Rotor Winding Fault Detection in the Wound Rotor Induction Machines Based on Magnetic Flux Distortion

Abstract

Induction machines are one the most important electrical equipment in industrial plants and manufacturers. They have a great variety of features, structures and performance. So, proper and secure protection design and fault detection are important issues. The induction machines are divided into two categories as squirrel-cage and wound-rotor types. Due to many advantages of squirrel cage motors (e.g. simplicity, lower cost, less maintenance and etc.) they are the most common choice for industries. However, wound rotor induction motors have their own applications, where they have high start-up torque for industries such as hoist, cement and steel ones. According to the fault destructive effects on induction motors, fault detection and the related region identification in early stages are the most important protection criteria. In this thesis, monitoring of the air-gap flux and measuring the rotational magnetic field distortion and asymmetricity caused by stator or rotor Turnto-Turn fault is regarded as the main parameter to detect such faults. To achieve such aim, some simple flux sensors are installed in the specified stator slots, so that each phase equipped by two flux sensors aligned with the magnetic axis of the related coils. In motor healthy condition, passing fluxes from the corresponding sensors are equal, which results in the identical induced voltages in them and then difference of them (i.e. Δ ) is equal to zero. In case of stator or rotor winding Turn-to-Turn fault, Δ will no longer be zero due to rotational magnetic field distortion and such phenomenon is the main criterion to detect the faulty condition and discriminate the stator or rotor winding faults, in this thesis. To demonstrate the capability of this method, not only the simulation results but also the experimental test results have been used. The ability to detect and discriminate rotor and stator winding Turn-to-Turn fault, faulty phase detection in the cases of stator faults, low processing burden and high sensitivity in the face of permissible voltage unbalance and mechanical load variation are the undeniable advantages of this method, which can be pointed out.