SORTiA™ Series Automatematic model update function（SORTiA-IMB）

This function uses AI technology to estimate models for control targets based on operational data.Model estimation is executed periodically, so it is possible to build models that always reflect the latest plant characteristics. If the tool determines that a model update is recommended based on estimation results, it notifies the operator. The operator can review detailed information, such as the expected effect of applying the estimated model and decide whether to update the model. SORTiA-IMB consists of the following modules.
 

 Overview of Model Update and Construction Technology by SORTiA-IMB
 

Procedure	Evaluation criteria
1. Determining the Appropriate Interval for the Model	The parameter probability distributions must be narrow, and the data segment estimation must be suitable for model estimation.
	Image
2. Verification of Reproducibility via Simulation	The predicted outputs of the newly estimated model must track the plant measurements more accurately than those of the current model.
	Image
3. Verification of Estimated Parameter Values	The parameters of the newly estimated model must not deviate significantly from the statistical distribution derived from the existing data.
4. Notification to the Operator	If all the above conditions are met, the operator is notified and can decide whether to update the model.