engleski

Batch rectification modelling and control using hybrid neural model

Hybrid neural model of the batch rectification process is developed in a way that the material balance equations are used for the description of the still pot and condenser dynamics while the dynamic of the column is modelled by neural net. Necessary data for the training of the net are gathered on the laboratory batch rectification column. Trained hybrid neural model is used for off line simulation of the process and on line in generic model control of distillate composition.

batch rectification; process control; hybrid neural model

nije evidentirano