Konstantinos N. Gyftakis

A Novel Approach for Broken Bar Fault Diagnosis in Induction Motors Through Torque Monitoring

The cracked or broken bar fault constitutes about 5-10% of total induction motor failures and leads to malfunction as well as reduction of the motors life cycle. This is the reason why there is continuous research on techniques for prompt detection. In this study, a study of the influences of the broken bar fault to the electromagnetic characteristics of the induction motor is presented, using an asynchronous cage motor and finite element method analysis. To this direction, two models have been created and studied: a healthy and one with a broken bar. Additionally, a new approach on the detection of the broken rotor bar fault through the electromagnetic torque monitoring is suggested and validated through experimental results.

The Zero-Sequence Current as a Generalized Diagnostic Mean in Δ-Connected Three-Phase Induction Motors

In this paper, a new diagnostic mean is introduced for accurate, reliable, and effective diagnosis of faults in delta-connected induction motors. This mean is the zero-sequence stator current spectrum. The study is carried out with analytical calculations and finite-element method simulations. Most common faults are examined for a 4-kW cage induction motor. The analysis will reveal that the proposed mean is able to successfully identify the static eccentricity fault, even in induction motors which produce principal slot harmonics, the broken bar fault, the unbalanced voltage supply, and the stator interturn short-circuit. It will also be shown that through the proposed method, it is possible to distinguish the interturn short-circuit from the unbalanced voltage supply.

A Novel and Effective Method of Static Eccentricity Diagnosis in Three-Phase PSH Induction Motors

Eccentricity-related faults are very common in induction motors and that is the reason for the extensive and worldwide research on their prompt diagnosis during the past decades. Previous works have shown that traditional diagnostic methods cannot offer reliable results for principal slot harmonics (PSH) induction motors. Moreover, most works were focused in the detection of dynamic and mixed, rather than static eccentricity. This paper is focused on the detection of the static eccentricity fault and its severity in PSH induction motors. The proposed method is original and proves to be reliable for the static eccentricity fault detection, even when the severity of the fault is low. The work was carried out with both finite element method (FEM) simulations and experimental testing.

Power Spectrum-Based Detection of Induction Motor Rotor Faults for Immunity to False Alarms

It has recently been shown that spectrum analysis of instantaneous power can provide sensitive online detection of rotor faults in induction motors compared with current, torque, speed, or vibration spectrum analysis. However, it was reported that monitoring of the twice slip frequency, 2sfs, components induced by the fundamental component can produce false rotor fault alarms due to asymmetry in the rotor or low frequency load oscillations. In this paper, the rotor fault components induced in the power spectrum by the stator fifth and seventh space harmonics are derived to evaluate their immunity to false alarms. It is shown that the (6 - 8s)fs component can provide reliable detection of rotor faults under cases where existing methods produce false alarms. An experimental study performed on custom built rotor samples shows that the new components are capable of detecting rotor faults immune to false alarms produced by rotor axial air ducts, rotor anisotropy, and low frequency load oscillations for cases where existing methods fail. The components derived in this paper can also be applied to vibration, speed, torque, or acoustic monitoring for reliable detection of rotor faults.

Comparative Experimental Investigation of Broken Bar Fault Detectability in Induction Motors

It has been shown in the past that the zero-sequence current spectrum can be reliably used to detect broken bar faults in induction motors. Previous work was carried out with extensive FEM analysis. Although it allows detailed study of spatial and time-dependent electromagnetic characteristics of induction motors, FEM is a heavily time-consuming tool and this limits full study. So, in this work, extensive experimental testing has been performed to validate the zero sequence current spectrum for detecting rotor asymmetries. Three identical induction motors have been used: one healthy, one with a broken rotor bar, and one with two broken rotor bars. The motors were tested under different voltage supply levels and with different mechanical loads. The zero-sequence current spectrum was calculated after measuring the three phase currents. It is for the first time experimentally shown that this approach offers greater diagnostic potential than traditional MCSA.

Dielectric Characteristics of Electric Vehicle Traction Motor Winding Insulation Under Thermal Aging

The electric motor is the heart of the electric vehicle. It is crucial that any occurring faults are detected promptly so that a catastrophic failure is avoided. At the same time, deep knowledge of the degradation mechanisms is required to allow maximum performance at minimum cost. This paper focuses on this balance. Statistical results from measurements of unaged and accelerated aged winding insulation samples provide information about the degradation processes, enabling steps toward a reliable prognosis model of the motors remaining life.

Comparative experimental investigation of broken bar fault detectability in induction motors

It has been shown in the past that the zero-sequence current spectrum can be reliably used to detect broken bar faults in induction motors. Previous work was carried out with extensive FEM analysis. Although it allows detailed study of spatial and time-dependent electromagnetic characteristics of induction motors, FEM is a heavily time-consuming tool and this limits full study. So, in this work, extensive experimental testing has been performed to validate the zero sequence current spectrum for detecting rotor asymmetries. Three identical induction motors have been used: one healthy, one with a broken rotor bar, and one with two broken rotor bars. The motors were tested under different voltage supply levels and with different mechanical loads. The zero-sequence current spectrum was calculated after measuring the three phase currents. It is for the first time experimentally shown that this approach offers greater diagnostic potential than traditional MCSA.

The Impact of Thermal Degradation on Properties of Electrical Machine Winding Insulation Material

Interturn stator short circuits can develop quickly leading to serious damage of an electric machine. However, degradation mechanisms of winding insulation material are not yet fully understood. Therefore, the main contribution of this paper is the analysis of the impact of thermal ageing on the electrical properties of the thin-film winding insulation. The insulation samples have been aged thermally at 200 °C-275 °C and for 100-1600 h. After ageing, impedance spectroscopy measurements were undertaken on the samples and equivalent-circuit model (ECM) parameters fitted for each measurement. This allows the impact of thermal ageing on ECM parameters to be analyzed, giving insight into the changes in the electrical properties of the insulation. Finally, high voltage was applied to the samples aiming to identify the breakdown-voltage (BV) characteristics of the insulation material.

Evaluation of the broken bar fault detectability depending on the rotor bar number

This work aims to evaluate the diagnostic ability of different means for the identification of the broken bar fault in induction motors with different rotor bar numbers. The stator is the same for all cases. The investigation is carried out with transient FEM simulations under constant speed.

Broken bar fault diagnosis in single and double cage induction motors fed by asymmetrical voltage supply

The diagnosis of faults in induction motors is a crucial scientific subject since it affects the industrial production functionality and reliability. Lately, a new scientific area is born: the electrical machines prognostics. So, aiming to this direction the complete knowledge of all abnormal parameters during the induction motor operation should be identified. In this paper, a methodology is described in order to identify not only if there is a broken bar fault, but also if the induction motor voltage supply is asymmetrical at the same time. The investigation is carried out with transient FEM simulations. The proposed methodology comes to supplement the traditional MCSA method with desired satisfaction.