Improving the accuracy of carbon nanotube yarn 3D printing using machine learning

Abstract

To overcome the limitations of conventional continuous carbon fiber 3D printing in achieving precise curved printing and intricate shaping, a 3D printing technique based on carbon nanotube (CNT) yarn was proposed, offering finer and more accurate fabrication capabilities. However, the contributions of two critical features of CNT yarn—its fine diameter and yarn twist—to enhanced printability remain inadequately understood. This study explores the impact of these features on printing precision through a combination of experimental methods and machine learning approaches. The findings reveal that yarn twist plays a more significant role than diameter in reducing radius errors during single-layer circular printing. A predictive model developed in this study achieved an $R^2$ value of 0.888 and reduced radius error magnitude by approximately 79.3% when feedback was incorporated into the printing process. These results highlight the potential of CNT yarn to advance the precision of 3D printing technologies.