PR #897 introduces the possibility to define device-level power constraints as sensors. This is interesting enough to create a small section in the documentation about how to apply this new feature to some use-case.

Some examples:

• Model availability of a device. This could be industrial processes or EVs, for instance.
• Model dynamic power limits due to outside temperature. For example, power lines capacity depends on temperature: https://en.wikipedia.org/wiki/Dynamic_line_rating_for_electric_utilities.
• Due to grid congestion, one could limit the charging power of EV (or any device) during some periods only. This probably makes more sense to do at site-level.
• Setting a max and min limits really close so that the device is made to follow a given power profile (with a margin).
• Implement PV curtailment as a storage device. The maximum power of the device is the expected generation and the minimum is 0. The initial SOC should be 0 and the size of the storage should be an upper bound of the PV production on the schedule window.