Tech Tip #2

[ Follow Ups ] [ Post Followup ] [ Control Consulting ] [ FAQ ]

Posted by Tod Peterson on November 07, 1999 at 23:30:34:

Tech Tip #2 NEWPV detection for Intermittent Signals

For intermittent variables (analyzers) it is necessary to tell the controller when a new measurement is received. Ideally a digital signal from the analyzer itself is set to one indicating a new measurement. The controller then resets the signal to zero until the next new measurement is available.

If a digital signal from the analyzer is not available then the application engineer has to develop some kind of calculation to detect a new measurement. The simplest form is to look at a combination a) elapsed time since the last new signal and b) the change in analyzer signal. This calculation is usually done on the DCS system.

With the latest calculation capability in DMCplus, it is also possible to do the analyzer detection directly in the .CCF file. I have included a sample calculation as an example. Just so you understand the terms:

AN001ADEL (Delta, difference between successive measurement that triggers a NEWPV signal

AN001ALAST (Storage spot for last Analyzer PV measurement)

AN001APV (Current AnalyzerPV measurement)

AN001ANPV (NewPV Calculation Result, 0 = use prediction, 1= update prediction)

AN001ALL (Low Limit, No update if analyzer is lower than this limit)

AN001AUSE (Operator Switch to deactivate NEWPV update, for analyzer calibration, etc.)

AN001AHL (High Limit, No update if analyzer is higher than this limit)

AN001ATIM (Counter since last analyzer update)

AN001ASAM (# of samples before a NEWPV is forced)

This calculation will not update if the USE flag is off or if the variable is outside the high and low limits, defined above. This calculation will also automatically update a NEWPV for this analyzer if the "initialize prediction switch" is set. The two main "tuning" parameters for this calculation are the difference between measurements (DEL) and the sample time (SAM). Usually you should set the sample time equal to one sample more than the actual sample time.

I have used this type of calculation in multiple controllers. There is really only one disadvantage I have come across. That is that you cannot force the analyzer update from the "VIEW" terminal. In fact the NEWPV parameter in view will not have any effect if you change it with these calculations in place.

I would appreciate getting any feedback or improvements on this calculation that you might come up with.

Tod

Follow Ups:



Post a Followup

Name    : 
E-Mail  : 
Subject : 
Comments: Optional Link URL: Link Title: Optional Image URL:


[ Follow Ups ] [ Post Followup ] [ Control Consulting ] [ FAQ ]