This project has moved and is read-only. For the latest updates, please go here.

Range Testing for Alarms

Sep 20, 2012 at 3:31 PM

I noticed that there isn't an explicit range test in the alarm subsystem, but I would like to know if it is advised to use the 'Not Equal' test to get the same purpose? For instance, a phase angle test could have the Operator='Not Equal', Alarm Point=0, Tolerance=180. A frequency test could have Alarm Point=60, Tolerance=0.05. Voltage magnitude could have Alarm Point=206000, Tolerance=1500.

Is it expected that this could be a use of 'Not Equal', and is there any problem with using it this way?

Sep 20, 2012 at 8:19 PM

Yes, 'Not Equal' can be used this way. The only real problem with that approach is the lack of hysteresis. Values hovering around the alarm point may cause a lot of alarm chatter.

If you do need hysteresis, then you can also create two separate alarms monitoring the same signal with the same severity; one for the high limit and one for the low limit. Of course, this solution is less convenient. So I would say to use whichever works best for your purposes.

Sep 21, 2012 at 2:47 PM

I wasn't aware that Hysteresis wasn't available for Equal and Not Equal. Not mentioned in the documentation (Alarming in the openPDC and openPG). But if that is true, I will probably just create the two alarms.


For the next upgrade of openPDC, I would beef up the interface to alarming, including a range check. Allowing wild-cards like the ones in FlatlineTest and RangeTest are wonderful, and shouldn't be too hard to implement.