Artificial neural network modeling for predicting the carbon black content derived from unserviceable tires for elastomeric composite production

Abstract

Given the increasing need for sustainable solutions and the large amount of improperly discarded end-of-life tires, recovered carbon black (rCB) from tire pyrolysis was investigated as a filler for rubber composites. This study considered rCB as an alternative to commercial carbon black due to its sustainability and CO2 emissions reduction. Composites with varying rCB contents (0 to 50 per 100 rubber) were produced and assessed for mechanical properties, such as hardness, abrasion resistance, and rheometric tests. The findings were used to train artificial neural networks (ANNs) with MATLAB software to predict rCB contents. Input parameters included optimal curing time, minimum and maximum torque, and results of mechanical tests like Shore A hardness and abrasion loss. The model was trained on data from 90 samples, with 10 reserved for validation. The predicted outcomes closely matched the experimental data, with a maximum prediction error of less than 3%. This indicates that ANNs are effective tools for intelligently modeling the curing process of natural rubber mixtures, minimizing material waste, optimizing production time, and determining suitable carbon black contents for desired mechanical properties.