An analysis of induction motor incipient-failure prediction and the development of systems concepts for electrical diagnostics in mining industry applications.

摘要

The safe and economical winning of mineral resources is increasingly dependent upon electric machinery. A failure within the machine's electrical system, either in the control or the power circuits, usually results in significant production losses. In the case of control circuit failures, on-board diagnostics are needed to quickly identify the failed component. However, in the power circuit, specifically the electric motors, a predictive capability is needed to schedule the deteriorating component for replacement. An objective of this work was to develop system concepts for such a diagnostic system.; Traditional motor insulation tests are unable to detect all failure modes and require that the motor under test be taken off line, both of which present serious practical constraints. Thus, there is considerable interest in a technique which can overcome one or both of these shortcomings. A preliminary examination of concepts for detecting incipient failure in ac motors revealed the existence of on-line techniques that had the potential to overcome the limitations of traditional ones. However, the realization of these techniques was impossible due to gaps in enabling technology.; Specifically, issues related to the deterioration processes within the stator winding were not understood. Another significant unknown was the affect of monitoring a motor that was connected to an unbalanced voltage bus. Attaining a theoretical and practical understanding of these issues was the major thrust of this research.; The major research issues were investigated using analytical and experimental methods. A computer model was used to investigate motor behavior while the stator winding was undergoing deterioration. Laboratory investigations were conducted to study deteriorating motors in an effort to understand and then model the salient behavior. These investigations included both destructive testing and simulation of deterioration within several motors.; Viable systems concepts for machine and motor diagnostics have been developed. The need to detect turn and coil insulation failures has been demonstrated. A mathematical technique to correct for unbalanced bus voltages has been developed and an alternative fault-detection predictor has been suggested. These results comprise the remaining enabling technology base needed to engineer and construct practical devices for industrial use.