Cognite Charts and CDF: Alarms

Related products: Charts

We are connecting our ALARMs Data Source to Cognite Data Fusion. It would be extremely valuable that the alarm setpoints and their metadata can be integrated in Cognite Charts. 

 

We may need to treat the alarm limits as time series, not constants – if the alarm limit changes, that shouldn’t change the historical value.  Ie, while performing an RCA, the alarm limits reflect what they were at the time of the incident on the display.  That being said, they will look like horizontal lines most of the time.
Changes in alarm state should be called out as instantaneous events on the chart with the ability of toggling them on and off.  Knowing when alarms come in and what action is taken (or not taken) is important.

 

 

Hi @ibrahim.alsyed

Once again, this is a very helpful guiding use case – thank you for reaching out about it.

Just to clarify, are the High, Low, Low-low, etc. alarm setpoint values something you would expect to be automatically available in CDF? (meaning, no additional configuration or manual setup needed by a person at Celanese)

And how do folks check these values today? Is it something everyone knows for the equipment or sensors they’re working with? Or do they typically need to check the data source for these values when performing an RCA today?

-----


@Sebastian Skar @Magdalena Rut @Arun Arunachalam FYI regarding monitoring, alerting, and RCA.
@David Liu @Anders Hafreager, perhaps this could be a good case for leveraging FDM functionalities?


NewGathering Interest

@Eric Stein-Beldring  my answers below:

 

The alarm setpoint values will come from the data source - in this case we are using Plant State Integrity or Dynamo. These data sources will be connected and the alarm setpoints should show up including metadata. For example, Alarm setpoint will typically have descriptions, inside and outside actions etc. 

I also believe users in Charts should put their own manual setpoints or alerts independant of the alarms, for their own purposes without impacting the alarm setpoints coming from the data source that is used in DCS.

 

Today, its really challenging, they have to go check the data source seperately or go walk over to the control room to find out the setpoints in the DCS or ask operators. We would like to liberate the data from this data source and make it available and usable in CDF/Charts.

 


Got it, thanks for the added information. 

Do you know yet how these alarm setpoints will be stored or modeled in CDF? (i.e. as events, time series, using relationships, etc.) Or is that something that is being discussed now?