Abstract
Abstract The additive manufacturing process for continuous fibre-reinforced polymer composites enables the development of lightweight complex structures tailored to specific mechanical properties. The impact performance of such composites is difficult to predict due to the complex interactions between the reinforcement materials and the additive manufacturing process. The impact performance of Onyx-based composites reinforced with fibres and processed via continuous filament fabrication (CFF) using a Markforged Mark Two additive manufacturing machine was studied. Using a Taguchi L27 orthogonal array, 324 specimens were experimentally evaluated. Impact strength varied significantly across fibre systems, ranging from 285 to 936 J/m. Analysis of variance was carried out to assess statistical significance and the contributions of the principal process parameters. It was observed that fibre type and configuration were significant in determining impact properties. A regression model was in good agreement with experimental observations. To enhance the accuracy of this analysis, a machine learning model based on a decision tree ensemble, also known as a bagged trees model, was developed to understand the impact of processing parameters on the properties of fibre-reinforced composite structures. The model was trained using an 80:20 data split, and regression analysis was performed across various parameter settings, yielding R² = 0.90 and RMSE = 0.54 J during validation, indicating that the model is reliable. Scanning electron microscopy was also carried out during this study to analyse fibre failure properties; various failure mechanisms were observed and are presented in this paper. This study developed a framework for analysing the impact properties of fibre-reinforced composite structures processed by additive manufacturing.