基于监测信号边际谱和双谱特征融合的孔系钻削质量分析

Holes drilling quality consistency analysis based on the fusion of marginal spectrum characteristics and bispectrum characteristics of monitoring signals

从钻削监测信号数据中挖掘与加工质量相关的信息,可有效实现孔系钻削质量检测。提出一种基于融合钻削过程三向加速度振动和声发射监测信号时频特征的孔系钻削质量一致性评估方法。首先采用振动传感器和声发射传感器监控孔系钻削过程;然后对各钻孔监测信号进行Hilbert-Huang变换和高阶谱分析,提取各孔监测信号的边际谱和双谱特征;应用主成分分析方法进行特征降维,特征融合聚类分析,直观获得各钻孔钻削过程监测信号时频特征波动状况。基于钻削过程质量波动与监测信号边际谱的频率能量特征和双谱特征数值变化的耦合关系,并与孔系钻削加工人工质量检测对比表明:融合孔系钻削监测信号边际谱特征和双谱特征进行数据聚类研究可有效进行孔系加工质量的一致性检测,快速分析和识别质量异常钻孔。

The information mined from the drilling process monitoring signals data could be helpful to inspect the holes drilling quality. A holes drilling consistency inspection method was presented based on the fusion of marginal spectrum characteristics and bispectrum characteristics of monitoring signals. Three acceleration vibration sensors and an acoustic emission sensor were used to monitor holes drilling process. The Hilbert Huang transform and a high order spectrum estimation were used to analyse each hole＇s drilling monitoring signals,of holes drilling monitoring signals marginal spectrum and double spectrum features of rach hole were extracted from the monitoring signals. Finally,the principal component analysis（ PCA） method was used to realize features dimension reduction,features fusion and features clustering. the computer conclusion show the change condition of these features directly and clearly. Based on the coupling relationship between the drilling process quality fluctuation and the numerical changes of these features,and comparing with the artificial quality test results of drilling holes,it is concluded that the data clustering analysis on the fusion of marginal spectrum characteristics and bispectrum characteristics of holes drilling monitoring signals can realize the holes drilling quality consistency detection effectively and also analyse and identify the abnormal drilling quality rapidly.