Sensitivity of the Rolling Bearings Diagnostic Method Depending on the Number of Measuring Points of a Multipoint Temperature Sensor

This work is aimed at research in the field of developing methods for the diagnosis of rolling bearings. This paper evaluates the influence of a number of measuring points of a multipoint temperature sensor on the performance of a rolling bearing diagnostic method. The study considers a rolling bearing diagnosis method based on analyzing a bearing temperature field and applying a neural network with an “autoencoder” architecture. The research of efficiency of the method is made on the temperature field data of the rolling bearing, which is a part of the technical equipment of the metal rolling line. As a result of this work, the hypothesis that the efficiency of the diagnostic model increases with the increase in the number of measuring points was confirmed.