考虑数控机床性能退化条件的立铣刀服役寿命在线预测关键技术研究

Due to the diverse manufacturing conditions, complex structure and dynamic process, there are vast significant scientific questions remaining not to be solved in the study of tool service life prediction. Currently, the tool condition and tool life are mostly predicted offline through indirect measurement, ignoring the effect on the performance of cutting tool caused by degradation of manufacturing process system. Consequently, corresponding investigation on tool life prediction should be strengthened. This project attempts to forecast end milling tool service life online. Specifically, temperature, current, vibration and the cutting force are collected as condition monitoring data, and the evolution rule of tool service life is revealed through the cutting experiment under coupling of various factors, such as work piece, machine performance and processing conditions. Additionally, modal test , kinetic analysis and Dictionary Learning Models are effectively used to decompose the signals of tool life. Then the intrinsic features indicating tool life are extracted by deep Auto-Encoder neural network, and an online tool life prognostic method is proposed. Finally, framework and design criteria of tool service life prediction will be established. This study is in line with the strategic requirements of the development of national intelligent manufacturing, which can not only enrich the basic theory of intelligent manufacturing, but also improve production efficiency and reduce manufacturing costs.

制造工况多样性，机床结构复杂性和加工动态性等特点导致刀具服役寿命预测研究尚有许多重要的科学问题没有解决。当前，更多是采用间接检测方法对刀具状态或刀具寿命进行预测，忽略了数控机床性能衰退对刀具寿命的影响，成为制约智能制造发展的巨大障碍，刀具服役寿命预测相关研究亟待加强。本项目拟系统开展数控机床性能退化条件下立铣刀服役寿命在线预测研究。利用温度、电流、振动和切削力传感器采集信号，通过不同切削实验阐明工件、机床性能及加工条件等多因素耦合下的铣刀服役寿命演变规律，采用模态测试、动力学分析和字典学习相结合的方法有效分离刀具寿命信号，基于深度自编码神经网络提取反映刀具服役寿命衰退的本质特征，并提出刀具寿命在线预测方法，最终建立刀具服役寿命系统。相关研究符合国家智能制造发展的重大战略需求，其成果不仅可以丰富智能制造的基础理论，而且也有助于提高生产效率和降低制造成本。

制造工况多样性，机床结构复杂性和加工动态性等特点导致刀具服役寿命预测研究尚有许多重要的科学问题没有解决。当前，更多是采用间接检测方法对刀具状态或刀具寿命进行预测，忽略了数控机床性能衰退对刀具寿命的影响，成为制约智能制造发展的巨大障碍，刀具服役寿命预测相关研究亟待加强。本项目以立铣刀为研究对象，开展了数控机床关键部件性能退化条件下刀具磨损及失效机理研究，提出了多因素耦合条件下的刀具服役寿命在线预测方法，最终能够实现立铣刀服役寿命的准确可靠预测。通过开展大量实验，研究了数控机床关键部件性能退化对刀具服役寿命的影响规律，能够准确揭示立铣刀磨损和失效机理；建立了多种能够有效进行刀具服役寿命信号分解的信号处理方法；提出了基于深度学习理论的刀具服役寿命特征选择方法；提出了面向多模式、多工况下的刀具服役寿命建模方法，能够完成在机/在线实时构建刀具服役寿命预测模型；增加了基于工件纹理和机器视觉的直接测试方法研究，能够在线/在机实现刀具磨损量的直接测试，提出了几种适合不同工况的机器视觉处理算法，有效改进了测试精度，为在线实时检测提供了可靠方法；相关研究符合国家智能制造发展的重大战略需求，其成果不仅可以丰富智能制造的基础理论，而且也有助于提高生产效率和降低制造成本。