Machine learning, response surface method, microscopic analysis, and optimization for mechanical properties of electrically conductive polymer composite fabricated via additive manufacturing

Abstract The benefits of electrically conductive polymer composites are recognized in various industries, including wearable electronics, medical, and sensors, where the combination of mechanical and electrical properties is required. In this study, we investigated the effects of fused filament fabrication (FFF) process parameters on the mechanical properties of conductive polylactic acid (CPLA) composite. The investigated process parameters included nozzle temperature (NT), bed temperature (BT), and cooling/fan speed (FS). After mechanical testing, it was found that the investigated parameters were significantly affecting the mechanical properties (statistically). The process parameters were optimized using a composite desirability based multi response optimization, to balance five mechanical properties simultaneously. The optimization suggested that NT = 220°C, BT = 45°C, and FS = 85% yielded the optimum values of Young’s modulus, ultimate tensile strength, break strain, and strain energy density of 825.63 MPa, 26.37 MPa, 9.24%, and 102.19 × 104 J/m3, respectively. To enhance prediction accuracy, machine learning (ML) models were applied to the same experimental dataset, and Gaussian process regression achieved the lowest validation error of less than 4% across all properties. A functional conductive contact terminal was then fabricated under the optimum parameters as qualitative proof of the concept, which could light the LEDs up. The combined statistical analysis and optimization, SEM, ML predictive approach, and the fabrication of the conductive terminal will serve as a guideline for researchers in additive manufacturing, ensuring the ever-growing use of FFF in various applications.