I'm pretty new to PI AF as I only started with it 4 months ago. In the meanwhile, I was lucky to spend quite some time on it but last week I noticed something I could not explain immediately. An ideal moment to post something on the forum...
On the image underneath, you will see an analysis I created to summarize quarter-hourly data into daily values. 5 of the 7 calculations are mapped to attributes which are set as pi point reference.
- The first 2 calculations ('ACEventCount' and 'APEventCount') are there to count the events in the timeframe defined.
- Calculation 3 and 4 ('ACTotal' and 'APTotal') are summarizing the quarter-hourly values into daily data (TagMean * eventcount).
- Calculation 5 and 6 ('ACTotalFinalNet' and 'APTotalFinalNet') are there because I want to substract them from each other in calculation 7 and therefor I couldn't use calculation 3 and 4 because I need a 0 instead of NoOutput() to perform the calculation correctly.
- Calculation 7 ('FinalNetTotal') is - as mentioned above - there to substract 2 earlier calculated variables and write the value into an attribute.
As you can see, all calculations are executed without errors.
For your informatoin, the attributes 'E4U_AC_active' and 'E4U_AP_active' are booleans. They are both set to 'True' in this example. I only use them because not every instance of the template will have both active.
The image underneath is a screenshot of the attribute screen. As you can see, 4 out of 5 mapped signals (colored green) are writing something back to the PI-tag but calculation 7 ('FinalNetTotal' - colored red) is writing nothing back as 'Pt created' as signal state is still visible.
Does anybody know why 4 out of 5 calculations perform as expected and one isn't?
Thanks in advance for your feedback.