Integration of principal component analysis and neural classifier for fault detection and diagnosis of Tennessee Eastman process

Fault Detection and Diagnosis are two important areas of research interest in knowledge based expert systems. This paper examines the suitability of neural classifier for fault detection and diagnosis. New methodologies for improving the performances of fault detection and diagnosis systems have been proposed. Within this framework, Principal Component Analysis has been applied, as a feature extraction method for the classification steps. These techniques are applied to simulated data which collected from the Tennessee Eastman chemical plant simulator, that was designed to simulate a wide variety of faults occurring in a chemical plant based on a facility at Eastman chemical. The whole set of the Tennessee Eastman Process faults was evaluated and an improved detecting and diagnosis performance was obtained for all of them. These results exhibit the improved capability of proposed method and the promising potential for the fault detection and diagnosis of industrial applications.