Outlier Handling
Figure 65. Outlier properties
To access this property sheet expand , expand a definition and double-click Outlier.
| Property | Value | Description |
|---|---|---|
| Outlier, Status | Status check boxes (disabled, fault, down, stale, and null are checked by default) | Configures filtering behavior to remove records from a dataset based on the status flags or value of each record. Status values
include: disabled, fault, down, alarm, stale, overridden, null, unackedAlarm and NaN (Not a Number. This is another way, similar
to the InvalidValueFilter block, to filter records based on bad status conditions.
If you check all boxes, the framework filters out all records except those with a status of {ok}, which is always enabled. If you check no box, the framework filters out no records based on status. You may configure additional properties for High Limit and Low Limit (defaults to null check box selected, which does not enforce a limit). This filtering does not apply to value requests. Algorithm blocks may perform additional filtering based on statuses or values. After the framework filters out the records with invalid data, use a missing data strategy to interpolate valid data. |
| Outlier, RawDataFilter, High Limit | null check box (defaults to checked) or numeric value | Optionally, defines a number above which a data value (an outlier) should be excluded from an analytic calculation.
Setting a limit is similar to using a RangeFilter block in an algorithm where values greater than the high limit are excluded from being processed and using an InvalidValueFilter to filter NaN numbers without the benefit of also filtering infinite values. A use case might be that you have a sensor with a range of 0-150 deg F, any readings above 150 would be anomalies or suspect values, which need to filtered out. |
| Outlier, RawDataFilter, Low Limit | null check box (defaults to checked) or numeric value | Optionally, defines a number below which a data value (an outlier) should be excluded from an analytic calculation.
Setting a limit is similar to using a RangeFilter block in an algorithm where values lower than the low limit are excluded from being processed and using an InvalidValueFilter to filter NaN numbers without the benefit of also filtering infinite values. A use case might be that you have a sensor with a range of 0-150 deg F, any readings below zero would be anomalies or suspect values, which need to filtered out. |
| Outlier, Delta Values, High Limit | null check box (defaults to checked) or numeric value | Sets a high limit that applies when Analytics is calculating a delta value.
You might have a history for electrical energy consumption (KWH) that is totalized, which means that an ever increasing value
is being logged. Analytics gets the delta values (difference between each record) to show the electrical consumption for a
period like 15 minutes or a day. The |
| Outlier, Delta Values, Low Limit | null check box (defaults to checked) or numeric value | Sets a low limit that applies when Analytics is calculating a delta value.
You might have a history for electrical energy consumption (KWH) that is totalized, which means that an ever increasing value
is being logged. Analytics gets the delta values (difference between each record) to show the electrical consumption for a
period like 15 minutes or a day. The |