Automated identification of kinetics model in elastomer vulcanization using a rheometer MDR and Rubber-K pattern recognition software

Abstract

This research developed a novel tool for the automatic characterization of elastomers during vulcanization. For this task, robust algorithms for digital image processing and pattern recognition related to curing kinetics were generated and integrated into Rubber-K software. Rheometer tests were performed under ASTM D2084 standard. Experimental data are imported in portable document format and fit the models established by Kamal-Ryan and Isayev-Deng, using a system of first-order ordinary differential equations. During the fitting of the experimental data, the reduced gradient method is used to minimize the error in the objective function. The models used in the algorithms allow higher accuracy and numerical stability compared to similar applications. Rubber-K software provides relevant information to the user during the injection molding parameterization process, allowing production optimization and significantly reducing non-conforming parts. In order to validate the results and accuracy of Rubber-K software, it was compared with data published by other software developers and similar models, whose findings confirm the power and applicability of the software.

Keywords

• Data fitting
• Pattern recognition
• Gradient descent
• Polymer rheology
• Cure kinetics