In their system, they have a series of binary indicators that they want to have trigger AML messages on both assertion and deassertion for acknowledgement, but that are only really in an Alarm in one state. They also want these points to each have unique Limit Value Text (which I believe rules out Reaction Matrices). These behaviors together are fine; you just have two limit values and set both to log in AML.

The additional request which I believe makes this impossible is that they also want aggregate alarms for areas (such as by tagging points to Equipment Modeling) which are based on only the "real" alarm state. As far as I can tell, since these points are in "alarm" in both states, then that means that the aggregate alarm for an equipment will always be active. You cannot separate the Limit Value with the "In Alarm Message List" behavior from the Limit Value behavior which drives the Equipment since the Equipment is attached to the entire variable.

Alarm Groups do not have any aggregate alarm behaviors as far as I can tell.

We are already using Alarm Classes on all of our limit values to set alarm priorities. As such, I don't think that I can use the Alarm Area functionality since that links to Alarm Classes, and each Limit Value can only have a single Alarm Class.

Does anyone have ideas for this sort of behavior? Or are we going to have to choose between "generate AML instances in both states" and "have aggregate alarms based on real alarm state"?