Skip to Main Content
AVEVA Product Feedback


Status Declined
Categories Data Archive
Created by Guest
Created on Aug 20, 2022

Thresholds for PI Points

It would be nice to have threshold attributes for pi points. Then we could configure that values received from the source outside this limits would be ignored (not stored), substituted by the respective threshold value or even by an "Over\Under Range" state. This discussion is pretty old in the community, and I find it quite odd this configuration was never considered (of course there could be specific reasons I am not aware of). The need for filtering negative values of a flow tag, for example, when summarizing data is very common. See this discussion below for better understanding: https://pisquare.osisoft.com/thread/31212-datalink-calculated-filtered-data-odd-behavior Users are led to believe that Zero and Span constitute these thresholds, but in fact they do not filter any value neither impact archiving. Best regards
  • ADMIN RESPONSE
    Aug 20, 2022
    One of the goal of the Data Archive is to faithfully store raw streaming time series data from various data sources. When an interface streams data to PI Points, the server faithfully stores the raw data based on various settings like point type, etc. But by design it does not alter the data (compression settings eliminate excess data, but doe not alter). We have no plans to change that design, but rather we suggest users to ensure the source systems produce the correct values instead of the Data Archive altering the raw data.
  • Attach files
  • Guest
    Reply
    |
    Aug 20, 2022
    What is the source of the data? One can argue that the source should be sending the proper data and not, for example, a negative value for flow, requiring the destination to examine every single value and applying a filter.
  • Vincent Witzel
    Reply
    |
    Aug 20, 2022
    I just had a related request from a customer. They intermittently have very large, erroneous values being generated by their control system, which are muddying up their ProcessBook displays. Right now, they are manually going into PI to delete these values, which is obviously not ideal. I thus asked about filtering functionality in a related PI Square thread (https://pisquare.osisoft.com/message/122852-re-filtering-in-pi-processbookpi-datalink) and was pointed to an existing enhancement request (https://techsupport.osisoft.com/Troubleshooting/Known-Issues/18617OSI8), so I definitely think there is an interest out there for this functionality.