The paper describes how mapping a neural network into a rule-based fuzzy inference system leads to knowledge extraction. This mapping makes explicit the knowledge implicitly captured by the neural network during the learning stage, by transforming it into a set of rules. By applying the method to transformer fault diagnosis using dissolved gas-in-oil analysis, one could not only develop intelligent diagnosis systems, providing better results than the application of the IEC 60599 Table, but also generate a new rule table whose application also leads to better diagnosis results.Index Terms-Fault diagnosis, fuzzy logic, neural networks.

This paper presents a new approach to incipient fault diagnosis in power transformers, based on the results of dissolved gas analysis. A set of autoassociative neural networks or autoencoders are trained, so that each becomes tuned with a particular fault mode or no fault condition. The scarce data available forms clusters that are densified using an Information Theoretic Mean Shift algorithm, allowing all real data to be used in the validation process. Then, a parallel model is built where the autoencoders compete with one another when a new input vector is entered and the closest recognition is taken as the diagnosis sought. A remarkable accuracy of 100% is achieved with this architecture, in a validation data set using all real information available. IndexTerms-Transformer fault diagnosis, Dissolved Gas Analysis, autoassociative neural networks, mean shift, information theoretic learning. I. INTRODUCTION his paper describes a new approach to the problem of fault detection and identification in power transformers, that reaches 100% accuracy: a diagnosis system based on a set of autoassociative neural networks. The new model gives indication of no-fault or normal condition of the transformer and, if a faulty condition is detected, it identifies the type of fault. This capacity has not been reached before. Power transformer incipient fault diagnosis based on dissolved gas [1] analysis (DGA) has been attempted many times, due to the economic importance of potential equipment failure. It is a problem prone to be addressed by researchers since the publication an IEC norm (IEC 60599 [2]) and a seminal paper [3] that included a data base for diagnosed failures denoted IEC TC10. A number of models have been proposed, adopting a diversity of techniques: expert systems [4], fuzzy set models [5], multi-layer feedforward artificial neural networks (ANN) [6][7], wavelet networks [8], hybrids fuzzy sets/ANN [9], radial basis function neural networks [10],

The paper describes a new methodology for mapping a neural network into a rule-based fuzzy inference system. This mapping makes explicit the knowledge implicitly captured by the neural network during the learning stage, by transforming it into a set of rules. The method is applied in transformer fault diagnosis using dissolved gas-in-oil analysis. Studies on transformer failure diagnosis are reported, illustrating the good results obtained and the knowledge discovery made possible.Index Terms-Fault diagnosis, fuzzy logic, neural networks.

This paper presents a new approach to electrical appliance identification for non-intrusive load monitoring (NILM). In the proposed method a set of autoassociative neural networks is trained so that each one is tuned with the characteristics of a particular electrical appliance. Then, the autoassociative neural networks are set up in a competitive parallel arrangement in which they compete with one another when a new input vector is entered and the closest recognition is accepted to identify the given electrical appliance. The system is trained to recognize specific types of electrical appliances and use the transient power signal obtained from the on/off events for each electrical appliance. To test the proposed method, three public datasets were used, they are, the reference energy disaggregation dataset (REDD), the United Kingdom recording domestic appliance-level electricity (UK-DALE) and the Tracebase dataset containing real residential measurements are used. The accuracy and F-score obtained for the three datasets show the applicability of the proposed method for NILM systems. INDEX TERMS Autoassociative neural network (AANN), electrical appliance identification, non-intrusive load monitoring (NILM).