Predictive retrieval (wwFilter)

AVEVA Historian supports predictive retrieval. Beginning with AVEVA Historian 2014 R2 Patch 01, the historian can return predictive data based on a "simple linear regression" (SLR) algorithm. More capabilities will be added in future releases.

With AVEVA Historian, you can create a query based on data you have stored to predict additional values in a trend. Historian returns predictive data based on a "simple linear regression" (SLR) algorithm.

For example, based on your currently stored values, you could use the predictive retrieval feature to help predict if a certain production target will be met by the end of the shift. Or, if the Historian loses communication with the data source, you could use predictive retrieval to determine whether and when a tank is likely to become empty.

You can predict:

• Values in between other values.

• Values that extend beyond stored values.

For example, suppose you already captured data for a tag with timestamps up to 3 p.m. on a certain day, but not for the rest of the shift, which ran until 5 p.m., because of a power cut. With predictive retrieval, you can view the interpolated results based between 3 p.m. and 5 p.m. These results are based on the data you received through 3 p.m.

The following is an example of a query that retrieves stored values and reports both those values and additional predictive data:

SELECT DateTime, Value, wwFilter

FROM History

WHERE TagName = 'Tag1'

AND DateTime >= '2014-01-01 0:00:00.000'

and DateTime < '2014-01-01 1:00:00.000'

and wwFilter = 'SLR()'

In this example, "SLR" stands for "simple linear regression," the algorithm used by AVEVA Historian to analyze currently stored values and predict other values within the detected trend.