Piezoresistive Effects in a Highly Sensitive Conductive Elastomer: Experiments and Artificial Neural Network Study

Extensive research is conducted on carbon-based conductive polymer composite sensors due to their excellent piezoresistivity and tunable conductivity. In this article, new carbon fiber/polydimethylsiloxane (CF/PDMS) composites with a high gauge factor of 772 are manufactured, surpassing standard carbon-based conductive polymer composites significantly. Two types of CF/PDMS composites are investigated, including single-length CF (5 mm) and double-length CF (5 and 2 mm) embedded in a 15 × 15 × 50 mm PDMS matrix. The fabrication process is straightforward and cost-effective. A neural network model is used to analyze the rates of strain and resistance change in composite materials. The model is validated through k-fold cross validation and has demonstrated remarkable precision and consistency with the experimental data. This research identifies an ideal proportion of 0.10 wt% 5 mm CF and 0.40 wt% 2 mm CF, which achieve enhanced sensitivity while maintaining the economic feasibility of the CF/PDMS composites. These findings contribute to expanding knowledge in CF/PDMS composite materials and highlight the innovative use of neural networks to enhance the performance characteristics of conductive elastomers.