基于动态KECA的工业过程故障检测  

Industrial process fault detection based on dynamic KECA

作　　者：郭金玉 朱明坤 李元 GUO Jinyu;ZHU Mingkun;LI Yuan(College of Information Engineering,Shenyang University of Chemical Technology,Shenyang 110142,China)

机构地区：[1]沈阳化工大学信息工程学院,沈阳110142

出　　处：《沈阳工业大学学报》2023年第5期576-581,共6页Journal of Shenyang University of Technology

基　　金：辽宁省教育厅项目(LJ2019007).

摘　　要：针对工业过程数据中存在的非线性特性和时间延迟性问题,提出了一种基于动态核熵成分分析(DKECA)的工业过程故障检测方法.将数据集按照时间序列构造增广矩阵,建立DKECA模型,并计算训练数据的Cauchy-Schwarz(CS)统计量及其控制限.将在线监测数据投影到DKECA模型上,其相应的统计量超出控制限的数据作为故障数据.实验结果表明,与传统的非线性方法相比,所提方法能够在保持较低误报率的基础上有效提升故障检测效果,通过引入时间延迟系数提取工业过程的动态变化信息,为传统故障检测方法在动态工业过程中的应用提供了参考.In order to solve the nonlinear characteristic and time delay problem of industrial process data,a fault detection method based on dynamic kernel entropy component analysis(DKECA)for industrial process was proposed.The data set was used to construct an augmented matrix according to the time series;the DKECA model was built;the Cauchy-Schwarz(CS)statistic and its control limit of training data were calculated.The online monitoring data were projected onto the DKECA model.The data whose corresponding statistic exceeded the control limit were taken as fault data.The experimental results show that the as-proposed method can effectively improve the fault detection effect while maintaining a low false alarm rate,compared with the traditional nonlinear methods.The dynamic information of industrial process was extracted by introducing the time delay coefficient through this novel method,providing a reference for improving the application of traditional fault detection methods in dynamic industrial process.

关 键 词：故障检测 非线性特性 时间延迟性 核熵成分分析 增广矩阵 Rényi熵 Cauchy-Schwarz统计量 田纳西伊斯曼过程 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]