Fuzzy fault detection and diagnosis under severely noisy conditions using feature-based approaches

This paper introduces an approach to fault detection and diagnosis scheme which uses fuzzy reference models to describe the symptoms of both faulty and fault-free plant operation. Recently, some approaches have been combined with fuzzy logic to enhance its performance in particular applications such as fault detection and diagnosis. The reference models are generated from training data which are produced by computer simulation of typical plant. A fuzzy matching scheme compares the parameters of a fuzzy partial model, identified using on-line data collected from the real plant, with the parameters of the reference models. The reference models are also compared to each other to take account of the ambiguity which arises at some operating points when the symptoms of correct and faulty operations are similar. Independent components analysis (ICA) is used to extract the exact data from variables under severe noisy conditions. A fuzzy self organizing feature map is applied to the data obtained from ICA for obtaining more accurate and precise features representing different conditions of the system. The results are then applied to the model-based fuzzy procedure for diagnosis goals. Results are presented which demonstrate the applicability of the scheme.