3 Replies Latest reply on Dec 1, 2015 1:00 PM by SGoodwin@auriga

    Configuring Analytics with PI AF Online Course November 2015 Run - Videos, Exercises, and Final Project Discussion

    dlong

      Hello everyone,

       

      I created this new topic for discussing the video lectures and exercises as part of the "Configuring Analytics with PI AF Online Course".

       

      • Do you have a question or comment about one the videos you watched?
      • Would you like to share an idea you have for one of the exercises?
      • Are you stuck trying to figure something out for a calculation, roll-up, notification, or event frame you are building?

       

      This is the place for you to ask questions, share thoughts, and even answer other students' questions with your own suggestions!

       

      Thanks,
      Dararith Long

        • Re: Configuring Analytics with PI AF Online Course November 2015 Run - Videos, Exercises, and Final Project Discussion
          SGoodwin@auriga

          One question that may be basic knowledge but here goes - we have points that go out of range (in this case, under range).  This causes issues with the calculations in the analytics.  Is there a way to mask the under range to be zero for the calculations I need.  It is essentially zero at that point, but has gone beyond the range value.  Thanks for your help.

            • Re: Configuring Analytics with PI AF Online Course November 2015 Run - Videos, Exercises, and Final Project Discussion
              dlong

              This is a good question!  Are the under/over range values hard coded (e.g., to a digital state), or is it an analog value that happen to be too high or too low (e.g., random spikes in data)?  If the value is is hard coded, please check to see if you're using Float16 tags.  These automatically get reported as under range or over range.  We discourage users from using Float16 these days, as it's a legacy feature to accommodate storage space at the time.  If you make a simple switch to Float32 on the PI Data Archive, it would then report the actual analog value.

               

              KB00472 - "Under range"/"Over range" behavior for Float16 and Float32 tags

               

              However, if the actual analog values are NOT hard coded to a digital state when under/over range, here are two options you can explore that works:

               

              1) Clamp the values with a formula.  Create a formula and map it to the Tag in question.  Set Min/Max for its settings.

               

              Here's a picture of a Formula called 'Hot-Spot Temperature (Filtered)' that references a tag called 'Hot-Spot Temperature'.  I can set the minimum and maximum values in the properties shown here.  As you can see from behind the configuration, the green box shows the actual tag values in comparison to the filtered one.  The reported value is 80 (because that's what the maximum value I set), even though the actual tag value is 88.6.

               

              Example.png

               

              And here's how it looks like when I raise the maximum to 200.  It would report the normal value of the tag, as expected:

              Example2.png

               

               

              2) Another option is to create multiple IF-Then statements in your calculation, where each is a variable that you can use in a calculation elsewhere (each line listed below is a variable).  You would need at least two variables in your Analysis, and then you can use these variables as INPUTS for your actual calculation.

              • IF 'Tag' < 0 THEN 0, else 'Tag'
              • IF 'Tag' > 100 THEN 100, else 'Tag'

               

               

              Either options are good, depending on how you want to see the data.  Hope this helps.

              2 of 2 people found this helpful