基于压缩采样的微铣削加工刀具磨破损在线监测与寿命预测

Micro-milling has wide applications in micro and ultra precision devices for its prominent capabilities in versatile material processing and complex 3D surface machining. However, due to the nature of ultra-high speed interupted cutting and micro-sclae tool, the tool faliure apprears quickly. This leads to low quality and precision product as a result. The online monitoring of tool conditions has alsyse been an important and difficult issue in micro-milling. The proposed compressive sampling based online tool condition monitoring and prognosis has raised a new idea. This proposal will systematically study the dynamic issues of micro-milling, develop the cutting force modeling approaches under tool wear, and study the mechanism of tool wear and breakage, and their relationships with cutting forces, tool vibration and acoustic emissions. It will further develop the algorithms for compressive sampling which could realize low-rate sampling and denoising while keeping the inner high-frequency properties of signals the at the same time. It will develop multi-scale Hidden Markov Models to model the tool wearing dynamics within and between passes. The developed online monitoring and prognosis system could intelligently estimate the tool conditions and predict tool life. This proposal fuses the advanced studies of precision machining, mechanics, signal processing and control, the results would provide an efficient way to the monitoring of high precision machining process and benefit to the development of high precision CNC control system.

微铣削因具有加工材料的多样性和能实现复杂三维曲面加工的优势在微细与超精密仪器加工等领域有广泛的应用前景。然而微铣刀在超高转速下进行不连续切削，刀具磨破损迅速且难于监测，严重影响加工精度与产品质量，刀具磨破损的在线监测一直是微铣削加工过程控制的难点。本课题提出的基于压缩采样的在线监测方法为解决这一问题提供了新思路。本课题将系统地研究微铣削加工的动态特性，探索在微尺度下刀具磨破损的发生机理及其与铣削力、振动、声发射等信号变化之间的内在联系；通过稀疏表示的数学理论实现对高频信号的低采样率获取，并在保持高频信号本质特征同时实现去噪及特征提取，简化状态监测的模块；建立具有多级多尺度动态特性的隐马尔可夫模型，实现刀具状态智能监测与寿命预测。本研究融合了精密加工、力学、信号处理和智能控制等前沿学科理论，研究成果将为超高速精密加工的在线监测提供有效手段，对新一代的高精密机床控制系统的开发具有积极的推动作用。

微铣削因具有加工材料的多样性和能实现复杂三维曲面加工的优势在微细与超精密仪器加工等领域有广泛的应用前景。然而微铣刀在超高转速下进行不连续切削，刀具磨破损迅速且难于监测， 严重影响加工精度与产品质量，刀具磨破损的在线监测一直是微铣削加工过程控制的难点。本课题提出的基于压缩采样的在线监测方法为解决这一问题提供了新思路。本课题将系统地研究微铣削加工的动态特性，探索在微尺度下刀具磨破损的发生机理及其与铣削力、振动、声发射等信号变化之间的内在联系；通过稀疏表示的数学理论实现对高频信号的低采样率获取，并在保持高频信号本质特征同时实现去噪及特征提取，简化状态监测的模块；建立具有多尺度动态特性的隐马尔可夫模型，实现刀具状态智能监测与寿命预测。本研究融合了精密加工、力学、信号处理和智能控制等前沿学科理论，研究成果将为超高速精密加工的在线监测提供有效手段，对新一代的高精密机床控制系统的开发具有积极的推动作用。.本项目完成了项目规定的所有目标，取得了丰富的研究成果，申请发明专利4项，培养博士生3人，硕士生2人，发表学术论文11篇，其中SCI收录9篇。