Industrial Automatic Control Systems and Controllers Annotation << Back Application of Artifi cial Neural Networks to Detect Anomalies in the Information Exchange of Protective Automatics of a Digital Substation Ivanov S.O., Kopisheva T.N., Nikandrov M.V. The article is devoted to the possibility of using artifi cial neural networks to identify anomalies caused by various distortions of transmitted values in the information exchange of protective automation. The operation of maximum current protection against short circuits in a digital substation was selected as the object of study. To obtain data for machine learning, mathematical modeling of the current values of a three-phase transformer arising from short circuits was used. The optimal confi guration of an artifi cial neural network for recognizing the operating modes of short-circuit protective automation, as well as the conditions for its training, were determined. Using a trained artifi cial neural network model, computational experiments were carried out with various types of distortions in the transmitted values of protective automation. As a result of the computational experiments, it was shown that, using the deviation of the values calculated by the trained artifi cial neural network from the values of the protective automation signals, it is possible to identify anomalies in the transmitted valuesof the digital substation protection mechanisms. Keywords: artifi cial neural network, data distortion, anomalies of values, protective automation, digital substation. Pp. 36-43.