Ieee std 1159-1995, ieee recommended Practice for Monitoring Electric Power Quality

8.3 Critical data extraction
Often the summaries will not provide an actual solution to a problem. They should, however, help determine
what data needs to be examined more closely. This data is referred to as critical data. In the example of
8.2.3, the summaries showed that the sag and N-G voltage increase had a direct cause and effect relationship.
The transients did not. Thus, the sags and N-G increases would be considered critical data.
8.3.1 Determining critical events from multiple disturbances
The next step in interpreting monitored results is to take the critical data and combine them into events. An
event is the electromagnetic phenomena that resulted in one or more reports from the power monitor. For
example, during the short interruption, which happens while a fault is being cleared, the monitor may report
an L-N sag or interruption, one or more transients, and a waveshape fault or two. All of these describe the
one event of the interruption.
Practically speaking, determining critical events involves collecting all disturbances that appear to describe
the same event, and then analyzing each disturbance in light of the whole. If an L-N sag occurred, did an
increase in N-G voltage also occur indicating a load change on the monitored circuit? If an interruption
occurred were there any waveshape graphs indicating whether the interruption was local or from the utility?
Many times an event will be seen as a group of disturbances, each one providing a valuable piece of informa-
tion needed to put the whole puzzle together.
Isolating an event is done by correlating each graph or report to others with similar time stamps. Be careful
to include graph durations when looking at the time stamps. Figure 16 shows three graphs relating to the
same eventÑa brief local interruption. Notice in 16c) that the waveshape disturbance showing the restora-
tion of power has a time stamp equal to the initial waveshape disturbance time plus the interruption duration
from the sag graph. This is how events are determined.