3 Replies Latest reply on Sep 21, 2018 9:53 PM by kbobeck

    summing to the ordinal hour


      I have a meter that records values every 15 minutes on the quarter.  I would like to take all of the reading within an hour and sum them up to the next ordinal hour.  For example: summing the 01:15 reading + 01:30 reading + 01:45 reading + 02:00 reading.


      Any suggestions on how I can achieve this?

        • Re: summing to the ordinal hour
          Asle Frantzen

          Hi David


          You can create an AF analysis scheduled to be run at the full hour, or you can make it run event triggered by your meter tag and ignore the triggered runs where the minute part of '*' was not 0.


          To create a total in AF you have the standard TagTot() function, but that'll give you the time integral over the period and that's not what you want. Instead you can get the TagMean('meter tag', '*-1h', '*') and multiply it with the EventCount('meter tag', '*-1h', '*').



          So I'd recommend you implement the event triggered analysis like this:


          if Minute('*')<>0 then NoOutput()
          else TagMean('meter tag', '*-1h', '*') * EventCount('meter tag', '*-1h', '*')
          4 of 4 people found this helpful
          • Re: summing to the ordinal hour

            This is good but its not quite there yet.  Maybe we need to get into some more details to refine this some more.

            As I mentioned I have two meters; one measures every 15 minutes and the other measures on the hour.  Both of these meters are connected to the same transformer.  I want to sum up the measurements to the transformer.

            Meter A is my 15 minute meter.  Meter B is my hourly meter.  I created a PI Point for Meter A and applied the suggested analysis above to get an hourly value.  This hourly value, I believe, is time weighted averaged not a true summation of the four 15 minute intervals.  Thus my calculation for meter A is not equaled to the summation.  Meter A Calculated value does not equal Meter A Expected value. See table below for details.


            FYI...Originally, I set up a simple roll up analyses to sum all of the values but what I soon discovered was the Rollup analyses wanted to interpolate values on the quarter hour for meter B which increased the true summation.




            measurement time0:150:300:451:001:151:301:452:00
            Meter A raw read3.2323.2732.9942.6730.7240.6760.6640.635
            Meter A Expected Summation 12.172 2.699
            Meter A Calculated Summation 15.282 5.372
            Meter B raw readno readno readno read12.722no readno readno read10.21
            Meter B Expected Summation 12.722 10.21
            Meter B Calculated Summation 12.722 10.21
            Transformer expect rollup 24.894 12.909
            Transformer actual rollup 28.004 15.582
              • Re: summing to the ordinal hour

                Hi David,


                I think the issue here is that it's including the previous on-the-hour value in the total, as well:


                Measurement time1:001:151:301:452:00
                Meter A raw read2.6730.7240.6760.6640.635


                2.673 + 0.724 + 0.676 + 0.664 +0.635 = 5.372


                Instead of doing *-1h, you could try *-45m.



                1 of 1 people found this helpful