基于新息矩阵的独立成分分析故障检测方法

来源期刊:中南大学学报(自然科学版)2021年第4期

论文作者:孔祥玉 杨治艳 罗家宇 王晓兵

文章页码:1232 - 1242

关键词:故障检测;独立成分分析;新息矩阵;局部动态阈值;故障检测率

Key words:fault detection; independent component analysis; innovation matrix; local dynamic threshold; rates of fault detection

摘    要:针对核密度估计(kernel density estimation,KDE)方法获取的静态控制限可能不合理以及静态控制限无法跟踪信号动态特性的问题,提出一种基于新息矩阵的独立成分分析故障检测方法(innovation matrix-independent component analysis,IM-ICA)。首先采用正常样本数据建立ICA模型,然后在ICA的基础上引入基于移动窗口协方差矩阵的新息矩阵,最后考虑连续样本的相互影响,将静态控制限改进为局部动态阈值进行故障检测。研究结果表明:IM-ICA故障检测方法通过实时更新控制限能有效反映数据的动态特性;IM-ICA故障检测方法在受极小误报影响下,能有效提高故障检测率,在过程监控中具有较好的故障检测效果。

Abstract: Aiming at the problems that the static control limits obtained by the kernel density estimation (KDE) method may be unreasonable and the static control limits cannot track the dynamic characteristics of the signal, a fault detection method with independent component analysis based on the innovation matrix(IM-ICA) was proposed. First, the proposed method used normal sample data to establish an ICA model. Then, an innovation matrix based on the moving window covariance matrix was introduced on the basis of ICA. Finally, the mutual influence of continuous samples was considered to improve the static control limits to a local dynamic threshold for fault detection. The results show that the IM-ICA fault detection method can effectively reflect the dynamic characteristics of data by updating the control limit in real time. The IM-ICA fault detection method can effectively improve the fault detection rate under the influence of minimal wrong alarms. It has a good fault detection effect in process monitoring.

相关论文

  • 暂无!

相关知识点

  • 暂无!

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号