Lag Estimation

The estimated lag for a variable can be calculated in relation with other fields. Lag estimation is accessed from the Visualization view by clicking on the Lag icon on the list of fields or by clicking on the Auto Lag Configuration button.

To configure a lag for a selected field, users can use the lag estimation provided by the Continuous Troubleshooter to determine lags automatically, or enter known process delays (lags) manually on the Visualization view.

Automatic lag estimation
Manual lag entry

Automatic lag estimation

There are two ways to perform the automatic lag estimation.

Method 1:

Click the Lag icon on the Preparation view next to the desired field
From the Field selection dialog, select the target field(s) against which a lag estimation should be calculated. (NOTE: This is usually the target field for the purposes of the troubleshooting exercise).
Enter the following properties:
- Number of Steps: This is the number of points to be calculated in the specified timespan. Setting this value too low might cause the lag estimation calculation to miss or skip the process delay. Setting this value too high might cause the lag estimation calculation to take very long.
- Timespan: This is the maximum time delay to calculate and consider for process delays on the target fields. Set this value to the maximum known limit for a delay in the process. This value is always specified in seconds regardless of the sampling period in your data.
Click OK to begin lag estimation.
Once the calculations are complete, the lag estimation view is displayed. From this view, use the ruler provided to allocate the desired process delay. Typical use is to position the ruler on the peaks displayed on the graph. These peaks typically indicate a process delay. (Those points in the data where the highest correlation between a historical value for the displayed input and the selected target variable exists across the entire data set)

The steps indicated on the x-axis of the displayed graph indicate the time-steps calculated:

T-0 indicates the current value for the input variable as correlated to the current value for the target variable.

T-1 indicates the value of the input variable on time-step back as correlated to the current value for the target variable (at T-0).

etc.

Once the lag has been set by dragging the ruler, click OK to apply the desired lag to the selected input variable. The Preparation view now displays the configured delay (lag) as set by the ruler.

Method 2:

Click on the Auto Lag Configuration button on the Preparation view.
From the Auto Lag Configuration dialog select the target against which the lag estimation should be calculated. (NOTE: This is usually the target field for the purposes of the troubleshooting exercise). Also select the fields you wish to calculate the lags for.
Enter the following properties:
- Number of Steps: This is the number of points to be calculated in the specified timespan. Setting this value too low might cause the lag estimation calculation to miss or skip the process delay. Setting this value too high might cause the lag estimation calculation to take very long.
- Timespan: This is the maximum time delay to calculate and consider for process delays on the target fields. Set this value to the maximum known limit for a delay in the process. This value is always specified in seconds regardless of the sampling period in your data.
Click OK to begin lag estimation.
Once the calculations are complete, the lags are entered into each lag field on the Preparation view.

Manual lag entry

On the Preparation view, enter the known lag directly on the list of fields by clicking on the appropriate Lag cell on the list view. The Lag value sets a fixed delay to each field. The value is interpreted as a number of seconds to delay the output.