Reduced Kernel Principal Component Analysis for Fault Detection and Its Application to an Air Quality Monitoring Network

Fault detection of nonlinear processes using Kernel Principal Component Analysis(KPCA) method has recently prompt a lot of interest due to its industrial practical importance. However, this method cannot be applied for data sets with a large amount of samples. To overcome this deficiency, this paper proposes a reduced KPCA method based on K-means clustering. This method aims to find a reduced data set among the training data in the input space and uses this reduced data set to built the reduced KPCA model in the feature space. The relevance of the proposed method is illustrated on an air quality monitoring network. The simulation results demonstrate the effectiveness of the new method when compared to the classical KPCA technique.