Artificial Neural Network Modeling of AGET ATRP of MMA in an Emulsion Batch Reactor

Atom Transfer Radical Polymerization is known to produce 'tailor-made' polymers for special applications in biomedical areas. Running this polymerization in dispersed media is environmentally beneficial but presents challenges in developing a mechanistic model due to the heterogeneous medium and kinetic uncertainty. Alternatively, this thesis discusses an artificial neural network modeling approach to kinetically model the AGET ATRP process. This study develops a forward ANN to predict conversion and molecular weight and reverse ANN to predict reaction conditions using experimental data collected from past studies. Additional experimental work was done to confirm the previous data and the ANN's predicting capabilities. The results demonstrate that the proposed ANN models agree with experimental data with R values greater than 0.98 and can be used to predict AGET ATRP process performance and plan experimental scenarios. The study is a promising way for the design of an industrial ATRP process in dispersed media.