Modelling the Free Radical Polymerization of Styrene Using Artificial Neural Networks

Abstract

In this work, the modeling of free radical polymerization of styrene is performed using two approaches: phenomenological and empirical. In the last case, artificial neural networks are used, applying two working strategies. First, a single model with three outputs (monomer conversion, numerical molecular weight and gravimetrical molecular weights) based on time, temperature and initial concentration of the initiator (model inputs) was determined. Due to the unacceptable performance results for the molecular weights, a second strategy is developed, modelling each output, separately, as function of the three inputs. Different artificial neural network topologies were tested, the two hidden layer networks performing the best for all the outputs.

Authors and Affiliations

LUCIANA GHIBA, ELENA-NICULINA DRĂGOI, SILVIA CURTEANU

Keywords

Related Articles

Studies Regarding Au(III) Extraction in Aqueous Two Phase Systems Based on Polyethylen Glycol with Different Molecular Weights

This study presents the efficiency of Au(III) ions extraction in aqueous two phase systems (ATPS) based on polyethylene glycol (PEG) 40% (w/w) and (NH4)2SO4 40% (w/w), in the presence of chloride ions with a concentratio...

ANALYTICAL METHODS FOR ANALYZING SOME OILSUSED IN COSMETICS

In this paper three different oils have been analysed: sesame, olive and jasmine. For their analyses standard methods for oils and fats described by American Oil Chemists Society (AOCS, 1993) and pharmacopoeia assays (Br...

SYNTHESIS OF 4’-AMINOPHENYL BENZIMIDAZOLE

In order to synthesize new azomethines derivate from p-aminophenyl benzimidazole we tried to obtain the benzimidazole using synthetic paths from scientific literature. Three different synthetic methods have been used, th...

Thermal Characterization of Chocolate. The Polymorphism of Cocoa Butter

The thermal behavior of seven chocolate samples from various countries of origin, which cover all the three types of chocolate – dark, with milk and white – was analyzed. They revealed the influence of composition on the...

Biosorbent Based on Nonliving Biomass for Textile Dye Retention from Aqueous Media

Due to the benefits of using microorganisms or grafting natural materials as bio-adsorbents, the specialists’ attention has been oriented to developing new techniques of bioremediation, as a part of the environment biote...

Download PDF file
  • EP ID EP305502
  • DOI -
  • Views 56
  • Downloads 0

How To Cite

LUCIANA GHIBA, ELENA-NICULINA DRĂGOI, SILVIA CURTEANU (2017). Modelling the Free Radical Polymerization of Styrene Using Artificial Neural Networks. BULETINUL INSTITUTULUI POLITEHNIC DIN IASI, CHIMIE SI INGINERIE CHIMICA, 63(1), 67-81. https://europub.co.uk/articles/-A-305502