4 Replies Latest reply on Oct 11, 2018 12:39 PM by JimGavigan

    SQC Analysis Problems

    SteveGuepet

      Anyone using SQC analysis?  I'm finding that it's triggering event frames on 1 pt beyond 2 sigma or 1 sigma, despite being set for 1 out of 2 pts beyond 2 sigma, and 4 out of 5 points beyond 1 sigma.  Also I want it to trigger on every point outside control limits.  Right now it only notifies on the first point to violate a run rule.  Even if one point is above 3 sigma and the next below 3 sigma, it only triggers notification on the 1st point outside 3 sigma. 

        • Re: SQC Analysis Problems
          Roger Palmen

          When i doubt, i would recommend using the backfilling preview to look at the input data and the output produced from that. I use SQC, but have not seen any issues with the logic.

           

          What do you mean by "it only notifies on the first point to violate a run rule"? It depends how you setup your analysis what would happen if the SQL status changes from one fault status to the next.

            • Re: SQC Analysis Problems
              SteveGuepet

              Hi Roger,

              So this was part of my data set.  The UCL was set to 8.5 and the LCL set to 8.1.  Starting with the point 7.41 every point is outside 3 sigma and I want to receive a notification on every point.  The SQC analysis only issued a notification on the first point to go outside 3 sigma (7.41).  Looking at the event frame duration, it looks like it considered all the others to be part of the same event.  This may be because it considers the fault status to be the same?  I do have trigger on any input change selected, so it should be running the analysis on each point.

               

              8.36882782
              8.16429615
              8.286250114
              8.286250114
              7.415625572
              8.616015434
              8.56187439
              7.187031269
              8.02320385
                • Re: SQC Analysis Problems
                  Roger Palmen

                  Ah, now i understand. After the 7.41 value, all subsequent values 8.61, 8.56, 7.18, 8.01 are outside your 3-sigma value. Yes, the evaluation is triggered on every value, but the status of the EventFrame does not change: Outside Control Limits.

                  Therefore, whatever the setup of your SQC analysis might be, this will not generate a new EventFrame, and thus no new notification.

                    • Re: SQC Analysis Problems
                      JimGavigan

                      As Roger has stated, this is how event frames work. There really is no way to end the event frame unless you did some type of periodic analysis and the stop trigger was some time period after the above 3 Sigma limit. However, I think you would get irritated at how many events get triggered, depending on the scan rate of that tag. We are doing a fair amount with SQC analyses, and we often write the triggering event (outside 2 sigma, one side of centerline, etc.) into a string attribute and use only certain strings to generate event frames. We want operators to acknowledge certain events in PI Vision and give reason codes as to why they believed the event happened, so we are picky as to what generates the event. So, we don't want to generate new events every time the value goes outside of the limits set.