Development of a Neural Network Method for Identification of the Method to Detect Slag Flow in a Metallurgical Unit "Steel Ladle-Intermediate Ladle"

The paper presents the results of research on the development of an algorithm for identifying the moment of the beginning of slag lapse during the functioning of the metallurgical unit "steel ladle-intermediate ladle". In the paper were performed works on the formation of datasets of the vibration acceleration signal at the functioning technological object. After that, experiments on amplitude and frequency analysis of the signal were performed in order to form a criterion for determining the moment of the beginning of slag flow. Then the possibility of application of autoencoder artificial neural network for identification of metallurgical unit state was tested.