Model Based Stator Fault Detection in Induction Motors

Abstract

A new strategy for stator fault diagnosis in induction motors is presented in this paper. The proposed strategy is based on the generation of a vector of specific residual using a state observer. The residual allows the detection of the faulted motor phase and the quantification of the number of shortcircuited turns. This strategy is robust to motor load variation and power supply perturbations.

Online Model-Based Stator-Fault Detection and Identification in Induction Motors

“…1. In [21], a new dynamic IM model was proposed. Different from the one presented in [12], this model allows consideration of interturn short circuits in any motor phase by means of a vectorial fault factor.…”

“…Artificial neural networks [18] and neuro-fuzzy techniques [19] are used to automatically diagnose stator faults from current and voltage measurement, mainly in online condition-monitoring systems. For online diagnosis, observer-based [20], [21] and parameter-estimation techniques [22] have also been recently presented.…”

“…Since an early fault detection is the main objective of statorfault detection schemes, a model-based strategy for online stator-fault detection is presented in this paper. This strategy, previously proposed in [21], uses a state observer to obtain a vector of specific residuals, which allows for a rapid detection of incipient stator faults. The use of a vectorial residual makes the fault detection independent of the phase in which the fault occurs.…”

“…In this paper, the development of the strategy proposed in [21] is improved and extensively analyzed. A different severity factor is defined, and its ability to detect and diagnose very incipient faults is explored and tested by several experiments.…”

Transient modeling and analysis of induction motors with position effects in stator turn faults

“…The short-circuit may occur anywhere in the distributed winding, as shown in Fig 1. This paper also presents the derivation of the steady-state phasor model based on the transient model, suitable for analytical studies. The transient and steady-state models are useful for diagnostics [11], [12], [13], [14], as well as for load emulation [15].…”