Modelling of Thermal Degradation Process Dynamics of Bioproducts Using Artificial Neural Networks

摘要

A preliminary study aimed at applying artificial neural networks (ANN) to the modelling of thermal degradation process dynamics of labile materials is presented. Experimental data of process parameters gathered from the vibro-fluidised bed dryer was used for network training. A feedforward ANN was trained to predict inactivation kinetics of the dried thermolabile material (saccharolytic activity was used as a biomaterial quality index). It was shown that a small-size network is capable of predicting this quality index in the limited range of drying process parameters (constant static height of the bed, fixed composition of the material).