基于监测信号特征的批量钻削工序质量研究  

Cluster of Batch Drilling Process Quality Based on Monitoring Signal Features

作　　者：周友行[1] 董银松[1] 张海华[1] 唐稳庄[1] 

机构地区：[1]湘潭大学机械工程学院,湘潭411105

出　　处：《世界科技研究与发展》2010年第5期569-572,592,共5页World Sci-Tech R&D

基　　金：湖南省高校创新平台开放基金(10K063);教育部留学回国人员科研启动基金项目(2009-1590);国家自然科学基金项目(50975243)

摘　　要：针对批量钻削工序质量的快速监测和分析问题,利用批量钻削加工过程主轴功率信号和声信号的时域统计特征和频域能量特征构造了批量工序钻削过程特征矩阵。应用K均值聚类技术从时域统计、频域能量和时频域综合特征三个角度依据钻孔本身的质量特征对钻孔进行分类,分析批量工序过程特征分布状况,间接反映批量钻削工序质量。对比人工质量检测结果,分析批量钻削过程监控信号时频域特征矩阵聚类纯净度,结果显示其工序质量分布状况检测准确率高达94.19%In order to monitor and analyze the process quality of batch drilling, the character matrix of the batch drilling process is abstracted from statistic characters and frequency characters of the Spindle power signal and AE signal in batch-drilling process. According to the quality characteristics ,the holes are classified into several clusters using the clustering method of K-means. Then the distribution of the character ma- trix is gotten to analyse the batch drilling-quality indirectly. Compared with the manual quality inspection ,the conclusion shows the detection accuracy of the distribution of the batch drilling process quality is up to 94.19%.

关 键 词：批量钻削工序 工序质量 监测信号 特征矩阵 聚类分析 

分 类 号：O213.1[理学—概率论与数理统计]