The ultimate goal is to build two PI Vision events tables based on the same Digital PI Point. Table 1 would have events where Attribute1 <> State2, but only for durations >5m. Table 2 would have events where Attribute1 = State3 for any duration.
The current solution is an EF analysis with two start triggers with no "True for" condition. Then we can apply a duration filter for Table 1 inside PI Vision.
Trigger1 := Attribute1 <> State2
Trigger2 := Attribute1 = State3
The main problem is that this generates many event frames - roughly every 5 minutes. It would be nice to incorporate the duration limit on the event frame capture itself.
When I edit the current solution by adding a "True for 5 minutes" condition to Trigger1, it creates problems when Attribute1 has a sequence of State1 > State3 > State1. In this scenario, we want to capture the entire sequence as one event, and only the State3 portion as a separate event. In PI AF, the "True for" condition causes the event for State3 to close and a new event to open 5 minutes after the first State1 instance. The new event then closes at the end of the entire sequence. The result is that the State3 event is not fully captured, and the first state State1 instance is not captured at all due to the 5m-delayed start time.
Here's one example of such a sequence:
4/2/2024 1:09:08 PM 1
4/2/2024 1:10:04 PM 3
4/2/2024 1:25:54 PM 1
4/2/2024 1:27:51 PM 2
Ideally one event should last from 1:09:09 PM to 1:27:51 PM, and another from 1:10:04 PM to 1:25:54 PM. Instead, one event goes from 1:09:08 PM - 1:14:08 PM, and the other from 1:14:08 PM to 1:27:51 PM.
Is it possible to re-configure this to achieve the desired results? Thanks for your help.