Units about digitalbuildings HOT 6 CLOSED

The intention is for a translation to exist between our representation and that of the BMS. The code below is an example of what a config might look like (see here for further details):

FCU-123:
  ...
  translation:
    zone_air_temperature_sensor:
      present_value: "temp_1"
      unit_values:
        degrees_celsius: "degC"
        degrees_fahrenheit: "degF"
    supply_air_isolation_damper_command:
      present_value: "damper_1"
      states:
        OPEN: "1"
        CLOSED: "2"

There is a mapping of k:v to units that the device can send and what they translate to. In this example, the translation can handle unit values of degC and degF but no K. If you wanted it to, you could add another k:v pair. You would also configure this to contain your BACnet or base system address if you were using it for a traditional BMS, based on however your JSON is configured (the expectation is that the JSON conforms to UDMI).

from digitalbuildings.

The units.yaml file indicates which subfields are linked directly to specified units of measurement. Listed under each subfield is a list of equivalent units. We use SI as outlined in QUDT; it is not intended to be the local standard. For example, in the US we default to CFM for air or GPM for water, or, for a more convoluted example, rate of energy transfer (i.e. power) is KW, MBH, THERM/H, etc. depending on context; by converting everything to common units we remove the ambiguity of context for the purposes of storage. Upstream applications, knowing the units of storage and the conversion rates, can account for local custom.

Noted on the broken link. I have submitted a pull request to fix it.

from digitalbuildings.

Do you see the mapping between the actual BMS point to be external to your representation? For example if the application was an overlay over existing BMS(es) you'd need to know what the BMS units were (along with the object identifier/device identifier etc.). Haystack 4.0 for example seems to model the BMS as part of the ontology to some degree - they include tags to indicate if a point has curr and historised data for example. If the intent is to create a single data warehouse and exclude the ETL process from the RDF then this is probably unnecessary.

from digitalbuildings.

Thanks that's useful. Also I agree on normalisation of units - in the same way that it's considered best practise to convert timestamps to UTC in the persistence layer and to localtime in the presentation layer.

from digitalbuildings.

I'll reopen this rather than create a new issue. I noticed another broken link

https://github.com/google/digitalbuildings/blame/master/ontology/docs/building_config.md#L285

This link is actually broken in the faucetsdn/daq repo as well (i.e. https://github.com/faucetsdn/daq/blob/master/README.md refers to schemas/ and the same readme.md but neither exist in master)

from digitalbuildings.

Thank you !
Just updated with the new link: https://github.com/faucetsdn/udmi

from digitalbuildings.