Motor bearing failures detection by using vibration data

The use of methodologies for condition monitoring of rotating machines has been growing to reduce unplanned downtime and to increase the reliability of the industrial processes. The companies must select a correct maintenance strategy to follow the evolution of rotating machines. Condition monitoring is the collection of data related to the health status of the machine and it has been widely studied so far. Different methodologies have been developed to identify specific behaviors in the condition of induction motors. This paper proposes a methodology for bearing failure detection by using vibrations data, based on the frequency spectrum applied to induction motors. This methodology allows the use of vibration data obtained from motor bearings to establish their condition and therefore determine the type of damage to the bearing. The effectiveness of the proposed methodology is validated using a data set obtained from NASA (National Aeronautics and Space Administration). The results showed that this type of approach is very useful for analyzing bearings and in this way creating maintenance routes based on the condition of the electric machines.