Labeling Deviations

All new anomalies detected by FactoryTalk Analytics GuardianAI will appear as a deviation. Once a user labels a deviation, it becomes a failure risk until the user marks the anomaly as resolved. The deviation will have associated first principle recommendations. There are several ways to label the deviation:
  • First Principle
    : A user can select from the first-principle recommendations. The following image illustrates that the probable cause might be one of the items shown in the figure, such as Unbalance or Mechanical Looseness.
  • Normal Behavior
    : There might be a case where FactoryTalk Analytics GuardianAI may incorrectly detect the deviation. A user can select normal behavior to train FactoryTalk Analytics GuardianAI to recognize the detected pattern as normal.
  • Other
    : A drop-down menu displays existing labels stored in the FactoryTalk Analytics GuardianAI database for the user to select. Alternatively, the user can click
    Add New
     to create a new label for the detected deviation. For more information on creating a new label, refer to Create a New Label.
Deviation Detection and Grouping
Upon selecting the existing root cause or creating the new label, the user can associate severity and recommended time to resolve the deviation (in days). The time to resolve must be in between 1 and 365.
Provide Feedback
Have questions or feedback about this documentation? Please submit your feedback here.
Normal