Physics-guided machine learning modelling of reactive-extrusion manufacturing of graphene-reinforced silicone and tribological behaviour

A modified reactive-extrusion platform was developed to fabricate graphene-reinforced Ecoflex components with reliable deposition, controlled geometry, and functional surface quality. The system enabled continuous mixing and extrusion of a two-part silicone feedstock through dual-syringe delivery, overcoming common issues of premature curing, material backflow, and layer discontinuity. A D-optimal experimental scheme evaluated the influence of graphene loading, nozzle diameter, extrusion multiplier, and print speed on manufacturability and tribological response. All combinations achieved stable filament formation and acceptable dimensional deviation (∼2%–6%), indicating robust printability within the tested operating range. Graphene reinforcement enhanced layer consolidation and reduced surface irregularity, while retaining low processing resistance. Printed samples exhibited lower coefficients of friction and wear compared to neat silicone, without compromising shape fidelity. Independent validation showed prediction deviations below 2%, confirming consistent process behaviour. The results demonstrate that reactive extrusion can produce functionally enhanced elastomer parts with predictable form accuracy and durable sliding surfaces.