Fault detection of nonlinear dynamic processes using dynamic kernel principal component analysis

This paper proposes dynamic kernel principal components analysis (DKPCA) approach to bioprocesses monitoring. The basic idea of KPCA is to map the input data into a feature space first via a nonlinear mapping, and then perform a linear PCA in feature space F . The dynamic kernel matrix of DKPCA can capture the nonlinearity and the dynamics of bioprocesses. The proposed method was applied to the fault detection and diagnosis of a simulation benchmark of a biological treatment process. The simulation results clearly show the effectiveness of the proposed approach.