Savelii Salnikov - Data Science (FB5) Semesterprojekt

Development of a software complex for sensorless diagnostic of induction motor

The induction motor is the most popular and widespread type of electric motor today. It has found its application in various types of production, including oil, metallurgical, machine-building fields. This type of motor is so widespread due to its simple device and relatively high reliability. But, like any other type of machine, the induction motor is also subject to various breakdowns. The most common are inter-turn short circuits of the stator windings (such a state of the windings, in which the motor can still function, but already started an irreversible mechanism of winding destruction), as well as various failures associated with the motor rotor bearings. Now the asynchronous motor diagnostics is performed with the help of a sufficiently large number of measuring sensors (vibration sensors, temperature sensors, torque sensors, rotor speed). It is desirable to use the smallest number of sensors, while maintaining measurement accuracy. Therefore, this project is aimed at developing a diagnostic system for asynchronous motor, based on an intelligent analysis of the motor current. Which can be obtained using only two current sensors. The result is a classifier capable not only of determining the presence of a fault, but also its type (bearings or stator winding).