摘要
为解决航天产品因制造过程的随机性和动态性而在复杂的多变量加工过程中表现出非线性或多模态等特性,用传统多变量过程控制方法无法有效监控的困难,提出了一种基于自组织混合模型(SOMM)的过程监控方法,以实现非线性和多模态等过程的状态建模和在线异常监控。采用自组织映射结合混合模型,对非线性和多模态等过程状态进行建模,设置控制图阈值;提出了基于最小欧氏距离和负对数似然值的制造过程控制图,在线识别和评估制造过程运行状态。设计的基于SOMM的制造过程监控系统由离线建模和在线监控两部分组成。为验证所提方法在非线性和多模态制造过程中的监控效果,进行了两个仿真实验,并与主元分析、自组织映射等传统监控方法性能进行比较。实验结果表明所提方法在非线性及多模态过程中均有更好的过程监控性能,可用于航天产品加工过程的质量控制。
To solve effective monitoring difficulties in process control when using conventional multivariate process control methods because of nonlinear or multimodal characteristics shown in some complicated multi-variable manufacturing processes for aerospace products due to the randomness of the manufacturing process,aprocess control method based on self-organizing mixture model(SOMM)was proposed to implement the modeling and monitoring of nonlinear and multimodal processes in this paper.Self-organizing map and mixture model were combined in this method to accomplish the modeling of the nonlinear and multimodal processes.The threshold value of the control chart was set.The two control charts were proposed to identify and assess the process states online,which were the minimum Euclidean distance chart and the negative log-likelihood probability chart.The proposed manufacturing process control system based on SOMM was composed of two parts which were modeling out line and control on line.To verify the monitoring performances of this method in nonlinear and multi-modal processes,two simulation experiments were conducted,and then comparison was performed with traditional methods such as principal component analysis and self-organizing map.The experimental results show that the proposed method has better monitoring performances than the regular method in nonlinear and multi-modal processes and it can apply to the quality control of the aerospace products manufacturing.
出处
《上海航天》
2016年第5期42-49,共8页
Aerospace Shanghai
基金
国家自然科学基金资助(51375290)
中央高校基本科研业务费
关键词
多变量制造过程
工序质量
统计过程控制
过程监控
控制图
多模态
非线性
自组织混合模型
Multivariable manufacturing process
Process quality
Statistical process control
Process control
Control chart
Multimodal
Nonlinear
Self-organizing mixture model