Hybrid Model for Biopolymerization Process (ε-Caprolactone to Polycaprolactone)

A very attractive and accepted approach to the modeling problem is building a hybrid model, where certain amounts of both phenomenological and empirical information are used. In this paper the mechanistic model is created by using the mathematical equations which are represented in the MatlabTM Simulink environment so as to achieve a control over the bio-polymerization process. This mechanistic model was connected to a Feedforward Neural Network (FANN) model to complete the hybrid model of the process to predict the molecular weight distribution. The hybrid model in the Simulink environment was validated by comparing the results of the hybrid model with that of the experimental results carried out in a bioreactor.