Carbon fiber-reinforced thermoplastic composites (CFRTP) are widely used in automotive, aerospace, and other industries due to their lightweight, high specific strength, recyclability, and superior thermal properties. However, their non-homogeneity and anisotropy present challenging machining characteristics, often leading to damage that deteriorates component performance. It is imperative to conduct numerical simulation and experimental studies on CFRTP to systematically analyze the relationship between cutting mechanisms and the surface integrity of CFRTP. This study aimed to establish an innovative three-dimensional micro-scale cutting numerical model that integrates the differentiated constitutive behaviors and damage criteria of carbon fibers, matrices, and fiber–matrix interfaces—enabling precise characterization of micro-scale damage evolution during cutting. By combining simulation with experimental verification, it unveils the material removal mechanisms and processing damage causes of CF/PEEK, and further pioneers the quantification of the gradient correlation between fiber orientations (0°, 45°, 90°, and 135°) and fracture modes, cutting forces, and surface integrity, thereby addressing the gap of micro-mechanism and quantitative analysis in CFRTP machining. The micro-scale damage mechanisms revealed by the model directly reflect the intrinsic response of individual fibers in the tow, and the collective effect of these micro-behaviors determines the macro-scale machining performance observed in the experiments. A right-angle cutting experiment was conducted to validate the accuracy of the micro-scale numerical model. The mechanisms of fiber fracture, damage patterns, and chip morphology were systematically compared. The experimental results demonstrate good agreement with the outcomes of the numerical simulations. This study aims to bridge the gap between theoretical understanding and practical application of the cutting mechanisms in CFRTP, providing valuable insights for advancements in manufacturing processes.
Cao et al. (Thu,) studied this question.