Template sensor and filtering out occasional negative results from calculation

[Colwilliamsnz (Colin Williams)]

Hi All, I have a template sensor (home power) that derives it value from a subtraction from two other sensors (total power - ev charger power). The two source sensor values each come from a different integration so from a timing perspective the calculation sometimes yields a negative result.

Its not a huge problem just that the resultant graphs I create from it look a little odd with the occasional negative value.

Is there some way to filter negative results OR perhaps there is some other logical way of approaching this problem?

- platform: template
  sensors:
    power_house_consumption:
      friendly_name: House Consumption
      unit_of_measurement: "W"
      value_template: >-
        {% set total_consumption = (states('sensor.fronius_house_load') | float * -1) | round | int %}
        {% set tesla_consumption = states('sensor.twcmanager_1524_charger_load_w') | int %}

        {% set house_consumption = (total_consumption - tesla_consumption) | int %}

        {{ house_consumption }}

[ardysusilo (Ardy)]

- platform: template
  sensors:
    power_house_consumption:
      friendly_name: House Consumption
      unit_of_measurement: "W"
      value_template: >-
        {% set total_consumption = (states('sensor.fronius_house_load') | float * -1) | round | int %}
        {% set tesla_consumption = states('sensor.twcmanager_1524_charger_load_w') | int %}
        {% set house_consumption = (total_consumption - tesla_consumption) | int %}
        {{ ([0, house_consumption, 50000] |sort) [1] }}

See here for explanation-

Limit a value to fall within a range

I found a clever python example for limiting a value to fall within a range and converted it for use in a Jinja2 template. Let’s say you receive a value that must lie within the range of 0 to 100. If the value is less than 0 it should be reported as 0. If the value is greater than 100 it should be reported as 100. Otherwise, report the value as-is. The traditional way is to use arithmetic comparisons to check the value against the range-limits and, with some if-else involved, report the corr…

[Colwilliamsnz (Colin Williams)]

Thanks for this, it works well in so far as negative values are stopped. Very elegant! The reality is that power consumption in the house would never fall to zero (unless there is a power cut of course so I dreamed up a new way of filtering out what are effectively transient conditions (only lasting a single 60 second sample period) by feeding the template sensor “itself” (effectively its previous value) when the resulting computation is <=0.

- platform: template
  sensors:
    power_house_consumption:
      friendly_name: House Consumption
      unit_of_measurement: "W"
      value_template: >-
        {% set total_consumption = (states('sensor.fronius_house_load') | float * -1) | round | int %}
        {% set tesla_consumption = states('sensor.twcmanager_1524_charger_load_w') | int %}

        {% set house_consumption = (total_consumption - tesla_consumption) | int %}

        {% if house_consumption <= 0 %}
          {{ states('sensor.power_house_consumption') }}
        {% else %}
          {{ house_consumption }}
        {% endif %}