Modifiied repo to use PID Climate

[Karl-opec]

I have commited my modifications to this repo in Karl-Opec to use Climate PID and bluetooth Temperature Sensor for the Rooms.

Any Comments welcome

[nliaudat]

Hi,
Do you have good results with PID ?

How many times an hour your valves changes their status ?
I had bad results with PID who was making too many valve position changes.
Here you'll find mine for last 24h

I have pretty smooth temperature in living room according to outdoor temperature changes :

[Karl-opec]

See the following:

In the left side of the chart you see the stable Temp Soll/Ist mean desired-/measured- Temp. Then a window is opened and the measured temp drops, the Öffnung (valve open State) goes up to 100%. Steps are min-movement% appx. every 3-4 minutes. The PID's get a temp value ervery 60sec (heart-beat) of sensor. This means every 60sec PID loop is recalced.

It works fine

[nliaudat]

Glad it works well !

You mentionned github/Technologic in config.yaml to find the correct pid values. I cannot find that repo :(

# PID Parameter github/Technologic values kp=0.05 ki=0,0001 kd=0 ; aktuell update temp ist ca.2,5 min > cycle time 150 sec daher 0,05/(8h*3600s/150s)) 0,00026
  pid_kp: "0.06"
  pid_ki: "0.00011"
  pid_kd: "4.0"

How do you compute your pid values ?
I've used the formula from https://www.mdpi.com/1996-1073/13/8/2068 without good results

I''ll make a try on your solution to compare but my floor heating is low temperature and has a lot of inertia. I mean that I cannot see major differences between pid and "differential target temp" systems.

Can You ?

Regards

[nliaudat]

found repo sorry : https://github.com/Tecnologic/floor-heating-controller

[heinemannj]

I'm actually testing a return flow based thermostat.

1-wire temperature sensors for each heating circuit:
a) Supply flow
b) Return flow

Initially I've started with the "differential target temp" approach (return flow target) which was working more or less accurate - return flow based temperature measurements are extremely sensitive - you can measure temp diffs of ~0.1K within a minute!

Second I've tried your setup of PID => my ESP is rebooting latest after 16 mins ...

WARNING floor-heating-1: Connection error occurred: [Errno 104] Connection reset by peer
INFO Processing unexpected disconnect from ESPHome API for floor-heating-1
WARNING Disconnected from API
WARNING Can't connect to ESPHome API for floor-heating-1: Error connecting to ('192.168.x.x', 6053): [Errno 111] Connect call failed ('192.168.x.x', 6053) (SocketAPIError)
INFO Trying to connect to floor-heating-1 in the background
INFO Successfully connected to floor-heating-1 in 0.029s
INFO Successful handshake with floor-heating-1 in 0.172s

Switching back to the roots.

Any ideas:?

[nliaudat]

Perhaps your sensor must be less sensitive ...

https://esphome.io/components/climate/pid.html :
We recommend putting a filter on the sensor (see filters in Sensor Component) and using output_averaging_samples to calm the PID sensor from a noisy input sensor.

[nliaudat]

I have commited my modifications to this repo in Karl-Opec to use Climate PID and bluetooth Temperature Sensor for the Rooms.

Any Comments welcome

Perhaps it could be a good idea to make a separate branch for your Climate PID integration in that repo. What do you mean ?

[heinemannj]

I've made a lot of code changes:

• naming od IDs and names
• additional temp sensors
• different cron jobs
• adaptations based to my HA dashboard
• ...

Hard to sync.

I will publish my repro within the next days.

Then we can briefly discuss to develop a STANDARD.

[Karl-opec]

I'm actually testing a return flow based thermostat.

1-wire temperature sensors for each heating circuit: a) Supply flow b) Return flow

Initially I've started with the "differential target temp" approach (return flow target) which was working more or less accurate - return flow based temperature measurements are extremely sensitive - you can measure temp diffs of ~0.1K within a minute!

Second I've tried your setup of PID => my ESP is rebooting latest after 16 mins ...

WARNING floor-heating-1: Connection error occurred: [Errno 104] Connection reset by peer
INFO Processing unexpected disconnect from ESPHome API for floor-heating-1
WARNING Disconnected from API
WARNING Can't connect to ESPHome API for floor-heating-1: Error connecting to ('192.168.x.x', 6053): [Errno 111] Connect call failed ('192.168.x.x', 6053) (SocketAPIError)
INFO Trying to connect to floor-heating-1 in the background
INFO Successfully connected to floor-heating-1 in 0.029s
INFO Successful handshake with floor-heating-1 in 0.172s
`

Switching back to the roots.

Any ideas:?

Rebooting the ESP is happening when timing problems happen. Try using ESP-IDF. I have disabled api to homeassistant.
mine ESP-32 runs at least 4 to five days before it autostarts again.

I envision you use the dallas 18B20 sensors to mesure the Supply and Return Flow temperatures? If so, these sensors are very calculation time resource vampire's.

I use in a different situation in total 18 Dallas 18B20 with i2c DS2482-800 chip which communicates per i2c with the CPU and take a lot of computing burden away from thee ESP-32.

These are my quick thoughts

[nliaudat]

Have you tried to disable 1-wire and see if the reboots stops ? (1-wire is a complex protocol with a big library)

For sensors, I mean the best is NTC sensors :

• It's cheap
• You can easyly fix it on your tubes
• There no risk of corrosion
• Analog

You could use an analog to digital multiplexer with only one ESP32 ADC pin

[heinemannj]

Many thanks to both of you for your input and exellent support!

I've done a little bit of code cleanup and increaseed the update time of a couple of sensors.
NOT touched api to homeassistant.
NOT touched my 1-wire sensors.

Switched from Arduino framework to esp-idf:

#esp32:
#  board: esp32dev
#  framework:
#    type: arduino
esp32:
  board: nodemcu-32s
  framework:
    type: esp-idf
    version: recommended
    sdkconfig_options:
      CONFIG_COMPILER_OPTIMIZATION_SIZE: y

Since ~11 hours no reboot!

Because I've made a couple of changes at the same time I can not be 100% sure, if the change to esp-idf has resolved - but on 98%.

Any risks or disadvantages by using esp-idf?

Default ESPHome settings after ESP32 device setup:

esp32:
  board: esp32dev
  framework:
  type: arduino

[nliaudat]

In the past, esp-idf framework was missing functionnalities like web interface and async http requests.
Actually, I mean it perform better than arduino cause the memory management is very well done in esp-idf.
I have migrated all my project to idf without troubles.

[heinemannj]

Can you please provide your PID parameter?
Have you already an up and running PID setup?

[nliaudat]

Hi,
I've created a PID test in the latest release (v108)

You'll find a new template for channels : heating_channel_PID.yaml and a sample config in config.yaml

I mean it could help to stay sync with your repo.

Regards

[nliaudat]

You had BEMF triggers for opening and closing. I have not implemented that cause it's too complicated for lambda to set it up.

[Karl-opec]

I use different opening and closing stops, because the motor current curves for opening and closing are quite different. This
gives better control for pop off's of the valve drives.