Learning based model for predicting mechanical properties and sustainable filler band for NBR composites using lignin and carbon black

Experimental determination of mechanical properties of rubber composites, such as tensile strength and hardness, involves complex multistage preparation procedures that are laborious and expensive. In this study, a hybrid filler of carbon black (CB) along with a sustainable filler of lignin is added for reinforcement in the nitrile butadiene rubber (NBR) matrix, with the total filler content varying from 10 parts per hundred rubber (phr) to 80 phr. This work aims to develop a data-driven predictive model for the mechanical properties of rubber composites. An artificial neural network (ANN) model using multilayer feed-forward back-propagation has been created to forecast the tensile strength (Ts) and hardness (Hd) of rubber composites. The model predicts the uniaxial tensile response and hardness using input parameters that include total filler and lignin loading levels. The effectiveness of the suggested prediction method was demonstrated by statistical analysis using confidence intervals, showing a prediction error between 5.47% and 3.23% for the Ts and between 3.03% and 1.85% for Hd at 95% confidence intervals. A sustainable green band could be defined in the developed model, which is handy for designers to replace CB with lignin in various NBR based products, such as hoses, seals, etc., without compromising on tensile strength and hardness.