SORTiA-Intelligent Model Builder（SORTiA-IMB）is a Model Update and Construction Support
Function used in environments where SORTiA-MPC is installed.
component
SORTiA-IMB consists of a configuration tool, online model estimation logic, and a monitoring tool.
１） Configuration Tool
This tool is used to load the online model file for control and configure various settings necessary for executing the online model estimation logic.
２）Online Model Estimation Logic
Based on the settings configured with the configuration tool, this module accesses the ePREXION Continuous History database and periodically estimates models from operating data. When multiple controllers are involved, the system sequentially estimates a model for each controller and saves the results in separate files.
３）Monitoring Tool
Users can view the latest results of the online model-estimation logic by refreshing the Monitoring-Tool dashboard. If required they can also download the updated online-model file.
※Monitoring Tool
| Main screen | Image
|
| For each controller, the system displays update-recommendation notifications and the current model-update status. When an update is recommended, a <Detail> button appears; selecting it opens the [Select Submodel] screen. | |
| [Select Submodel] screen | Image
|
| Select a submodel from the matrix to view its details. | |
| [Submodel Detail] screen | Image
|
| Parameter reliability is evaluated with three metrics: • Posterior Distribution Variance Comparison※1 • Sum of squared errors (SSE) Comparison※2 • Variation Range Evaluation (min–max variation for each parameter)※3 |
※1.Posterior Distribution Variance Comparison
The posterior variance of each model parameter is examined. Lower variance indicates higher confidence in the estimated values.
※2.Sum of Squared Errors Comparison
Prediction accuracy is assessed by comparing the SSE between predicted and actual values for the currently deployed submodel versus the newly estimated submodel.
A smaller SSE denotes a better fit.
※3.Variation Range Evaluation
Each parameter’s current value is compared with its previous estimate. Small deviations suggest that the parameter remains stable, and the estimate is reliable.
Inquiries
