The aggregation functionality enables you to create a new tag of an existing tag based on certain operators.
These operators enable you to smooth out signals with noise or highlight certain parts of a trend, such as the minimum or maximum over a certain period.
Aggregated tags can be useful to make monitoring of your process more simple.
- Overview operators
- Creating aggregated tags
- Edit aggregated tags
- Sharing aggregated tags
- Visualize aggregated tags
- Example: averaged tag
- Example: derivative of a tag with noise
- Backward / forward aggregations
Overview operators
Available operators present within the aggregation functionality include:
- 'Average': The average operator returns the average value within the set time interval.
- 'Minimum': The minimum operator returns the minimal value within the set time interval.
- 'Maximum': The maximum operator returns the maximal value within the set time interval
- 'Range': The range operator returns the range between the maximum and minimum value within the set time interval.
- 'Delta': The delta operator returns the difference between the end and starting value of the set time interval. This option can be used as an approximation for the derivative tag.
- 'Integral (/1d)': The integral operator calculates the area under the curve of the set time interval, and returns the result expressed as 'units per day'. If the tag on which you base your calculation has another dimension (eg. tons/hour) you need to take into account a conversion factor (eg. x24 if your tag is in tons/hour) to obtain the correct integral value.
TrendMiner calculates these aggregations as a 'central moving operator', i.e. it will calculate the operator over a given aggregation level left and right of each point. If you are interested in just looking at the past time interval, you have to shift the aggregated tag by half the aggregation time. A step-by-step explanation can be found here.
Note: Averaging periods should be at least the index resolution and can be at most one day.
How to create an aggregated tag
Follow the procedure below to create a new tag using the aggregation functionality:
- Open the tag builder menu
. - Click on the aggregation submenu
. - Select the tag you wish to aggregate.
- Select the operator you wish to use. See above for more information about the different operators.
- Set the time interval.
- Enter a 'Name' and 'Description'.
- Click the 'Save' button.
How to edit an aggregated tag
- Open the tag builder menu.
- Click on 'Saved tags' hyperlink in the aggregation submenu.
- Select the aggregation tag you want to edit.
- Click the edit icon
at the top of the menu. - Edit the aggregated tag.
- Click the 'Save' button.
Note: only the creator of the aggregation tag can edit and delete the tag.
Sharing aggregated tags
When an aggregated tag is created all users can search it via the tag browser and add it to their active tag list. They will be able to use it on the chart and in searches but they will not be able to see the aggregation definition by default.
To allow other users to see the aggregation definition the owner needs to share the aggregation with them in the work organizer. After it has been shared, other users can see the aggregation in their 'Shared with me' tab. By clicking it there it will open the aggregation details panel where they can see the aggregation definition and add it to their view.
Click here to find out how to share items.
How to visualize an aggregated tag
Aggregated tags can be visualized as any other tag once created, alternatively the user can:
- Open the tag builder menu.
- Click on 'Saved tags' hyperlink in the aggregation submenu.
- Select the aggregation tag of interest.
- Click the 'Add to view' button.
Note: Aggregated tags are added automatically to your view after creating them for the first time.
Example: averaged tag
The aggregation functionality enables you to define signals that are smoothed versions of noisy raw data signals, such as hourly averages. Technically every point in time will be replaced by the average of all values left and right within the selected timeframe. Averaging periods should be at least the index resolution and can be at most one day.
In the example below presents an averaged tag called 'ATCAI averaged', as a representation of the 'ATCAI' noisy signal. The signals have been overlaid onto the same scale in the focus chart to show the result.
Backward/forward aggregations
The aggregation panel contains a "Direction" dropdown. The dropdown values include:
- Central
- Forward
- Backward
Aggregations calculates the aggregation value based on the operator for a given aggregation window (Aggregation per X min).
The 'direction' specifies the location of the resulting aggregated value within the aggregation window:
- Backwards: Start of window
- Central: Middle of window
- Forward: End of window
"Forward" calculates the aggregation based on the time period right of the source tag.
Example: derivative of a tag with Noise
Click here to view the example.