Condition monitoring of induction motors based on measurement and analysis of air gap flux spatial harmonics

Condition monitoring of induction motors is of much importance for their efficient and reliable operation. Most of the existing electrical techniques of condition monitoring and fault diagnosis are based on current, voltage, vibration and flux measurements external to the motor. Among these, most of them are based on motor stator current analysis. The most popular and well researched group of techniques are based on the motor current signature analysis (MCSA), which reveals motor condition through the changes of the stator current frequency spectrum. Although very powerful, MCSA has its shortcomings, including the need for long sampling times in steady state to achieve acceptable spectral resolution, dependence of successful diagnosis of some faults on the accurate knowledge of the motor slip, and potential ambiguity of interpretations. Besides, the characteristics of external signals such as current, leakage flux etc. are produced in the air gap flux density, which is a function of time and space. It is very information-rich, and contains ample indications of any imbalance or fault once it occurs. When transformed into a signal available for external measurements, this information is greatly reduced and distorted, as well as it loses space dimension. This thesis attempts to address the condition monitoring challenges by removing the limitations of a typical external sensor instrumentation. It is proposed to instrument a motor with an array of miniature Hall Effect Flux Sensors (HEFS) to measure the air gap flux density at multiple points around the air gap circumference. Internal instrumentation with an array of flux sensors is chosen in this thesis, because it allows to directly measure the air gap flux as a function of time and space, thereby conserving the space domain information to the benefit of condition monitoring. It also does not interfere with the motor operation and unlike search coils (which measure flux derivative), HEFS measures flux and their signal is less susceptible to noise distortions, which is particularly important for inverter driven motors. One important point to note is the additional cost due to sensor instrumentation. However, for production-critical large motors used in industries such as mining, petrochemical etc. are produced in fewer numbers, which are expensive to replace in case of a fault or failure. Also, the down time cost of these motors in terms of production loss is much higher than the extra instrumentation cost. Besides, instrumentation cost can be reduced if it is integrated in the manufacturing process of the motor. Additional functionality of the proposed instrumentation is described in the thesis. The resulting certainty about each motor condition at any point in time significantly reduces the risk of unplanned downtime and offsets the instrumentation cost by a large margin. Fundamental theory behind the air gap flux density as a function of time and space, and its distortions introduced by various faults, is analyzed in this thesis. Based on this theory, the thesis proposes a comprehensive condition monitoring approach based on the analysis of internal flux measurement to diagnose induction motor faults including stator turn-to-turn shorts, rotor bar damage, static and dynamic eccentricity, which may be present individually or at the same time. Moreover, the proposed approach is able not only to detect each fault at its early stage but also to determine its location and severity where applicable. The principles proposed in this thesis are illustrated by extensive simulations and are experimentally validated on a prototype online condition monitoring system based on National Instruments real-time platform.