摘要
为了实现机械加工过程中刀具寿命在线准确识别,采用时域、频域和小波变换等信号分析方法,提取切削力信号和振动信号中与刀具寿命变化敏感的多个特征,系统输入特征向量通过主向量分析(PCA)方法根据累积贡献率进行优化选择;监测系统根据加工条件自动选择对应的,由两个寿命计算模型构成的动态监测模型,两个模型根据输出精度交替实现刀具寿命计算、在线学习和模型参数更新,最终实现了刀具寿命的在线预测。长期运行结果证明,建立的刀具寿命监测系统能够准确预测刀具的寿命状态,具有良好的自学习能力,在线计算速度高,具有较强的工业推广价值。
In order to accurately predict tool life in the process of manufacturing operation, time domain analysis, frequency domain analysis and wavelet analysis are adopted to extract a series of features from cutting force signal and vibration signal, which are automatically selected and optimized as the system input vectors with principle component analysis (PCA) according to the accumulative contribution rate. The monitoring system could select the corresponding dynamic model according to machining condition, which is composed of two life computational models. The two models alternately accomplish tool life calculation and update model parameters, finally realize online tool life prediction. Long term operating results show that the proposed system can predict tool condition accurately; and has the characteristics of good self-learning ability, high calculation speed and better industrial value.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2010年第11期2416-2421,共6页
Chinese Journal of Scientific Instrument
基金
国家科技重大专项资助项目(2009ZX04014-103-03)(2010ZX04015-011)
中央高校基本科研业务费专项资金资助项目(SWJ-TU09CX019
SWJTU09ZT06)
西南交通大学校基金(2008B13)资助项目