AnsweredAssumed Answered

Issues with 'True for' in the Start Trigger of an Event Frame Generation Analysis

Question asked by LuisT on Aug 22, 2017
Latest reply on Oct 4, 2017 by AMarques

Using PI Point ‘LT-Test01’ and an attribute with the same name to test, the analysis has a StartTrigger as follows:

EF Analaysis.PNG

(true for 1 minute)

 

The first case

  • Event Frame did not close when it should have.
    • Current time is 05:00:00, ‘LT-Test01’ = 1 with a timestamp of 04:00:00.
    • Event Frame was already generated with start time of 04:00:00.
    • If new value ‘LT-Test01’ = 0 comes in at 05:10:00 with a timestamp of 04:01:30, we expect the EF to close at 04:01:30 but it does not. The EF does not have an End Time.

The second case

  • Event Frame was not generated when expected
    • Current time is 05:00:00, ‘LT-Test01’ = 0 with a timestamp of 04:00:00.
    • At 05:10:00 we get a new value ‘LT-Test01’ = 1 with a timestamp of 03:50:00.
    • EF is not generated even though ‘LT-Test01’ = 1 for longer than 1 minute.

The third case

  • Event frame is generated when it should not
    • Current time is 05:00:00, new value ‘LT-Test01’ = 1 with a timestamp of 04:00:00 (an hour late).
    • EF is generated right away because ‘LT-Test01’ = 1 for longer than 1 minute.
    • At 05:00:10 a new value comes in, ‘LT-Test01’ = 0 with a timestamp of 04:00:10 (also an hour late).
    • The original EF should not have been generated since ‘LT-Test01’ = 1 for only 10 seconds.

 

The analysis works as expected if the data comes in without any data lags or in order.

Other than periodically backfilling, are there any other workarounds for this behavior?

I am happy to provide more details if needed.

Zev Arnold

Akash Naik

Outcomes