中文摘要：

针对核独立成分分析故障检测时忽略各独立成分分量对系统故障贡献度的差异,提出一种基于加权核独立成分分析的故障检测方法.使用核独立成分分析提取过程变量的独立成分,根据核密度估计衡量各独立成分分量对系统故障的贡献度,对各独立成分分量赋予不同权重,突出包含有用信息的独立成分分量,引入局部离群因子在特征空间构造统计量进行故障检测.基于数值仿真和Tennessee Eastman数据集的仿真结果表明了所提出方法的优越性.

英文摘要：

To solve the problem of missing different contribution of each kernel independent component for the system fault by using the kernel independent component analysis fault detection method, a weighted kernel independent component analysis is proposed for fault detection. The kernel independent component analysis is performed to extract the kernel independent components. Kernel density estimation is used to evaluate the contribution of each kernel independent component. According to the contribution, different weighting values are set up to highlight the kernel independent components with more useful information. Finally, the local outlier factor is used to establish the monitoring statistic in the feature space. The advantages of the proposed method are demonstrated by the results based on a numerical and the Tennessee Eastman process simulation.