一种基于动态时间弯折和功率信息的在线加工工件自动识别和监控方法

Automatic Identification and Monitoring Method for Workpiece Based on Dynamic Time Warping Technology and Power Information

针对目前车间工件加工进度信息主要靠人工统计、且在混类加工时经常出现统计出错以及工人为加快加工进度擅自改变加工参数造成浪费等现象,在分析数控机床主轴功率特性的基础上,提出一种基于动态时间弯折技术和功率信息的在线加工工件自动识别和监控方法。该方法首先将各种工件加工时可能的主轴功率信息经过特征矢量的选取、特征序列的形成及其时间规正和匹配距离计算以及K均值法的聚类等处理过程,形成特征模板库;在工件加工时实时采集功率信息并进行处理形成特征矢量序列,将该序列与特征模板库中的模板进行时间规正和匹配距离计算等处理,根据匹配距离的大小以及所设定的门限值,判断加工工件的类别或监控加工过程加工参数的正确性。在某装备制造车间的应用验证表明,上述方法能很好地识别加工工件的类别,并起到监控加工过程的作用。

According to analysis of the properties of spindle-power of CNC machine tool,a kind of automatic identification and machining progress monitoring method for machined workpiece,based on dynamic time warping technology and spindle-power information,is proposed.Its purpose is to solve some problems in job shop,such as machining progress information mainly relying on manual statistics incurring frequent error in case of mixed type machining,some workpieces are scrapped because of unauthorized change of machining parameters.At first,through processing processes for the possible spindle-power information of various workpieces,such as selection of characteristic vector,forming of vector sequence,time warping,computation of matching distance,as well as clustering of K-means algorithm,characteristic template library is formed.During machining of workpiece,spindle power information is collected in real time and then processed to form characteristic vector sequence;this sequence is processed by time warping and computation of matching distance,etc.,together with template in the characteristic template library.According to the size of matching distance and the set threshold value,the type of machined workpiece or the correctness of machining parameter are judged.The case study in equipment workshop shows that the method is practical.