Component fault diagnosis using dynamic co-active neuro-fuzzy systems

This paper investigates the development of a new type of dynamic neuro-fuzzy system with neuronal rules and its application to fault detection and isolation (FDI) of components of a dynamic process. Hybrid learning based on fuzzy clustering algorithm and the steepest-descent method, is used to train the proposed neuro-fuzzy system. The experimental case study concerns the component fault diagnosis of a three-tank system. A neuro-fuzzy simplified observer scheme is used to generate the residuals (symptoms) in the form of one step-ahead prediction errors. These are then analysed by a neural classifier in order to take the appropriate decision regarding the actual process behaviour.