AnsweredAssumed Answered

Performance Equation point with event scheduling not generating event for scan before exception in source point

Question asked by fsloane on May 1, 2019

I have an issue occurring with some legacy Performance Equation points that do a simple scaling/unit conversion on a series of flow meters that have intermittent flow, i.e. they may run at 0 for hours, and then jump to 10 for 1 minute, then back to zero. The problem is that the PE point is generating a false ramp-up for minutes before the first non-zero value. This causes significant inaccuracy of downstream totalizers on our meters that see highly intermittent flow.


Raw instrument point:

excdev 0 gpm

excmax 10 min

compdev 1.64 gpm

compmax 1 day


Values read out of "archive editor" in SMT:

12:37:54 PM 0 gpm

1:14:04 PM 0 gpm

1:14:14 PM 7.7 gpm


PE scaled point:

scheduling is based on events in the raw instrument point and

excdev 0 gpm

excmax 10 min

compdev 0.5 gpm

compmax 1 day


Values read out of "archive editor" in SMT:

12:37 PM 0 lb/hr

12:48 PM 0 lb/hr

12:58 PM 0 lb/hr

1:08:14 PM 0 lb/hr

1:14:15 PM 1242 lb/hr


As you can imagine, since the zero value prior to the non-zero value in the "PE scaled point" is 6 minutes prior, the linear interpolation produces a six-minute ramp from 0 to 1242. This is false. The raw instrument tag generates the correct interpolated trend, because it has retained the zero value from the scan (this tag has 10 second scan class) immediately prior to the exception in which the 7.7 gpm showed up.


Turning off compression on the PE-scaled point doesn't resolve this issue. Turning on "step" on the PE scaled tag resolves the issue, but I believe step will be slightly less accurate than linear interpolation and can be misleading to users.


I'm looking to get a general understanding of what's happening here so I can choose the right solution. Does the scan prior to the exception not qualify as an "event" for PE scheduling although it winds up as a value in the archive for the source point of the PE calc?


Thanks for any help.