Design and performance assessment of custom static intermixers in extrusion 3D printing using machine learning–driven image analysis

Uniform blending in multi-material extrusion additive manufacturing is crucial for ensuring consistent material properties. This study evaluates the performance of five static intermixer designs, Split Path, Helix Array, Full Turn Helix, Half Moon, and Cross Bars, integrated into a coaxial extruder system for enhancing the blending of multi-colored polylactic acid (PLA) pellets. Each mixer was tested using a 50/50 mixture of red and blue PLA under controlled extrusion conditions at 210 °C. Mixing performance was assessed through microscopic imaging and machine learning-based analysis, including histogram evaluation, clustering algorithms, and standard color uniformity indices. Results showed that the Split Path and Full Turn Helix mixers provided the most uniform color distribution, with minimal segregation. In contrast, the Helix Array, Half Moon, and Cross Bars designs produced moderate to inconsistent mixing, showing visible streaking and uneven blending. All mixer configurations, however, significantly outperformed the control (no mixer) setup. These findings offer quantitative insights into the effectiveness of various mixer geometries, providing a basis for optimizing mixing strategies in multi-material extrusion additive manufacturing. The study contributes to the development of more reliable extrusion systems for applications such as functionally graded materials, and advanced polymer composites.