Sensor Fault Detection and Isolation of an Air Quality Monitoring Network Using Nonlinear Principal component Analysis

Recently, fault detection and process monitoring using principalcomponent analysis (PCA) were studied intensively and largely appliedto industrial process. PCA is the optimal linear transformation withrespect to minimizing the mean squared prediction error. If the datahave nonlinear dependencies, an important issue is to develop atechnique which can into account this kind of dependencies.Recognizing the shortcomings of PCA, a nonlinear extension of PCAis developed. This paper proposes an application for sensor failuredetection and isolation (FDI) of an air quality monitoring network vianonlinear principal component analysis (NLPCA). The NLPCA model isobtained by using two cascade three layer RBF-Networks. For trainingthese two networks separately, outputs of the first network areestimated by using principal curve algorithm (Harkat, 2003)...