Monitoring variables excluded from evolution models in SIMCA®-online (Q929)
Summary:
Sometimes you want to show variables in batch evolution plots, while you don't want those variables affecting the evolution model.
This article describes how you can do this without using the workaround weighting variables close to zero since that does affect the model. Using the description here you can display charts for variables not part of the evolution models, and configure target value alarms for them.
Workaround:
How it works
In short, this solution uses a dummy batch level model to bring evolution variables to SIMCA®-online without including these variables in the actual phase models.
To enable viewing excluded variables, or new variables, in SIMCA®-online:
- In SIMCA®, optionally add a new variable to your batch evolution dataset (right-click and Add variables).
- On the Batch tab, Create the batch level dataset 'Raw data as specified...' and make sure you include the 'Raw Data for Excluded Variables'. Note that here you can select to specifically include exactly the ones you want to monitor.
- Create a batch level model of all newly created batch level variables.
- Important: If the Phase iteration options-button is available on the Select dataset page, select 'As observations'.
- Save and upload the project in SIMCA®-online, either directly using File | Share in SIMCA® or using Upload SIMCA® project in SIMCA®-online.
- In SIMCA®-online, in the configuration, associate the excluded variables to tags.
- Optionally: Specify Target value alarms in the configuration.
More Information:
NOTE! This does not work for generated variables. With generated variables, see the related KB-article:https://umetrics.com/kb/monitor-variables-without-model-influence-simca-online.
