深孔啄钻振动信号DTW算法改进及应用

针对深孔加工质量检测难、效率低和成本高等问题,提出一种基于振动监测信号分割的改进动态时间规整算法,实现深孔加工质量一致性快速无损检测。首先,以啄钻间歇进给式加工采集的振动信号为研究对象,利用抗噪性能优良的双窗双谱算法对其进行分割;其次,对分割信号在时域和频域内进行特征降维;然后,针对信号长度不等引起的特征不等情况,采用改进的动态时间规整算法(dynamic time warping,简称DTW)进行规整对齐,同时达到减小时间复杂度和防止病态规整的目的;最后,利用求得的累积最短距离评估啄钻阶段振动信号的相似性程度,从而判别啄钻加工质量的一致性。仿真和试验结果表明,该方法能快速有效完成对深孔加工质量一致性无损检测,振动信号分析结果与实际物理检测结果相吻合。

New Approach to Inspection of Deep Hole Peck Drilling Quality Using Improved DTW

Aiming at the difficulty of efficiency and cost in deep hole quality inspection, a method using vibration signal segmentation and improved dynamic time warping (DTW) is proposed to test deep hole peck drilling quality quickly. Firstly, the vibration signal of peck drilling is segmented by the double window bispectrum algorithm with excellent anti-noise performance. Secondly, features of the sub-signal are extracted after segmentation in the time and frequency domain for dimensionality reduction. Then, there being the unequal length of feature vector caused by sub-signal, an improved DTW is used to perform regular alignment, reduce the time complexity and prevent ill-conditioned regularity. Eventually, the similarity between the sub-signals is represented by the cumulative shortest distance, so as to determine whether the quality of the corresponding hole segment of the sub-signal is abnormal. The simulation and experimental results show that the method can detect the quality of deep hole machining quickly and effectively, and the analysis results of the vibration signal is consistent with the actual physical inspection, which shows the reliability and practicability of the method.