摘要
为了分析火炮自动供弹系统运行状态并实现系统故障的快速诊断,研究开发了基于小波分析法和主元分析法的故障诊断系统。提出了利用小波分析供弹系统中变化剧烈的非平稳监控信号,这种方法对监控信号进行小波分解、非线性阈值化降噪处理完成了信号的重构。采用主元分析法(PCA)从监控变量中提取出主元,按照置信界限85%选取主元个数。在此基础上通过计算SPE和T2统计量,并分析统计量与控制限的状态判断系统是否正常工作。分析过程变量对SPE的贡献图识别故障源,并通过实验验证了该方法的有效性。
In order to analyses working state of artillery automatic feeding system and implement fast fault diagnosis, a fault diagnosis system based on wavelet analysis and principal component analysis be researched. In this method the non-stationary monitoring signal be reconstructed by the wavelet decomposition and nonlinear threshold de-noising reduction, and extract principal compo- nents from monitoring variables by PCA principal, then select the number of components in accordance with the confidence limits of 85%. On this basis the statistics of SPE and T2 can be calculated, through analyze statistics and status of control limits to judge the working state of system. Finally find out the fault source by analyzing contribution plot of process variables on SPE. The effectiveness and robustness of the system has been verified by experiments.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第12期4746-4750,共5页
Computer Engineering and Design
关键词
自动供弹
故障诊断
小波降噪
主元分析法
统计量分析策略
automatic feeding
fault diagnosis
wavelet de-noising
principal component analysis
statistics analysis
