基于MSVM的质量控制图异常模式识别
首发时间:2019-08-16
摘要:在质量控制图应用过程中如何及时、准确地对控制图异常模式进行识别,对于质量异常诊断具有深刻意义。针对控制图模式识别准确率低的问题,本文以控制图样本统计数据为基础,提取出能有效识别控制图六种模式的形状特征与统计特征,并提出了基于特征的MSVM控制图模式识别方法。该方法的创新点在于其采用有向无环图的思想对多个支持向量机进行分组,首先利用第一组特征识别控制图的总体趋势,再利用第二组特征识别控制图的具体模式。最后以某发动机缸体的质量监测数据为例,验证了该识别方法对常见的六种控制图模式有较高的识别率,结果表明了该方法的可行性和有效性。
For information in English, please click here
Research on Abnormal Pattern Recognition of Quality Control Chart Based on MSVM
Abstract:It is of great significance for quality diagnosis to identify the control chart pattern accurately during the control chart application process. Aiming at the problem of low accuracy of control chart pattern recognition, the feature-based MSVM control pattern recognition method is proposed in this paper. Based on the sample statistical data set, the method extracts the shape features and statistical features that can effectively identify the control chart pattern as the algorithm input variables. The innovation of this method is that it uses the idea of directed acyclic graph to group MSVM. Firstly, the first set of features is used to identify the overall trend of the control chart, and then the second set of features is used to identify the specific mode of the control chart. Finally, the quality monitoring data of an engine block is taken as an example to verify the high recognition rate of the six common control diagram patterns, and the results prove the feasibility and effectiveness of the method.
Keywords: Quality diagnosis Control chart pattern Features MSVM
引用
No.****
动态公开评议
共计0人参与
勘误表
基于MSVM的质量控制图异常模式识别
评论
全部评论0/1000