may need to reboot AppDaemon as it sometimes reads the config wrongly during the update (If this happens you will get a template configuration error).
- Edit in HomeAssistant config/appdaemon/apps/predheat/config/apps.yaml to configure
- You must delete the 'template: True' line in the configuration to enable Predbat once you are happy with your configuration
- Note that future updates will not overwrite apps.yml, but you may need to copy settings for new features across manually
First you need to edit apps.yaml to configure your system.
Set the mode (mode) to 'gas' or 'pump' depending on if you have a gas boiler or heat pump Set the external temperature sensor (external_temperature) either to a real sensor or create one from the open weather map by adding this sensor to your configuration.yaml file for HA:
template:
- sensor:
- name: "external_temperature"
unit_of_measurement: 'c'
state_class: measurement
state: >
{{ state_attr('weather.openweathermap', 'temperature') }}
Set internal_temperature to point to one or more internal temperature sensors, if you have a heating thermostat then ideally link it to this or to a sensor at least in a similar area of the house.
The weather configuration points to the Open Weather Map sensor by default so should work as-is.
Set the target_temperature to point to a sensor that indicates what your boiler thermostat is set to, or manually enter the temperature setting here.
Set smart_thermostat to True if your thermostat starts the boiler ahead of time for the new target temperature or False for regular options.
Set heating_energy To point to a sensor that indicates the energy consumed by your boiler/heat-pump in kWh. If the sensor isn't accurate then using heating_energy_scaling to adjust it to the actualy energy consumed.
Now you need to make a list of all your radiators in the house, measure them and look up their BTU output at Delta 50 and their volume in Litres. The links below maybe useful for various standard radiators:
- https://www.bestheating.com/milano-kent-straight-chrome-heated-towel-rail-various-sizes-91119
- https://www.bestheating.com/milano-compact-type-11-single-panel-radiator-multi-sizes-available-74174
- https://www.bestheating.com/milano-compact-type-22-double-panel-radiator-multi-sizes-available-74176
- https://www.bestheating.com/milano-compact-type-21-double-panel-plus-radiator-multi-sizes-available-74175
Add up all the BTUs and divide by 3.41 to gain the heat output in Watts and set that in heat_output configuration option. Add up all the litres of water, add in some extra for the piping and an expansion vessle if present (e.g. 5-10 litres) and set heat_volume accordingly.
Set the heat_max_power and heat_min_power to the minimum and maximum power output of your boiler/heat-pump in watts.
Set heating_cop to the nominal COP of your system. For a gas boiler use 1.0 (as the effiency will be based on flow temperature) or for a heat pump set it to the best value which is likely around 4.0 (it will be scaled down for cold weather).
Set flow_temp To the target flow temperature of your system, either via a sensor or as a fixed value. E.g. gas boilers are often set to say 60 or 70 degrees while heat pumps are much lower e.g. 30 or 40.
Set flow_difference_target to be the difference in flow temperature (in vs out) where your heating system will run at full power if it is above. e.g. for gas boilers this maybe something around 40 while on a heat pump it could be much lower e.g. 10.
For your energy rates either have metric_octopus_import point to the current energy rate sensor (gas or electric) or comment it out and enter your rate(s) using rates_import
If you want to account for standing charge set metric_standing_charge to a sensor or enter it manually, if not comment it out.
Now comes the tricky part, we need to calculate the heat loss for your house:
What will help here is historical temperature data, find a time period in the last few weeks when your heating was turned off (for a few hours beforehand) and the house is cooling down. Measure the number of degrees the house drops by in a given time period. Divide that figure (e.g. 1.5 degrees) by the time period e.g. (3 hours) and then again divide it by the average difference between the inside and outside temperature (e.g. 19 degrees inside, 9 degrees outside, so a temperature difference of 4 degrees) = 1.5 degrees / 3 hours / 10 degrees difference = 0.05. Set that figure to heat_loss_degrees. It maybe best to compute this when it's cold out and if you have your heating turned off overnight.
Note in future versions of Predheat I might make this calculation automatic.
Next we need to work out the number of watts of heat loss in the house, this can be done by looking at the energy consumed when the heating comes on. Pick a period of heating, ideally from the time the temperature starts increasing for a complete hour of increase, looking at the increase in temperature in degrees, add to that static heat loss which is heat_loss_degrees * (internal temp - external temp) * 1 hours to get the total degrees accounted for. Now divide that by the external temperature difference again / (internal_temp - external_temp) and multiply the final figure by the energy your system consumed in Watts during that period (can be found either from your sensor or just by looking at your energy bill for the same 1 hour period).
The final figure should be the number of watts your house loses per 1 degree of external temperature difference and be set to heat_loss_watts
Then you can set heat_gain_static to be the static heat output of other things in your house eg. computers and people. You can figure this out by looking at how many degrees of temperature difference your house can maintain without any heating and multiply up your heat loss watts figure by this.
The following are entity names in the Octopus Energy plugin. They are set to a regular expression and auto-discovered but you can comment out to disable or set them manually.
- metric_octopus_import - Import rates from the Octopus plugin
Predbat also include the daily standing charge in cost predictions (optional)
- metric_standing_charge - Set to the standing charge in pounds e.g. 0.50 is 50p. Can be typed in directly or point to a sensor that stores this information (e.g. Octopus Plugin).
Can you manuall set your energy rates in a 24-hour period using these:
rates_import:
- start : "HH:MM:SS"
end : "HH:MM:SS"
rate : pence
rates_export:
- start : "HH:MM:SS"
end : "HH:MM:SS"
rate : p
start and end are in time format of "HH:MM:SS" e.g. "12:30:00" and should be aligned to 30 minute slots normally. rate is in pence e.g. 4.