车削加工中刀具磨损对表面纹理的影响：基于ISO 21920-2的特征表征方法
Effects of Tool Wear on the Surface Texture in Turning: A Feature Characterization Approach Based on ISO 21920-2
中文总结 AI 辅助
本研究基于ISO 21920-2特征表征方法，通过车削表面纹理分析刀具磨损，提出平均特征分解法，发现单个物理参数即可解释83%-90%的磨损指标方差，为刀具磨损可靠评估提供新途径。
中文摘要 AI 辅助
车削零件的表面纹理如同工艺参数与刀具磨损状态的指纹，可映射切削刃的状态。经典表面参数如$R_\mathrm{a}$或$R_\mathrm{q}$仅能从全局层面描述形貌，无法对工艺诱发的确定性结构进行空间分辨评估。本研究探究了ISO 21920-2中标准化的特征表征能在多大程度上提取该磨损信息并赋予其物理解释。研究数据库包含12片AlTiN涂层硬质合金刀片（CNMG120408）加工正火AISI 1045钢的粗糙度轮廓，在整个刀具寿命周期内的9个磨损状态下进行测量，每个状态采集3个重复轮廓。研究首先分析了标准化的场参数与特征参数和月牙洼磨损、后刀面磨损及切削时间的相关性。随后采用改进的分水岭分割算法提取旋转刀具沟槽并对其几何特征进行统计评估。新开发的平均特征方法将轮廓分解为确定性分量和随机分量。磨损引起的变化几乎完全由确定性分量承载，且在该分量中主要由切削沟槽的后刀面侧体现。与共聚焦测量结果的对比证实，平均特征可重构参与切削的切削刃几何形状，后刀面侧的陡峭化可归因于副切削刃的缺口磨损。对超过92万种特征表征组合及多变量模型的全面评估表明，仅沟槽级平均最大绝对梯度$\overline{R_\mathrm{dt}}_\mathrm{groove}$一个参数就能解释磨损指标83%-90%的方差，因此单个具有物理意义的参数即可实现可靠的磨损评估。后续研究将探究利用散射光传感器进行在线监测的可行性。
英文摘要
The surface texture of a turned component acts as a fingerprint of both the process parameters and the tool wear condition, imaging the cutting edge. Classical surface parameters such as $R_\mathrm{a}$ or $R_\mathrm{q}$ describe the topography only globally and allow no spatially resolved evaluation of the process-induced deterministic structures. This work investigates how far the feature characterization standardized in ISO 21920-2 makes this wear information accessible and physically interpretable. The database comprises roughness profiles of twelve AlTiN-coated carbide inserts (CNMG120408) machining normalized AISI 1045 steel, measured at nine wear states over the entire tool life, with three replicate profiles per state. The correlation of standardized field and feature parameters with crater wear, flank wear, and cutting time is first examined. Watershed segmentation is then adapted to extract the rotational tool grooves and evaluate their geometry statistically. A newly developed mean-feature approach decomposes the profile into a deterministic and a stochastic component. Wear-induced changes are almost entirely carried by the deterministic component, and within it by the trailing flank of the cutting groove. A comparison with confocal measurements confirms that the mean feature reconstructs the engaged cutting edge geometry, with the trailing-flank steepening attributable to notch wear on the secondary cutting edge. An exhaustive evaluation of more than 920,000 feature characterization combinations and multivariate models reveals that the groove-level mean maximum absolute gradient $\overline{R_\mathrm{dt}}_\mathrm{groove}$ alone explains 83-90% of the variance of the wear indicators, so that a single, physically motivated parameter suffices for robust wear estimation. A follow-up study will investigate inline monitoring using scattered light sensors.
