田 颖,王文豪
AuthorsHTML:田 颖，王文豪
AuthorsListE:Tian Ying,Wang Wenhao
AuthorsHTMLE:Tian Ying，Wang Wenhao
Unit:天津大学机械工程学院，天津 300072
Unit_EngLish:School of Mechanical Engineering，Tianjin University，Tianjin 300072，China
Abstract_Chinese:在高速、高精、强适应性的智能制造生产模式下，刀具状态监测既要有效适应多种工况，也要兼顾监测方 法的便捷性，才能满足灵活生产的需求．在刀具状态的各种在线表征信号中，功率信号以其获取方便、信号处理简 洁并便于成本评价等优势逐渐受到关注．针对以机床主轴功率作为单一刀具状态评价指标的可行性和有效性问题， 首先分析多种信号特征与刀具磨损的相关性．分析表明：在多种工况下，主轴功率与刀具磨损的相关系数平均为 0.961，相对其他信号(振动与声发射信号)提升 20%以上．同时，依据与刀具磨损的相关性，筛选出其他信号的最优 特征，并以主成分分析(principal component analysis，PCA)提取主成分，其与刀具磨损的相关性平均低于主轴功率 的 10%以上．主轴功率与刀具磨损具有强相关性，故选其作为刀具磨损状态的单一评价指标．在此基础上，关注刀 具磨损后期，采用粒子群优化算法(particle swarm optimization，PSO)优化模型系数以建立能耗模型，获取多种工况 下主轴功率与刀具磨损的映射，以主轴功率在线监测刀具状态．实验表明此方法基于主轴功率可有效实现刀具状态 监测．
Abstract_English:In high-speed，high-precision，and highly adaptable，intelligent manufacturing systems，tool condition monitoring should not only effectively adapt to multiple working conditions，but also be convenient to meet the needs of flexible production. Among various online sensor signals that characterize tool conditions，power signal has gradually attracted attention because of its convenience of signal acquisition，signal processing，and cost evaluation. Aiming at the feasibility and effectiveness of using spindle power as a single index to evaluate tool conditions，the correlation between various signal features and tool wear was analyzed. The correlation analysis showed that under multiple working conditions，the correlation coefficient between spindle power and tool wear was 0.961 on average，which was more than 20% higher than other signals(vibration and acoustic emission signals). In addition，the optimal fea\u0002tures of other signals were selected by their correlation with tool wear，and principal component analysis was applied to obtain principal components. The correlation coefficient between spindle power and tool wear was more than 10% higher than that of each principal component of other signals. Therefore，spindle power had a strong correlation with tool wear，and it can be selected as an effective indicator to measure tool wear. Based on this，we focused on the advanced stages of tool wear. To obtain the mapping between spindle power and tool wear under multiple working conditions，particle swarm optimization was used to optimize model coefficients to establish an energy consumption model，and then，tool conditions were monitored online based on spindle power. In conclusion，the experiments showed that this method for tool condition monitoring based on spindle power is effective.
Keyword_Chinese:刀具状态监测；主轴功率；刀具磨损；相关性分析；能耗模型
Keywords_English:tool condition monitoring；spindle power；tool wear；correlation analysis；energy consumption model

PDF全文下载地址:http://xbzrb.tju.edu.cn/#/digest?ArticleID=6707