Induction motor windings faults detection using flux-error based MRAS estimators

Abstract

The paper is concerned with detection of a stator and rotor winding faults in a squirrel-cage induction motor. The idea of the fault detection is based on a hypothesis that each of windings faults results in a sharp increase or decrease of internal parameters’ values of the machine, therefore it can be treated as a suitable fault symptom. Resistances of the stator and rotor windings seem to be adequate quantities due to their direct relationship with the machine windings. An observation and analysis of the parameters’ changes in a realtime domain enables to an incipient detection of the fault. It is evident that internal parameters of the machine can’t be measured directly during operation on the drive system thus the only way is an estimation by specialized algorithms. In the paper two estimators based on Model Reference Adaptive System (MRAS) were utilized to achieve this goal. Two simple algorithms for faults detection are proposed as well. Detailed description of fault detection systems is included in the paper. Proposed systems were tested on computer simulations performed by MATLAB/Simulink software. Then, experimental tests were carried out on the laboratory setup to confirm usefulness of proposed approaches.