基于加权统计特征KICA的故障检测与诊断方法

针对核独立元分析（kernel independent component analysis， KICA）在非线性动态过程中对微小故障检测率低的问题，提出一种基于加权统计特征KICA（weighted statistical feature KICA， WSFKICA）的故障检测与诊断方法。首先，利用KICA从原始数据中捕获独立元数据和残差数据；然后，通过加权统计特征和滑动窗口获取改进统计特征数据集，并由此数据集构建统计量进行故障检测；最后，利用基于变量贡献图的方法进行过程故障诊断。与传统KICA统计量相比，所提方法的统计量对非线性动态过程中的微小故障具有更高的故障检测性能。应用该方法对一个数值例子和田纳西-伊斯曼（Tennessee-Eastman， TE）过程进行仿真测试，仿真结果显示出所提方法相对于独立元分析（ICA）、KICA、核主成分分析（kernel principal component analysis， KPCA）和统计局部核主成分分析（statistical local kernel principal component analysis， SLKPCA）检测的优势。

Fault detection and diagnosis method based on weighted statistical feature KICA

Aiming at the problem of low early fault detection rate in the nonlinear dynamic process of kernel independent component analysis (KICA), a fault detection and diagnosis approach based on weighted statistical feature KICA (WSFKICA) is proposed. First, the independent component data and residual data are captured from the original data by using KICA. Then, the improved statistical feature data set is obtained by weighted statistical feature and sliding window, and the statistics is constructed from the data set for fault detection. Finally, the method based on contribution chart of monitored variables is used for process fault diagnosis. Compared with traditional KICA statistics, the statistics of the proposed approach has higher fault detection performance for incipient faults in nonlinear dynamic processes. The proposed approach is tested in a simulated case and in the Tennessee-Eastman (TE) process. The simulation results show that the proposed approach has an advantage over ICA, KICA, KPCA and statistical local kernel principal component analysis (SLKPCA).