Advanced Control Foundation : Batch and Continuous Processes — Interactive Source for Process Control

Advanced Control Foundation

Tools, Techniques and Applications

Neural Networks for Property Estimation

This Neural Networks for Property Estimation workshop provides several exercises that are used to further explore neural network configuration, model development and verification. The prediction of the Kappa number for the Kamyr digester process is illustrated in this workshop.

Step 1: Examine the module used in this workshop to learn how the neural network block was configured for the Kappa number prediction.

Step 2: Open the neural network application that is used to neural network mode. Position the cursor over the trend window to choose the data that will be used in development of the neural network model.

Step 3: Select the Autogenerate button and then observed the model that is created from the selected data.

Step 4: Choose the Verify selection to determine how closely the predicted Kappa value matches that provided by the on-line analyzer.

Step 5: Download the module that contains the neural network block and observe the on-line current and future Kappa number prediction.