Revolutionizing Motor Maintenance A Comp

International Journal of Power Electronics and Drive Systems (IJPEDS)
Vol. 15, No. 3, September 2024, pp. 1968~1989

ISSN: 2088-8694, DOI: 10.11591/ijpeds.v15.i3.pp1968-1989  1968
Revolutionizing motor maintenance: a comprehensive survey of

state-of-the-art fault detection in three-phase induction motors
Bahgat Hafez Bahgat1, Enas A. Elhay1, Tole Sutikno2,3, Mahmoud M. Elkholy1

1
Electrical Power and Machines Engineering Department, Faculty of Engineering, Zagazig University, Zagazig, Egypt
2
Master Program of Electrical Engineering, Faculty of Industrial Technology, Universitas Ahmad Dahlan, Yogyakarta, Indonesia
3
Embedded Systems and Power Electronics Research Group, Yogyakarta, Indonesia
Article Info ABSTRACT

Article history: This comprehensive review delves into electrical machine fault diagnosis
techniques, with a particular emphasis on three-phase induction motors. It
Received Nov 11, 2023 covers a variety of faults, including eccentricity, broken rotor bars,
Revised Apr 13, 2024 and bearing faults. It also covers techniques like motor current signature
Accepted Jun 4, 2024 analysis (MCSA), partial discharge testing, and artificial intelligence (AI)-
based approaches. This review focuses on fault diagnosis techniques for
electrical machines, specifically eccentricity faults, squirrel cage rotor faults,
Keywords: and bearing faults. It discusses their efficacy, applications, and limitations, as
well as the role of AI and neural network techniques in modern fault detection
Artificial intelligence applications. The review covers not only eccentricity faults, but also stator or
Fault detection armature faults caused by insulation failure, as well as bearing faults classified
Fault diagnostics as ball, train, outer, and inner races. It focuses on early detection to ensure
Induction motors optimal machine performance and reliability, while also providing insights
Motor maintenance into fault detection mechanisms. Modern ways of finding problems with
Neural network machines, like non-negative matrix factorization, rectified stator current
analysis, incremental broad learning, and AI-based methods, make machines
work better and stop money from being lost. The review is a valuable resource
for practitioners and researchers in the field, allowing them to make better
decisions about maintenance strategies and increase machine efficiency.
This is an open access article under the CC BY-SA license.
Corresponding Author:
Bahgat Hafez Bahgat
Electrical Power and Machines Engineering Department, Faculty of Engineering, Zagazig University
Shaibet an Nakareyah, Zagazig 2, Al-Sharqia Governorate 7120001, Egypt
Email: b.hafez020@eng.zu.edu.eg
1. INTRODUCTION
The increasing complexity of electrical machines and their growing integration into critical
infrastructure have heightened the demand for robust fault diagnosis techniques to ensure operational
reliability, safety, and efficiency. Overcurrent, overvoltage, and earth-fault protection have been used as
standard safety measures for decades. However, faults that are not found can disrupt operations, cause schedule
delays, and result in large financial losses, so we need more proactive and accurate fault detection methods.
The significance of three-phase induction motors in various industries lies in their high motive power
generation, durability, and low maintenance costs. Approximately 90% of industrial equipment worldwide uses
these motors as the prime mover, highlighting their critical role in industrial operations [1]–[4]. They are known
for their robust nature and reliable performance, making them indispensable for industrial processes.
Additionally, a wide range of industrial applications favor these motors for their ability to provide flexible
production control and soft motor start-up, making them a versatile and efficient solution.
Journal homepage: http://ijpeds.iaescore.com

Int J Pow Elec & Dri Syst ISSN: 2088-8694  1969
Electrical machine faults can be roughly grouped into five areas: problems with the stator, bad
connections between the stator windings [5]–[9], changes in the air gap that are either dynamic or static, shorted
rotor field windings, and bearing and gearbox failures. These faults can manifest in various forms, including
overheating, vibration, noise, and changes in motor current signatures. Early detection and accurate
identification of these faults are crucial for implementing timely maintenance strategies, preventing catastrophic
failures, and ensuring optimal machine performance. The challenge of fault detection in electrical machines has
led to the development of numerous diagnostic techniques. We can broadly classify these techniques into two
categories: non-invasive and semi-invasive methods. Non-invasive techniques do not require direct contact with
the machine, relying instead on external measurements such as temperature, vibration, acoustic noise, and radio
frequency (RF) emissions. Examples of non-invasive techniques include temperature measurement, infrared
thermography, vibration analysis, acoustic noise analysis, and RF emission monitoring [6]. Semi-invasive
techniques, on the other hand, require some level of physical access to the machine, such as partial disassembly
or the insertion of probes. Examples of semi-invasive techniques include motor current signature analysis
(MCSA), online partial discharge (PD) testing, and axial flux component analysis [9].
In recent years, advanced diagnostic approaches based on artificial intelligence (AI) and neural
network techniques have emerged as promising tools for fault detection in electrical machines [10]–[15] These
methods use machine learning algorithms to look at complicated patterns in motor data. This allows them to
find small problems that may indicate a problem more accurately and sensitively than other methods. AI-based
fault diagnosis techniques are particularly well-suited for large datasets of motor data, allowing them to
continuously monitor machine health and detect trends that may lead to future faults. This comprehensive
review article provides a detailed overview of the diverse range of fault diagnosis techniques for electrical
machines, examining their effectiveness in detecting and classifying various types of faults. We will delve into
the principles, applications, and limitations of each technique, highlighting the advantages and disadvantages
of each approach. We will also talk in depth about the growing role of AI and neural network techniques in
modern fault diagnosis applications. We will look at how they can improve the accuracy of fault detection,
give us more information about what might go wrong, and help us come up with proactive maintenance plans.
Through this comprehensive review, we aim to provide a valuable resource for engineers, researchers, and
practitioners involved in the design, operation, and maintenance of electrical machines. By understanding the
strengths and limitations of each fault diagnosis technique and the potential of AI-based approaches, we can
make informed decisions about selecting the most appropriate methods for specific applications and achieving
optimal machine reliability and performance.
The review article revolves around the comprehensive exploration of various fault types and the
corresponding detection methods employed in three-phase induction motors. The article delves into the adverse
impact of eccentricity faults on motor efficiency, with a focus on the use of MCSA for static and dynamic
eccentricity detection. The article also carefully talks about faults in the stator or armature that happen because of
insulation failure. It does this in a number of ways, such as using PD tests, axial flux component analysis, and
statistical process control. Furthermore, the article delves into the intricate domain of bearing faults, classifying
them into ball defects, train defects, outer bearing race defects, and inner bearing race defects, and provides a
comprehensive overview of advanced fault detection methods for three-phase induction motors that incorporate
artificial intelligence-based approaches. The review article conducted a comprehensive exploration into various
fault types and the corresponding detection methods employed in induction motors. Their research looked at how
eccentricity faults hurt motor efficiency, how to find broken rotor bars using MCSA and sideband component
analysis, and how to fix stator or armature faults caused by insulation failure in a number of different ways.
Additionally, the contributors delved into the intricate domain of bearing faults, categorizing them into distinct
categories such as ball defects, train defects, outer bearing race defects, and inner bearing race defects. They
showcased a comprehensive overview of advanced fault detection methods for three-phase induction motors,
incorporating artificial intelligence-based approaches. This collaborative effort consolidated a wealth of insights
into fault types and their detection mechanisms, providing a valuable resource for practitioners, researchers, and
enthusiasts in the realm of induction motors. The review article conducted an in-depth exploration into various
fault types and the corresponding detection methods employed in induction motors. Their findings encompassed
a comprehensive understanding of fault types such as eccentricity faults, squirrel cage rotor faults (broken bars
and end rings), and bearing faults. They delved into the causes, detection techniques, and effects of each fault
type, providing valuable insights for practitioners, researchers, and enthusiasts in the realm of induction motors.
Furthermore, the authors presented cutting-edge methods for finding faults in three-phase induction motors that
use artificial intelligence and up-to-date diagnostic techniques.
2. THE COMPREHENSIVE THEORETICAL BASIS OF IM FAULTS

Figure 1 serves as a visual compass, offering a comprehensive overview of the myriad fault types and
the diverse array of detection methods applied in the intricate realm of three-phase induction motors. These
Revolutionizing motor maintenance: a comprehensive survey of … (Bahgat Hafez Bahgat)
1970  ISSN: 2088-8694
faults have ramifications that extend beyond mere operational glitches, potentially culminating in both reduced
efficiency and the specter of catastrophic failure. This review aims to highlight a specific facet of this complex
landscape, focusing on internal mechanical faults and the cutting-edge realm of AI-based methods meticulously
designed for fault detection.
As we navigate through the complexities of AI-infused fault detection, our journey will delve into not
only the merits and advantages of these methodologies, but also candidly address their limitations and the
challenges that lie ahead. The following sections of this review article will serve as an exploratory voyage,
shedding light on the nuanced landscape of internal mechanical fault detection with a keen eye on AI-based
techniques. Furthermore, we will cast our gaze toward the future, outlining potential directions for
advancements and innovations in the realm of fault detection methods.
Induction Motor Faults
Internal Faults External Faults
Electrical Mechanical Electrical Mechanical Environmental
• • Magnetic circuit •
Eccentricity • Earth Faults • Crawling Temperature
• Faults •
Broken rotor bar • Overload • Transient voltage Cleanliness
• • Inter-turn short •
Bearing faults` • Pulsing Load • Unbalance supply Foundation defects
• circuit
Stator winding • Poor mounting voltage or current
• Mass unbalance • Dielectric failure
Figure 1. Classification of the most frequent induction motor faults [16]
The expansive landscape of three-phase induction motors unfolds within the purview of this reviewed
article, where a comprehensive exploration ensues into a spectrum of faults that can disrupt their seamless
operation. These include the elusive eccentricity, intricacies within the squirrel cage rotor, potential stator or
armature aberrations, and the critical domain of bearing faults. Table 1, a trove of insights, serves as a structured
repository that meticulously dissects the causes, detection techniques, and ripple effects each fault can induce.
The intricate dance of eccentricity, the subtleties within the squirrel cage rotor, and the vulnerabilities
embedded in the stator or armature find their place under the scrutiny of this article's lens. Delving into the root
causes, the article unravels the mechanisms and intricacies involved in the detection of these faults, ensuring a
comprehensive understanding of their manifestations. Dedicated attention is paid to bearing faults, which are
essential to the operation of induction motors. Table 1 shows the complex landscape of their causes, how to
find them, and the effects they have on motor performance. As we traverse the contents of Table 1, a mosaic
of insights emerges, offering not only a panoramic view of fault intricacies but also a practical guide for
engineers, researchers, and enthusiasts navigating the labyrinth of three-phase induction motor faults. This
review article, through its meticulous exploration and presentation, endeavors to contribute to the collective
knowledge base, fostering a deeper comprehension of the challenges and solutions within this dynamic realm.
2.1. Eccentricity faults

Diving into the wealth of insights provided by studies conducted by the IEEE-industry
applications (IAS), the intricate landscape of failure mechanisms in induction machines comes to light. These
studies reveal a breakdown of common failure types, with stator faults accounting for 38%, rotor defects at
10%, bearing defects taking a significant share of 40%, and the remaining 12% attributed to various other
defects [9]. This comprehensive breakdown serves as a valuable benchmark, providing a nuanced
understanding of the vulnerability points within these motors.
Eccentricity is a common mechanical problem that can happen in induction motors [11], [17]–[19]. It
can show up in static, dynamic, or mixed forms, as shown in Figure 2. Static eccentricity, a notable subtype,
finds its origins in the stator core's ovality or improper rotor and stator alignment during the commissioning
phase. On the other hand, dynamic eccentricity, another part of this fault, occurs when bearings wear out,
mechanical resonance occurs, or a bent rotor shaft causes rotational misalignment. The interplay of these
factors culminates in a nonuniform air gap, triggering instability in air-gap voltage and line current.
The consequences of eccentricity extend beyond the mechanical intricacies, directly impacting the
motor's efficiency. As eccentricity worsens, the average motor current and losses increase. This shows that
Int J Pow Elec & Dri Syst, Vol. 15, No. 3, September 2024: 1968-1989
there is a direct link between how bad the fault is and how badly it affects the motor's performance. This
phenomenon underscores the importance of vigilance in monitoring and addressing eccentricity in induction
motors, as it plays a pivotal role in maintaining operational stability and efficiency. The insights gleaned from
these studies not only shed light on the prevalence of eccentricity but also emphasize its critical role in the
broader landscape of induction motor performance [20].
Eccentricity turns out to be a key factor in how failures show up in three-phase induction motors,
having a big effect on how well they work. The consequences of eccentricity extend beyond mere performance
degradation, as it induces vibration and uneven magnetic forces (UMF), thereby shortening the machine's
lifespan. The mechanical stress that eccentricity puts on different machine parts, especially making bearing
wear worse, makes its bad effects even worse [21].
In the complexities of real-world scenarios [22], a notable observation emerges: both static and
dynamic eccentricities often coexist within induction machines. Even in the nascent stages of machine life,
static eccentricity, an inherent byproduct of production and assembly processes, is prevalent. This results in a
consistent unbalanced magnetic pull (UMP) in a specific direction, laying the groundwork for potential issues
such as accelerated bearing wear or a misaligned rotor shaft. Simultaneously, these natural causes have the
potential to lead to dynamic eccentricity. Failure to fix these issues can lead to a severe machine breakdown,
such as a stator-to-rotor hub failure. MCSA emerges as a crucial diagnostic tool for identifying both static and
dynamic eccentricities. The (1) encapsulates the essence of MCSA, providing a framework for detecting
relevant frequency components [23].
𝑓ℎ = (𝑘 ∗ 𝑍2 ± 𝑛𝑑) ∗ 𝑓𝑟 ± 𝜐 ∗ 𝑓𝑠 (1)
Here, fh denotes the frequency of the shaft's lateral vibration, fr represents the rotor frequency, fs represents the
frequency of stator vibration, and nd denotes the frequency of the bearing's outer race defect. Z2 denotes the
bearing's characteristic frequency; k serves as a variable multiplier dependent on the bearing type and operating
conditions; and υ accounts for any misalignment between the shaft and bearing.
Various models, including the DFT and Modified Prony's Method, are instrumental in discerning
eccentricity faults. While DFT is user-friendly, its drawback lies in being time-consuming and ill-suited for
dynamic models. On the other hand, the Modified Prony's Method shows promise as a practical alternative,
accurately identifying fault characteristic frequencies based on the stator's power and current outcomes [24].
Table 1. IM faults, causes, detection techniques, and effects

Fault Cause Detection techniques Effect
Eccentricity Static eccentricity caused by MCSA, Vibration analysis, AI-based methods, Vibration and imbalance,
faults ovality of the stator core or Harmonic analysis at stator terminal voltages, increased stress on
improper rotor or stator NFM-IBL, Measurement of space and time bearings, shaft
alignment and dynamic dependencies of air gap flux misalignment, reduced
eccentricity resulting from performance, heat
bearing wear, mechanical generation, increased noise
resonance, or rotational
misalignment due to a bent rotor
shaft.
Squirrel cage Broken rotor bars caused by MCSA, Vibration analysis, AI-based methods, Increased vibration, altered
rotor faults electromagnetic forces, residual discrete Fourier transform (DFT), modified magnetic fields,
(broken bars tensions, thermal strains, Prony’s method, Online PD test methods, overheating, abnormal
and end rings) mechanical strains, statistical process control (SPC), rectified noise, reduced torque and
environmental stressors, and stator current analysis, artificial intelligence- power output, electrical
dynamic stresses. based stator winding fault estimation in three unbalance, increased
phase induction motor, asynchronous motor’s current and energy
fault detection using artificial neural network consumption
(ANN) and fuzzy logic methods.
Stator or Stator faults caused by insulation MCSA, vibration analysis, AI-based methods, Abnormal current flow,
armature faults failure, such as phase-to-ground online PD test methods, rectified stator current unbalanced magnetic
or phase-to-phase faults. analysis, artificial intelligence-based stator fields, motor overheating,
winding fault estimation in three phase voltage imbalance, reduced
induction motor, asynchronous motor fault motor efficiency, risk of
detection using ANN and fuzzy logic electrical hazards
methods.
Bearing faults Bearing failures due to fatigue, MCSA, vibration analysis, AI-based methods, Increased vibration,
improper lubrication, inadequate SPC, empirical mode decomposition (EMD), excessive noise, reduced
placement, contamination, or sample entropy (SampEn), hall effect flux efficiency.
corrosion. sensors.

1972  ISSN: 2088-8694
Figure 2. Rotor position effect on stator
2.2. Squirrel cage rotor faults: broken bars and end rings
Furthermore, delving into the complexities of rotor failures reveals a nuanced dichotomy between cast
and fabricated rotors. Cast rotors, renowned for their enhanced durability [25]–[27], have witnessed an
expanded domain of application, particularly in larger machines reaching power capacities up to 3000 kW,
courtesy of innovations like cast ducted rotors. In contrast, fabricated rotors find their niche in larger or
specialized application machines, delineating a tailored approach based on specific operational requirements.
Mechanics, electromagnetic forces, residual tensions, thermal strains, mechanical strains,
environmental stressors, and dynamic stresses all play a part in how rotor malfunctions happen, especially
when rotor bars and end rings break. Cast rotors, while celebrated for their durability, pose a unique challenge
when afflicted with faults such as cracked or fractured rotor bars, rendering them practically impossible to
repair [28]. Within the diagnostic realm, MCSA emerges as a key methodology for detecting broken rotor bars.
This technique relies on identifying sideband components, denoted as 'fb,' around the fundamental frequency, as
encapsulated in the (2). The lower sideband signifies the presence of broken rotor bars, while the upper sideband
is indicative of speed oscillations. The (3) further underscores the versatility of this approach by
allowing multiple sidebands to occur due to broken bars, each corresponding to different values of 'k' (1, 2, 3,
and so forth).
𝑓𝑏 = (1 ± 2𝑠) 𝑓 (2)
𝑓𝑏 = (1 ± 2𝑘𝑠) 𝑓 (3)
A notable advancement in this field is the integration of a soft sensor [29]–[32] designed to identify
broken rotor bars. This new idea uses a group learning method called dynamic weighted majority (DWM) and
combines drift detection with automatic structure evolution based on the entropy criterion. The membership
degree of the fuzzy classifier is important for figuring out entropy for drift detection. This makes for a
sophisticated and flexible way to find rotor faults [33]. It is important to note that uninsulated rotor cages,
which feature robust contact between the rotor core and bars, may pose challenges in accurately pinpointing
broken bars. To complement these advanced techniques, diagnostic tools such as harmonic analysis at stator
terminal voltages after motor shutdown offer additional insights into the intricacies of rotor health [34]. This
holistic exploration not only reveals the complexities of rotor failures but also underscores the imperative for
sophisticated diagnostic approaches to ensure the reliability and longevity of induction motors.
2.3. Stator or armature faults

When you look more closely at stator faults, you can see that they are very common and cause a lot
of induction motor failures [35]. Because of this, you need to fully understand them and be proactive about
diagnosing them. These faults, often stemming from insulation failure, unfold in the form of phase-to-ground
or phase-to-phase faults, gradually evolving from minor aberrations to major disruptions. The multifaceted
etiology of stator or armature insulation failure encompasses a variety of contributing factors, including but
not limited to short circuits during motor startup, environmental stressors, insufficient bracing of end windings,
compromised core lamination, leaks in the cooling system, electrical arcing, and elevated temperatures within
the stator core or winding [36].
The diagnostic arsenal deployed to identify these insidious faults is as diverse as the faults themselves.
While online PD test methods are reliable stalwarts, especially for large generators and motors with stator
windings rated at 4 kV and above, a conspicuous gap exists in standardized stator defect detection methods for
low-voltage motors. To bridge this gap, innovative techniques involve the analysis of the machine's axial flux
component through a strategically positioned large coil wound tightly around the shaft, as demonstrated by the
work of [37]. The (4) governs the installation of four symmetrically positioned coils in each of the motor's
quadrants, further refining the pinpointing of fault location.
𝑛(1−𝑠)
(𝑘 ± )𝑓 (4)
𝑝
Insightful research, as conducted by [10], has identified errors causing machine impedance asymmetry
leading to unbalanced phase currents, while [38] attributed such imbalances to negative sequence currents.
A modeling endeavor has meticulously mapped these imbalances, revealing a "bolted" flaw among 648 turns.
Building on these foundations, [39] proposes a diagnostic index based on the ratio of faulty to healthy positive
sequence current. In addition to these methods, techniques like SPC and signal processing [40] help find stator
problems more reliably. This shows how important it is to use a flexible and multifaceted approach to fault
diagnosis in the ever-changing world of induction motor operations. This thorough study not only sheds light
on the complexity of stator faults, but it also stresses how important advanced diagnostic methods are for
making sure that induction motors are reliable and last a long time in a wide range of operating conditions.
2.4. Bearing faults

In the intricate realm of electrical machines, the ubiquitous use of ball or rolling element bearings with
an inner and outer ring, housing a set of rolling components rotating within raceways, is a common design
paradigm [41]. Despite efforts to maintain load balance and alignment, the specter of fatigue failures looms,
potentially giving rise to heightened vibration and noise levels [42]. External factors such as improper
lubrication leading to abrasion and heating, suboptimal bearing placement resulting in raceway indentations,
and the insidious influence of contamination or corrosion by water and abrasive particles can precipitate these
often insidious and multifaceted failures. Such failures extend beyond the mechanical intricacies of the
bearings, as they have the potential to cause substantial damage to the overall motor system. The detection of
faults associated with bearings, a critical yet under-documented aspect, assumes paramount importance due to
their significant contribution to motor failures, accounting for a staggering 40–50% of reported cases. Ball
defects, train defects, outer bearing race defects, and inner bearing race defects systematically categorize the
intricate landscape of bearing-related faults, which manifest as rotor asymmetry, often symptomatic of
eccentricity.
Understanding the nuanced manifestations of these faults becomes imperative for maintaining the
operational integrity of electrical machines. Ball defects, which arise from anomalies within the rolling
components, can introduce irregularities in the rotational motion, resulting in a cascade of detrimental effects.
Train defects, which are characterized by issues throughout the entire set of rolling components, present a
broader challenge that necessitates comprehensive diagnostic strategies. Outer bearing race defects, which
occur in the outer ring raceway, and inner bearing race defects, which manifest in the inner ring raceway,
represent distinct fault categories, each with its own unique set of challenges and diagnostic considerations. As
the need for more comprehensive diagnostic methodologies in bearing fault detection becomes evident, it
underscores the urgency for further research and development in this domain. The intricate interplay between
mechanical factors, external influences, and fault manifestations necessitates a holistic approach to ensure the
reliable and efficient operation of electrical machines across a variety of applications and operating conditions.
3. METHODS FOR ADVANCED FAULT DETECTION

In this section, we present a MATLAB-created M-file that allows for the visualization and comparison
of various fault detection methods for induction motors. This tool streamlines the analysis process, enabling
researchers to assess the effectiveness of different techniques efficiently. By synthesizing findings from diverse
studies, it offers valuable insights into the most suitable IM fault detection method for specific industrial
applications. In the realm of advanced fault detection techniques for 3-phase induction motors, a diverse array
of methodologies emerges, each contributing to the evolving landscape of motor diagnostics. One such
approach entails the use of incremental broad learning and non-negative matrix factorization. This advanced
method combines the power of incremental broad learning with the insights gained from non-negative matrix
factorization in order to improve the ability to find faults. The fusion of these methodologies aims to
provide a comprehensive and nuanced understanding of motor faults, enabling more accurate and timely
detection [43], [44].
1974  ISSN: 2088-8694
Rectified stator current analysis stands as another noteworthy methodology in the arsenal of fault
detection techniques. By scrutinizing the stator current in a rectified manner, this approach provides unique
insights into the motor's health and performance. The rectification process allows for a more granular
examination of the current waveform, facilitating the identification of subtle deviations that may indicate
underlying faults. This method proves to be a valuable tool in the diagnostic toolkit for 3-phase induction
motors [45], [46]. The measurement of the air gap flux's space and time dependencies represents a cutting-edge
approach to fault detection. By analyzing the intricate interplay of spatial and temporal variations in the air gap
flux, this methodology unveils valuable information about the motor's condition. This holistic perspective
allows for the identification of faults that may manifest in complex ways, contributing to a more comprehensive
understanding of motor health [47], [48].
The three-phase induction motor utilizes artificial intelligence to estimate and predict faults in the
stator winding. The motor takes advantage of artificial intelligence's prowess to predict and estimate faults in
the stator winding. Leveraging advanced algorithms and machine learning techniques, this approach offers a
predictive framework for fault estimation that enables proactive maintenance and minimizes downtime [49].
ANN and fuzzy logic methods tackle fault detection for asynchronous motors. Intelligent systems, designed to
learn and adapt, offer a dynamic approach to fault detection in asynchronous motors. By combining ANN and
fuzzy logic, the diagnostic system can handle the complicated and changing nature of asynchronous motor
faults, which makes the system more reliable overall.
A comprehensive framework, encompassing feature extraction, broad learning, and incremental broad
learning, outlines a diagnostic methodology for 3-phase induction motor (TPIM) faults [50]. We train the system
using processed experimental data, highlighting its adaptability through retraining with incremental learning.
The study underscores the effectiveness of incremental broad learning and advocates for future research to focus
on improving feature extraction and developing automatic selection methods for incremental nodes. This
approach's versatility extends beyond TPIM faults, making it applicable to a range of motor diagnostic issues
with the potential for increased accuracy through the augmentation of feature nodes or inputs. Because they
focus on always getting better and being able to adapt, these advanced fault-finding methods are at the top of
the list for making sure that 3-phase induction motors work well and reliably in a wide range of situations.
3.1. Incremental broad learning and non-negative matrix factorization (NFM-IBL)

TPIMs stand as intricate machinery, subject to an array of potential faults arising from the dynamic
interplay of stator and rotor conditions. The complexity inherent in these motors necessitates a diagnostic
system that can swiftly and accurately respond to emerging issues [50], [51]. While machine learning-based
diagnostic systems have been developed for induction motors, their intricate design often entails prolonged
training times and necessitates complete retraining in the event of inaccuracies. To address the nuanced
challenges posed by TPIM faults, a novel approach is proposed, leveraging the adaptable incremental broad
learning (IBL) method. This method encompasses the integration of NMF-IBL, along with feature extraction
techniques employing empirical mode decomposition (EMD) and sample entropy (SampEn). The
amalgamation of these methodologies forms the backbone of the proposed TPIM fault diagnostic framework,
which comprises four essential sub-modules: data collection and processing, broad learning, incremental broad
learning, and structure simplification by non-negative matrix factorization (NMF) [52].
The diagnostic framework relies on the first sub-module, data collection and processing, which
systematically collects and processes relevant data related to the TPIM's operational conditions and
performance. This crucial step lays the groundwork for subsequent analysis and fault detection. The second
sub-module, Broad learning, involves the application of advanced learning algorithms to glean insights from
processed data. This phase aims to build a foundational understanding of the motor's normal behavior and
performance, serving as a reference point for anomaly detection. The third sub-module, incremental broad
learning, introduces adaptability into the diagnostic system. This incremental approach allows the system to
learn and adjust over time, avoiding the need for extensive retraining in the face of inaccuracies or evolving
motor conditions. This flexibility is particularly crucial in addressing the dynamic nature of TPIM faults. The
NMF's final sub-module, structure simplification, adds a layer of sophistication to the diagnostic framework.
We employ non-negative matrix factorization to simplify the complex data structure, enabling a more
streamlined and interpretable representation. This step enhances the diagnostic system's ability to identify and
isolate faults accurately.
The suggested TPIM fault diagnostic framework uses an all-around and flexible approach to find
faults. It is based on simulating the findings in [53] using the m-file shown in Figure 3. By integrating
innovative methodologies and leveraging adaptability through incremental learning, this framework aims to
overcome the challenges associated with TPIM faults. As the diagnostic landscape continues to evolve, the
emphasis on real-time responsiveness, accuracy, and adaptability positions this framework at the forefront of
ensuring the reliable and efficient operation of three-phase induction motors in diverse operational scenarios [54].
Figure 3. Feature extraction by IBL and NMF [53], [55]
3.1.1. Data acquisition

The proposed TPIM fault diagnostic system has a lot of moving parts. The data acquisition sub-module
is the most important part because it lets you learn more about how the motor works. The design of this sub-
module includes the detection and capture of four crucial signals: sound waves and the currents in windings A,
B, and C, represented as x1, x2, x3, and x4, respectively. However, due to the inherent involvement of only
two stator currents in TPIM operation, we strategically select only three signals (x1, x2, and x3). We judiciously
apply a band-limiting filter to further refine the data and mitigate interference, enhancing the accuracy and
reliability of the subsequent diagnostic processes.
The sound signal, denoted as x1, assumes a pivotal role in the diagnostic framework. In alignment
with best practices gleaned from studies [56], this signal is meticulously divided into distinct training,
validation, and test datasets. This segmentation is a very important step that helps train and test the diagnostic
system thoroughly, making sure that it can correctly identify the subtleties of TPIM faults. The judicious
selection and processing of signals within the data acquisition sub-module laid the groundwork for subsequent
phases of the diagnostic framework. This sub-module improves the accuracy and dependability of the TPIM
fault diagnostic system as a whole by focusing on important signals and taking steps to cut down on
interference. Paying close attention to signal processing based on previous research shows a dedication to a
method that is based on science and has been proven to work in the real world when trying to understand three-
phase induction motor faults [54].
3.1.2. Data Processing

When it comes to TPIM fault diagnosis, the EMD technique is one of the most important steps in
figuring out what the raw signals mean [57]. This technique yields three distinct datasets: xk-EMD-Train, xk-
EMD-Vali, and xk-EMD-Test. These datasets, each encapsulating unique insights derived from the raw signals,
lay the foundation for a more nuanced and granular understanding of the motor's operational dynamics. We
enlist the Sample Entropy (SampEn) statistical approach to distill pertinent information and eliminate
redundancy within these datasets. SampEn effectively extracts relevant features, focusing subsequent analysis
on the most discriminative and informative aspects of the signals. We then subject the resultant features to
normalization, a crucial step that ensures each feature contributes equally to the diagnostic process. We store
these normalized features as xk-SE-Train, xk-SE-Vali, and xk-SE-Test, respectively, encapsulating the refined
and normalized representations of the original signals.
Understanding the importance of domain knowledge (DK) in improving defect detection, we
strategically incorporate domain knowledge-informed attributes into the processed datasets. These DK
attributes serve as additional layers of information, enriching the datasets with insights that complement the
statistical features derived from EMD and SampEn. In the final stages of preprocessing, the datasets undergo
a renaming process, signifying the completion of the intricate signal processing journey. We now denote the
preprocessed datasets as xk-Proc-Train, xk-Proc-Vali, and xk-Proc-Test. This nomenclature reflects the
comprehensive nature of the datasets, encompassing the insights derived from EMD, SampEn, and domain
knowledge attributes. The preprocessed datasets are ready for deeper analysis, paving the way for the next
stages of the TPIM fault diagnostic framework. This careful and organized preprocessing pipeline, which is
based on tried-and-true methods and statistical approaches, shows how dedicated the team is to finding the
causes of TPIM faults with precision and accuracy. The diagnostic framework leverages the power of EMD,

1976  ISSN: 2088-8694
SampEn, and domain knowledge to establish a robust and comprehensive understanding of the underlying
signals for the subsequent analysis.
3.1.3. Board learning [54]

The most important part of the TPIM fault diagnostic framework is model training, which comes after
signal processing and preprocessing. A wide learning model, trained using the meticulously prepared xk-Proc-
Train dataset, serves as the initial foray into understanding and predicting the intricacies of motor faults. We
rigorously evaluate this model, scrutinizing its accuracy to determine whether it aligns with the predetermined
goal percentage. The evaluation of the wide learning model serves as a crucial checkpoint, ensuring that the
initial training phase meets the desired level of accuracy. However, if the model fails to meet the predetermined
goal percentage, we initiate a strategic pivot. The diagnostic system seamlessly transitions into the incremental
broad learning phase, an adaptive approach designed to enhance the model's performance dynamically.
In the incremental broad learning phase, a pivotal adjustment is made by increasing the number of
enhancement nodes. This augmentation is a strategic intervention aimed at refining the model's understanding
and predictive capabilities. The model gets a better understanding of the underlying patterns in the data by
adding more nodes. This lets it change and adapt to the changing nature of TPIM faults. This iterative and
adaptive process exemplifies the commitment to continuous improvement within the TPIM fault diagnostic
framework. The seamless transition from wide learning to incremental broad learning highlights the diagnostic
system's flexibility, ensuring that it can effectively navigate evolving motor conditions and fault
manifestations. With a goal-oriented approach and an adaptive learning strategy, the TPIM fault diagnostic
framework is a strong and flexible tool for making sure that three-phase induction motors work reliably and
efficiently in a wide range of operational situations.
3.1.4. Incremental broad learning [54]

Within the TPIM fault diagnostic framework, the IBL submodule assumes a crucial role in refining
the accuracy of the model. Using the information from the xk-Proc-Vali dataset, this submodule changes the
number of enhancement nodes on the fly. This is done on purpose to make the model better at making
predictions over time. However, the adaptive nature of the IBL submodule brings forth a challenge—the
potential for overfitting. If the number of nodes is excessively high, the model may start fitting the training
data too closely, leading to a reduction in its generalization capabilities. To circumvent this issue, a meticulous
trial-and-error process is instituted. This process is geared towards determining the optimal number of nodes,
denoted as N, striking a delicate balance between enhancing accuracy and preventing overfitting. The trial-
and-error methodology is a nuanced approach that involves systematically experimenting with different node
configurations until the desired validation accuracy is achieved. This iterative process is informed by insights
presented in a comprehensive review article, providing a theoretical and empirical foundation for optimizing
the IBL submodule.
The overarching goal is to harness the model's adaptability and dynamic learning capabilities without
compromising its ability to generalize to new data. The IBL submodule aligns with the broader objective of
the TPIM fault diagnostic framework by fine-tuning the number of enhancement nodes through a systematic
trial-and-error approach, resulting in a robust and accurate model that can effectively detect and predict motor
faults. The focus on avoiding overfitting, along with the evidence-based approach described in the review
article, shows that the TPIM fault diagnostic system is dedicated to accuracy and dependability. With its
iterative nature, the trial-and-error process encapsulates the adaptive spirit of the diagnostic framework, making
sure that it always changes to meet the complex and challenging needs of three-phase induction motor faults.
3.1.5. Structure simplification [53]

In order to optimize the TPIM fault diagnostic system, a critical post-processing phase comes into
play. The learning system, having undergone the adaptive and dynamic processes of wide learning and
incremental broad learning, may inadvertently harbor redundant nodes. To streamline and refine the model, a
systematic approach is employed to identify and eliminate these redundancies, ensuring the efficiency and
interpretability of the diagnostic system. Node reduction involves simplifying the learning system through low-
rank approximations, a technique aimed at retaining the essential features while discarding unnecessary
redundancies. This reduction not only enhances the computational efficiency of the model but also contributes
to a more interpretable and streamlined representation. We introduce non-negative matrix factorization (NMF)
as a compression mechanism to further optimize the IBL structure. NMF is a strong mathematical tool that
makes it easier to find meaningful patterns in the learning system. This leads to a clearer representation without
lowering the accuracy of the diagnosis [58].
Using NMF to compress the IBL structure is a smart way to cut down on system errors and make the
TPIM fault diagnostic system better at finding problems. However, the pursuit of optimal diagnostic accuracy
remains paramount. The incremental broad learning submodule initiates an additional layer of compression if
the diagnostic accuracy exceeds the predefined range of [TP 0.025]. This iterative compression process ensures
that the TPIM fault diagnostic system continually adapts to achieve the delicate balance between model
complexity and diagnostic precision. By using these compression techniques that are based on well-known
mathematical methods, the TPIM fault diagnostic system works in a way that is efficient, easy to understand,
and flexible. The iterative nature of compression, along with the strategic use of NMF and incremental broad
learning, makes the commitment to constant improvement in the diagnostic framework even stronger. This
careful post-processing step is an important part of making sure that the TPIM fault diagnostic system is a
strong, useful, and correct way to figure out what's wrong with a three-phase induction motor.
3.2. Rectified stator current

In the realm of fault detection for induction motors, stator current analysis [59] stands as a prominent
and widely employed technique. Renowned for its simplicity and the minimal hardware and software
prerequisites it demands, this non-invasive method has proven effective across various motor sizes and
operational conditions. However, the practical application of stator current analysis encounters challenges,
particularly when dealing with large induction motors operating at extremely low slip rates.
One of the primary challenges inherent in stator current analysis is the potential concealment of fault
harmonics by the fundamental component, particularly in situations where the motor operates at extremely low
slip. This concealment phenomenon poses a significant obstacle, often resulting in delayed fault detection until
the damage reaches a severe stage. In response to this limitation, a novel approach based on rectified motor
current analysis emerges as a pragmatic and effective solution. The proposed rectified motor current analysis
approach introduces a simple yet powerful strategy to overcome the challenges associated with fundamental
component leakage. By rectifying the motor current, the fault harmonics are distinctly revealed at a lower
frequency, untethered from the interference of the fundamental component. This innovative method not only
circumvents the delayed detection issue but also offers a versatile solution that can be seamlessly implemented
through both software and hardware.
The practical efficacy of the proposed approach is robustly demonstrated through experimental
verification [60], wherein the method proves instrumental in detecting a broken bars defect in a large induction
motor. This empirical validation underscores the real-world applicability and effectiveness of the rectified
motor current analysis, positioning it as a valuable addition to the arsenal of fault detection techniques for large
induction motors. Stator current analysis remains a stalwart in fault detection, the proposed rectified motor
current analysis emerges as a complementary and pragmatic approach tailored to address the nuances of large
induction motors operating at extremely low slip rates. Its simplicity, effectiveness, and demonstrated success
in real-world scenarios mark it as a noteworthy advancement in the pursuit of reliable and timely fault detection
strategies for induction motors.
3.3. Measuring the variations of air gap flux with respect to both space and time
Ensuring the efficient and dependable operation of induction motors hinges on robust condition
monitoring systems, a critical facet addressed in this section. Contemporary diagnostic systems predominantly
rely on external measurements such as current, voltage, vibration, or flux to glean insights into the
motor's health. However, this study advocates for a groundbreaking online fault diagnostic method, as proposed
in [52], [61]–[64], which introduces a paradigm shift by utilizing an array of Hall effect flux sensors to measure
the internal main air gap flux density of induction motors.
The crux of this method lies in its applicability to specialized motors with elevated reliability
requirements. Unlike traditional diagnostic approaches, the proposed method offers a comprehensive solution
for diagnosing induction motor failures by tapping into the internal dynamics of the motor's air gap flux density.
This internal measurement not only provides a unique perspective but also allows for the detection of issues at
an early stage, enabling timely intervention to prevent the escalation of faults.
A distinctive feature of the proposed method is its capability to not only detect faults but also identify
their precise location and gauge their severity [58]. This granular level of diagnostic information is instrumental
in formulating targeted maintenance strategies, enhancing the overall reliability and lifespan of the induction
motors. To substantiate the efficacy of the proposed method, the study undertakes extensive simulations and
develops a prototype online condition monitoring system based on the National Instruments real-time
platform [65]. The results of these simulations and the functioning prototype serve as empirical evidence,
confirming the effectiveness of the proposed method in real-world scenarios. This empirical validation
reinforces the viability of integrating Hall effect flux sensors for online fault diagnosis, establishing it as a
potent tool for ensuring the health and longevity of induction motors.
This section not only highlights the pivotal role of condition monitoring in induction motor operation
but also introduces a pioneering online fault diagnostic method that leverages Hall effect flux sensors. The
method's applicability to specialized motors, early fault detection capabilities, and the ability to pinpoint fault
1978  ISSN: 2088-8694
location and severity underscore its significance in advancing the field of induction motor diagnostics. The
combination of theoretical proposals, simulations, and a functional prototype serves as a robust foundation for
considering the proposed method as a valuable addition to the repertoire of condition monitoring techniques
for induction motors.
3.3.1. Faulty stator winding

Stator winding shorts pose a severe threat, with the potential to rapidly escalate into catastrophic
failures [60]. When a single turn of the stator winding is shorted, it becomes disconnected from other turns
within the same phase coil. This disconnection induces a shift in the phase magnetomotive force (MMF),
altering the magnetic field distribution. Simultaneously, the isolated turn carries a loop current that exceeds the
rated current, resulting in excessive heating. The swift identification and localization of stator winding
problems is imperative, as evidenced by an alarming experiment in which the temperature of a shorted turn
soared from 60 °C to 120 °C in just 6 seconds. This stark reality emphasizes the critical need for effective fault
diagnosis methods.
A groundbreaking study [61] proposes an innovative fault diagnosis method capable of detecting and
precisely locating stator winding problems. This method functions by measuring the distortion that stator shorts
cause directly within the main air gap flux. It provides a direct and efficient means of identifying faults in the
stator winding. Furthermore, the proposed method excels at pinpointing the exact location of the fault by
analyzing the rise in primary wave magnitude during rotation. This strategic approach not only mitigates the
uncertainties associated with time harmonic interpretation, but it also optimizes the use of extensive
instrumentation in investigative processes.
We wrote an m-file to simulate the experimental results [66] in Figure 4, which show a substantial
increase in the primary wave's magnitude during the short rotation. This remarkable rise acts as a reliable
indicator, allowing for precise fault location. The implications of this method go beyond just finding faults; it
provides a complete answer to the urgent need to quickly identify and pinpoint problems with stator windings,
which can prevent disastrous outcomes in induction motors.
Figure 4. Stator winding with a turn-to-turn short
3.3.2. Eccentricity, both static, and dynamic

This groundbreaking article [67] introduces a pioneering approach aimed at improving the
instrumentation critical for effective induction motor operation. The key innovation is the use of a Hall effect
sensor array strategically positioned around the stator circumference within the motor air gap. This strategic
placement enables the sensor array to capture vital information about the motor's internal dynamics and
performance. We dedicate our study to unraveling the fundamental theory underlying AC motor faults,
specifically focusing on detecting critical issues like stator winding defects, static and dynamic eccentricity,
and rotor bar faults [67].
The proposed approach extends beyond theoretical considerations, providing concrete algorithms
designed for the accurate detection of these various faults. Notably, the study employs a comprehensive
condition monitoring system, leveraging the capabilities of the National Instruments Compact RIO real-time
platform. This system has demonstrated exceptional fault sensitivity, noise immunity, and dynamic variation
tolerance, making it a robust tool for diagnosing and monitoring the health of induction motors.
This approach's empirical validation further solidifies its credibility. The study underscores that this
innovative methodology is particularly well-suited for dynamic applications involving inverter-fed large
induction motors with elevated reliability requirements. This approach's systematic development and validation
contribute significantly to the field of fault diagnosis in induction motors [7]. It not only enhances the
understanding of motor faults but also presents a practical and effective solution for real-world applications,
promising heightened reliability and performance in dynamic settings.
3.4. Estimation of stator winding fault in three-phase induction motors using artificial intelligence
The challenge of detecting faults in induction motors through multiple optimization techniques has
spurred innovative research. In an interesting study [68], researchers came up with a smart way to find short-
circuit faults in the stator winding of a three-phase induction motor by using fuzzy logic. This method is
different because it uses a fuzzy logic controller to carefully look at the stator current and give clues about the
type of motor failure based on a clear set of rules and an inference mechanism [69]–[71]. MATLAB-Simulink
software facilitates the implementation of this intelligent approach, enabling the simulation of both healthy and
short-circuit fault conditions in a three-phase induction motor, as illustrated in Figure 5. The fuzzy logic-based
approach has a unique benefit in that it can find faults based on the inputs that are available. This makes it an
intelligent and flexible way to keep an eye on motor condition and analyze faults.
Figure 5. Membership functions for both input and output variables
The significance of this approach goes beyond its technical prowess. By facilitating the detection of
faults at an earlier stage, it contributes to the establishment of a safer working environment in industries. The
proactive nature of fault detection aligns with the broader goal of ensuring the reliability and safety of induction
motors in industrial settings. This study not only introduces an intelligent methodology for fault detection but
also underscores its potential impact on industrial operations, emphasizing the importance of early fault
detection in maintaining a secure and efficient working environment.
To lessen the impact of load changes on defect detection, researchers propose artificial intelligence
(AI) algorithms for detecting interturn short circuit (ITSC) faults in three-phase induction motor stators, with
a focus on the use of ANN and fuzzy logic systems [72]. The ANN algorithm is capable of detecting and
locating ITSC faults, while the fuzzy method can diagnose the severity of ITSC defects. Simulation and
experimental results validate the effectiveness of both techniques under ITSC fault and load change conditions.
In order to achieve reliable stator defect detection even under load changes, a combination of ANN
and Fuzzy Logic System (FLS) is proposed. Specifically, this research [73] utilizes a feedforward multi-layer
perceptron (MLP) Neural Network trained with the back propagation (BP) algorithm to automatically detect
and locate ITSC faults. The suggested approach and neural network architecture are illustrated in Figure 5,
with the NN's output number set at three to correspond with the three phases of the induction motor where the
ITSC fault could potentially occur.
The selection of fault indicators is a crucial step in developing a monitoring and diagnostic system.
Induction motors often detect interturn short-circuit (ITSC) faults using the phase shifts between line current
and phase voltage, which offer a wealth of fault information. Healthy motors exhibit identical phase voltage
1980  ISSN: 2088-8694
and line current magnitudes, offset by 120 electrical degrees. However, faulty operations can result in changes
in magnitudes and phase shifts. We simulated the faulty stator model from Section II under different load
torques (T=3 and 7 N.m.) to investigate the behavior of three-phase shifts under various ITSC faults. The output
of the m-file that was used to simulate the three-phase shifts [74] is shown in Figure 6. It shows how the results
change with the number of short-circuited turns and a 5 N.m. load. The results reveal that load changes have
an impact on the ITSC fault detection technique.
This study [75] presents two AI-based techniques for robust stator failure detection in the presence of
load variations. Artificial neural networks and fuzzy logic are utilized to eliminate the need for an induction
motor model during fault detection, resulting in adaptable and easily deployable systems. The ANN technique
identifies the faulty phase, while the fuzzy logic-based detector determines the severity of the problem. These
smart techniques have various applications, including alerting the workforce of a dangerous condition and
facilitating the repair of a defective stator.
Figure 6. The characteristics of phase shift due to interturn short circuit (ITSC) faults are observed for phase
A, phase B, and phase C under a load torque of 5 N.m.
4. MODELLING APPLICATIONS
As the popularity of artificial intelligence-based condition monitoring systems rises, fault detection
systems [10], [31], [65], [73], [76], [77] utilizing support vector machine (SVM), ANN, nave bayes classifier,
ensemble, and K-nearest neighbors (KNN) have become more prevalent. These systems can detect faults and
determine their severity level, but they require extensive data to train and malfunctioning machines are scarce.
Numerical methods can simulate faulty conditions that are difficult to test in the field or lab, providing fault data
for machine learning algorithms. Accurate induction motor (IM) malfunctioning models can minimize harmful
testing, lower costs, and validate new fault detection techniques [78], [79], making them beneficial for training
and testing artificial intelligence-based condition monitoring systems. This section covers recent developments in
IM models, which are categorized into electrical circuits, magnetic circuits, numerical approaches, and hybrid
models, and provides an overview of various fault diagnosis methods [80] as shown in Figure 7.
The coupling circuit model (MCC) represents a wide variety of fault modes, such as stator open circuit,
stator short circuit, broken rotor bar, broken end ring, static eccentricity, dynamic eccentricity, mixed
eccentricity, and defective bearing. In contrast, the d-q model simplifies the representation of faults and reduces
the number of equations required for simulation, but it does not provide specific information about individual
rotor bars or end ring currents. The magnetic circuits (MEC) model provides detailed magnetic modeling with
reluctances and permeances to accurately simulate faults. The mathematical procedures (FEM) approach, on
the other hand, requires a lot of computing power but can accurately model faults in induction machines by
taking into account the nonlinearities of magnetic materials and reproducing the machine's performance.
Models featuring MCC

coupled circuits
d-q
Magnetic circuits
MEC
based models
IM fault models
Models based on
mathematical FEM
procedures.
Models with hybrid

FEM-Analytical
components.
Figure 7. Recent developments in IM models
4.1. Models featuring coupled circuits

The d-q model is a widely used coupling circuit model that assumes symmetrical motors, linear iron
permeability, uniform air-gap, and no tangential induction. These simplifications allow for a fast and accurate
mathematical model but are problematic when dealing with faulty equipment. This section discusses recent
breakthroughs in MCC and d-q models illustrated on Table 2.
Table 2. compiles references addressing outlining methods for incorporating specific faults into MCC, d-q,
MEC and FEM models
Faults MCC Reference d-q MEC FEM Remarks
Reference Reference Reference
Stator open [81] [82] N/A N/A d-q Model: More effective
circuit (MEC) Model: More accurate
Stator short [72], [83], [84] [85]–[87] [88] [71], [89]– d-q Model: More effective
circuit [91] (FEM): More suitable
Broken rotor bar [83], [84], [92]–[95] [77], [85], [88], [97], [71], [99], (MCC) Model: More comprehensive
[96] [98] [100] d-q Model: More effective
(MEC) Model: More accurate
Broken end ring [101], [102] [96] N/A [58], [103], (MCC) Model: More comprehensive
[104] d-q Model: More effective
(FEM): More suitable
Static [18], [19], [105] [35], [106] [92], [107] [20], [108] (MCC) Model: More comprehensive
eccentricity d-q Model: More effective
(FEM): More suitable
Dynamic [18], [19] [55], [105] [107] [49], [99] (MCC) Model: More comprehensive
Mathematical Procedures (FEM):
More suitable
Mixed [18], [109] [105] [110] [17], [26] (MCC) Model: More comprehensive
More suitable
Defective [35], [42], [46], [111], [113] [114] [115]–[117] (MCC) Model: More comprehensive
bearing [112] d-q Model: More effective
More suitable
4.1.1. Multiple coupled circuit models

After determining model parameters, the expressions characterizing IM behavior (related to rotor
conductors) and requiring stator solving are known as in (5) to (8) [84].
ⅆ
[𝑢𝑠 ] = [𝑅𝑠 ][𝐼𝑠 ] + [∅] (5)
ⅆ𝑡
[𝑄𝑠] = [𝐿5𝑠 ][𝐼𝑠 ] + [𝐿5𝑟 ][𝐼𝑟 ] (6)

ⅆ
[𝑢𝑟 ] = [𝑅𝑟 ][𝐼𝑟 ] + [𝜙𝑠 ] (7)
ⅆ𝑡
[𝜙𝑟 ] = [𝑙𝑟𝑠 ][𝐼𝑠 ] + [𝐿𝑟𝑟 ][𝐼𝑟 ] (8)

1982  ISSN: 2088-8694
The equations consist of various vectors and matrices, including [Us] for stator voltage, [Is] for stator currents,
[Ir] for rotor loop current, [∅s] for stator flux linkage, [Rs] for stator phase resistances, [Lss] for stator windings
inductance, and [Lsr] for stator to rotor mutual inductance. Additionally, [Ur] denotes the rotor voltages vector,
[Ir] the rotor currents vector, [Ir] the rotor loop currents vector, [∅r] the rotor flux linkages vector, [Rr] the
rotor resistance matrix, and [Lrr] the rotor inductance matrix.
4.1.2. d-q model

The implementation of the space vector transformation technique is beneficial for depicting an
induction machine with structural symmetry in simulations. By utilizing this method, it is possible to represent
such machines using only four differential equations that are linked together. This leads to a reduction in the
overall number of equations that are required for simulation purposes. Consequently, the equations that define
the stator voltage can be expressed as:
1 ⅆ 𝑄 ⅆ𝑠
[𝜈ⅆ𝑠 ] = − 𝜔∅ 𝑞 + 𝑅𝑠𝑖 𝑑𝑠 (9)
𝑤𝑏 ⅆ𝑡
1 ⅆ 𝑄 ⅆ𝑠
[𝜈𝑞𝑠 ] = − 𝜔∅ 𝑞 + 𝑅𝑖𝑄𝑆 (10)
𝑤𝑏 ⅆ𝑡
The variables in the aforementioned equation 9 and 10 can be defined as follows: wb refers to the base per-unit
electrical speed, while Øds, Øqs, and Ø0s represent the d-axis, q-axis, and zero-sequence stator flux
connections. The stator resistance is denoted by Rs, and the d-axis, q-axis, and zero-sequence stator currents
are respectively represented by ids, iqs, and i0s. Additionally, vdr and vqr refer to the d-axis and q-axis rotor
voltages, while φdr and φqr are the d-axis and q-axis rotor flux linkages. Furthermore, ω denotes the per-unit
synchronous speed, and ωr refers to the per-unit mechanical or rotational speed.
4.1.3. Models based on magnetic circuits (MEC)

In this approach, discrete winding distributions, stator and rotor slotting, and magnetic material
saturation-induced saliency effects related to space harmonics are taken into account [113]. The nodal magneto-
motive forces [F] are connected to the reluctances [R] in the (11), which represent the fluxes of the rotor and
stator [Ø].
[𝜙] = [𝐹][𝑅] (11)
In conclusion, the MEC-based framework has demonstrated high precision in predicting machine performance
across diverse operating points, load conditions, and even unbalanced excitation and faulty conditions. Its
accuracy and computational efficiency make it an ideal alternative between the standard lumped parameter
models and FEM-based approaches [114]. The MEC technique has also proven effective in modeling a range
of induction motor faults.
4.1.4. Models based on mathematical procedures (FEM)

By utilizing the machine's actual magnetic and geometric properties, this technique calculates the
distribution of the magnetic field [85]–[103]. In general, faulty induction motor models are created in 2D,
which offers excellent accuracy in terms of magnetic phenomena. However, these models do not consider the
rotor's skewing behavior or the end rings, and the connection of the rotor bars is typically addressed through
an ideal current source in the electrical circuit [89]. The magneto-dynamic field equation for a standard
induction motor in 2D is defined as:
ⅆ 1 ⅆ𝐴𝑧 ⅆ 1 ⅆ𝐴𝑧 ⅆ𝐴2
( )+ ( ) = −𝐽0 + 𝜎 − 𝜎 𝑣⃗×(𝛻𝑦𝐴) (12)
ⅆ𝑥 𝑣 ⅆ+ ⅆ𝑦 𝑣 ⅆ𝑦 ⅆ𝑡
In the given (12), (A) represents the magnetization potential, Az is the z-component of the magnetic vector
potential, J0 refers to the applied density current source, v represents velocity, σ is the electric conductivity,
and v represents permeability
4.2. Models with hybrid components

Recent research suggests that combining FEM and analytical methodologies can generate models with
FEM level accuracy, which can be executed in real-time simulators [58]. A hybrid model based on the d-q
method and finite element analysis has been proposed to model short circuit defects in IM drives. In this model,
sparse identification is utilized to minimize the number of FEM simulations needed to compute the IM coupling
parameters. FEM is used to solve the entire geometry of the IM, and the coupling parameters are then imported
into the machine's analytical model [58]. The sparse identification method is effective in obtaining a defective
IM model while minimizing the number of FEM simulations needed, thus reducing computing expenses by
over 99.9%. However, the full FEM analysis is still required for each failure scenario, resulting in extensive
simulation periods and expensive computational costs. TSFEM-based models require lengthy simulation times
for small simulated spans, whereas hybrid models require approximately 25 minutes [104]. Even when the time
to execute one simulation is factored in, the time savings exceed 98%.
5. CHALLENGES AND FUTURE DIRECTIONS

The existing challenges in fault detection for three-phase induction motors encompass a range of
factors [118], including the limitations of current signature analysis (CSA), the need for specialized diagnostic
methodologies, and the impact of variable frequency drives (VFDs). Furthermore, the reliance on Fourier
analysis for signal interpretation and feature extraction presents certain drawbacks, such as the lack of transient
information and the absence of spectrum content variations over time. These challenges have fostered the
exploration of various data-driven prognostics and health management (PHM) methodologies driven by
artificial intelligence, machine learning, and deep learning, aiming to leverage current, vibration, and thermal
signals for effective fault detection and isolation
The challenges in fault detection for three-phase induction motors can include issues with accurately
identifying incipient faults, distinguishing between various fault types, and dealing with the effects of operating
conditions and external disturbances. Additionally, extracting fault signatures from noisy measurements and
developing reliable and automated fault detection methods also present significant challenges in this domain
[31]. The existing challenges in fault detection for three-phase induction motors are multifaceted and
encompass various aspects of signal analysis, system complexity, and operational conditions. Some of the
prominent challenges include:
(a) Complex operating conditions: Induction motors operate in diverse industrial environments where
operating conditions such as variable loads and speeds, temperature variations, and mechanical stresses
can influence the manifestation of faults and complicate the diagnostic process [119].
(b) Incipient fault detection: Early detection of incipient faults, such as broken rotor bars, poses a significant
challenge due to the limited availability of diagnostic techniques capable of identifying subtle changes in
motor behavior at the initial stages of fault development [120].
(c) Signal interpretation: The interpretation of motor current and vibration signals to differentiate between
normal and faulty conditions requires accurate analytical models and sophisticated signal processing
techniques to extract relevant fault signatures and mitigate false alarms [121].
(d) Transient regimes: Fault detection during transient regimes, such as startup and shutdown, presents
challenges due to signal variations and the need for suitable methods to differentiate between normal
transient behavior and actual fault conditions [122].
(e) Noise and interference: The presence of electrical and mechanical noise, as well as interference from
external sources, can mask fault signatures in the acquired signals, making it challenging to extract relevant
diagnostic information [102].
(f) Operational dependence: The effectiveness of fault detection techniques can be influenced by the
operational characteristics of induction motors, necessitating robust diagnostic methods capable of
accommodating varying operating conditions and loads [32].
(g) Need for expert identification: Traditional fault detection in induction motors often relies on the expertise
of skilled engineers to interpret diagnostic data, highlighting the need for automated and intelligent
diagnostic systems to overcome human-related limitations [123].
(h) Inadequate diagnostic techniques: The limitations of conventional diagnostic methods in detecting specific
fault types, such as broken rotor bars, and the lack of comprehensive fault detection solutions present
challenges in ensuring reliable and accurate fault diagnosis in induction motors.
(i) Varied operating conditions: The fault detection process can be complicated due to the diverse operating
conditions experienced by induction motors, such as variable speeds, load variations, and environmental
factors [85].
(j) Dynamic stresses: Induction motors used in variable-speed applications undergo dynamic stresses at high
power levels, leading to reduced lifetimes compared to constant-speed motors, adding complexity to fault
detection [124].
(k) Non-deterministic fault characteristics: Some faults, such as eccentricity and certain rotor faults, may
manifest in non-stationary and intermittent characteristics, making them challenging to detect and
diagnose accurately [85], [125].
(l) Reliability and complexity of parameters: Parameters such as induced rotor voltage exhibit non-reliability
and complexity, posing challenges for their usage in condition monitoring and fault diagnosis [126].
1984  ISSN: 2088-8694
To get around these problems, we need to keep studying and creating more advanced diagnostic methods that
use cutting-edge signal analysis, machine learning, and artificial intelligence to make fault finding in three-
phase induction motors more accurate and reliable. Furthermore, integrating comprehensive fault diagnostic
systems that account for various operational scenarios and noise sources can significantly contribute to
improving the effectiveness of fault detection methodologies.
6. CONCLUSION
In conclusion, the landscape of fault diagnosis techniques for electrical machines has undergone
significant evolution to ensure the secure and reliable operation of these critical systems. The comprehensive
exploration within this review article has delved into a myriad of diagnostic techniques, ranging from
traditional methods like temperature measurement and infrared recognition to more advanced approaches such
as vibration analysis, MCSA, and artificial intelligence and neural network techniques. The article provides a
comprehensive exploration of numerous fault detection techniques for electrical machines, covering
advancements in traditional methods and cutting-edge approaches. The article talks about advanced techniques
such as AI-based methods, rectified stator current analysis, incremental broad learning (IBL), non-negative
matrix factorization (NMF), talking about how air gap flux changes over time and space, and other machine
learning and signal processing methods. The article underscores the paramount importance of fault detection
methods tailored for different types of faults inherent in electrical machines. The article focused on a variety
of fault categories, such as eccentricity, rotor faults (broken bars and end rings), stator or armature faults, and
bearing faults. Recognizing the dynamic nature of modern electrical machines, the article delved into advanced
fault detection methods. These included cutting-edge technologies like non-negative matrix factorization and
incremental broad learning, as well as methods based on artificial intelligence and measuring how air gap flux
changes over time and space. These advanced techniques' efficiency lies in their ability to provide robust fault
detection solutions. By enabling the timely identification of faults, they play a pivotal role in preventing
unexpected machine downtime and mitigating potential financial losses. However, the article aptly
acknowledges that the field is a dynamic one, and it advocates for continued research and development. This
call to action emphasizes the need for ongoing efforts to enhance fault diagnosis capabilities for electrical
machines, ensuring that diagnostic methodologies stay ahead of evolving challenges in the ever-changing
landscape of electrical engineering.
ACKNOWLEDGEMENTS
The Science, Technology, and Innovation Funding Authority (STDF) under Grant Number 48203
supported the work that forms the basis of this paper.
REFERENCES
[1] C. N. Okwuosa, U. E. Akpudo, and J.-W. Hur, “A cost-efficient MCSA-based fault diagnostic framework for SCIM at low-load
conditions,” Algorithms, vol. 15, no. 6, p. 212, Jun. 2022, doi: 10.3390/a15060212.
[2] S. Kamar et al., “Performance analysis of three-phase induction motor for railway propulsion system,” International Journal of
Power Electronics and Drive Systems, vol. 14, no. 3, pp. 1433–1441, 2023, doi: 10.11591/ijpeds.v14.i3.pp1433-1441.
[3] M. Baniyounis, A. Agha, H. Al-Kashashneh, and A. Al-Omoush, “Economic evaluation of induction motor based on motor’s
nameplate data and initial cost,” International Journal of Power Electronics and Drive Systems, vol. 13, no. 3, pp. 1340–1351,
2022, doi: 10.11591/ijpeds.v13.i3.pp1340-1351.
[4] N. T. García, Y. A. Garcés-Gomez, and V. Henao-Cespedes, “Robust control technique in power converter with linear induction
motor,” International Journal of Power Electronics and Drive Systems, vol. 13, no. 1, pp. 340–347, 2022, doi:
10.11591/ijpeds.v13.i1.pp340-347.
[5] A. Naha, K. R. Thammayyabbabu, A. K. Samanta, A. Routray, and A. K. Deb, “Mobile application to detect induction motor faults,”
IEEE Embed Syst Lett, vol. 9, no. 4, pp. 117–120, Dec. 2017, doi: 10.1109/LES.2017.2734798.
[6] G. S. Elias, C. Guy, R. Hubert, and S. Abdenour, “Fault diagnosis and prognosis for reliability enhancement,” in Fault Diagnosis,
Prognosis, and Reliability for Electrical Machines and Drives, IEEE, 2022, pp. 345–414.
[7] S. A. Bednarz and M. Dybkowski, “On-line detection of the rotor faults in the induction motor drive using parameter estimator,” in
2018 International Symposium on Electrical Machines (SME), IEEE, Jun. 2018, pp. 1–5. doi: 10.1109/ISEM.2018.8443020.
[8] D. G. Dorrell and K. Makhoba, “Detection of inter-turn stator faults in induction motors using short-term averaging of forward and
backward rotating stator current phasors for fast prognostics,” IEEE Trans Magn, vol. 53, no. 11, pp. 1–7, Nov. 2017, doi:
10.1109/TMAG.2017.2710181.
[9] S. Nandi, S. Ahmed, and H. A. Toliyat, “Detection of rotor slot and other eccentricity related harmonics in a three phase induction motor
with different rotor cages,” IEEE Transactions on Energy Conversion, vol. 16, no. 3, pp. 253–260, 2001, doi: 10.1109/60.937205.
[10] A. Almounajjed, A. K. Sahoo, M. K. Kumar, and M. W. Bakro, “Condition monitoring and fault diagnosis of induction motor - an
experimental analysis,” in 2021 7th International Conference on Electrical Energy Systems (ICEES), IEEE, Feb. 2021, pp. 433–
438. doi: 10.1109/ICEES51510.2021.9383729.
[11] J. A. Corral-Hernandez and J. A. Antonino-Daviu, “Thorough validation of a rotor fault diagnosis methodology in laboratory and
field soft-started induction motors,” Chinese Journal of Electrical Engineering, vol. 4, no. 3, pp. 66–72, Sep. 2018, doi:
10.23919/CJEE.2018.8471291.
[12] O. Alshorman and A. Alshorman, “A review of intelligent methods for condition monitoring and fault diagnosis of stator and rotor
faults of induction machines,” International Journal of Electrical and Computer Engineering, vol. 11, no. 4, pp. 2820–2829, 2021,
doi: 10.11591/ijece.v11i4.pp2820-2829.
[13] R. Boudiaf, B. Abdelkarim, and H. Issam, “Bearing fault diagnosis in induction motor using continuous wavelet transform and
convolutional neural networks,” International Journal of Power Electronics and Drive Systems, vol. 15, no. 1, pp. 591–602, 2024,
doi: 10.11591/ijpeds.v15.i1.pp591-602.
[14] A. T. Radhi and W. H. Zayer, “Faults diagnosis in stator windings of high speed solid rotor induction motors using fuzzy neural
network,” International Journal of Power Electronics and Drive Systems, vol. 12, no. 1, pp. 597–611, 2021, doi:
10.11591/ijpeds.v12.i1.pp597-611.
[15] N. A. Dobroskok et al., “Neural network based detecting induction motor defects supplied by unbalanced grid,” International
Journal of Power Electronics and Drive Systems (IJPEDS), vol. 14, no. 1, p. 185, 2023, doi: 10.11591/ijpeds.v14.i1.pp185-198.
[16] M. E. H. Benbouzid, “A review of induction motors signature analysis as a medium for faults detection,” IEEE Transactions on
Industrial Electronics, vol. 47, no. 5, pp. 984–993, 2000, doi: 10.1109/41.873206.
[17] J. Faiz and S. M. M. Moosavi, “Detection of mixed eccentricity fault in doubly‐fed induction generator based on reactive power
spectrum,” IET Electr Power Appl, vol. 11, no. 6, pp. 1076–1084, Jul. 2017, doi: 10.1049/iet-epa.2016.0449.
[18] J. Faiz, B. M. Ebrahimi, and H. A. Toliyat, “Effect of magnetic saturation on static and mixed eccentricity fault diagnosis in
induction motor,” IEEE Trans Magn, vol. 45, no. 8, pp. 3137–3144, Aug. 2009, doi: 10.1109/TMAG.2009.2016416.
[19] N. Yassa, M. Rachek, and H. Houassine, “Motor current signature analysis for the air gap eccentricity detection in the squirrel cage
induction machines,” Energy Procedia, vol. 162, pp. 251–262, Apr. 2019, doi: 10.1016/j.egypro.2019.04.027.
[20] M. M. Mafruddin, S. Suwarno, and A. Abu-Siada, “Finite element simulation of a 126 MW salient pole synchronous generator with
rotor eccentricity,” in 2019 2nd International Conference on High Voltage Engineering and Power Systems (ICHVEPS), IEEE, Oct.
2019, pp. 1–96. doi: 10.1109/ICHVEPS47643.2019.9011072.
[21] F. Mansou, Induction Motors: Construction, Principle of Operation, Power and Torque Calculations, Characteristics and Speed
Control. 2020. doi: http://dx.doi.org/10.13140/RG.2.2.15490.71360.
[22] R. S. Kumar, K. K. Ray, and K. V. Kumar, “Fault diagnosis of industrial drives using MCSA techniques,” in 2009 International
Conference on Control, Automation, Communication and Energy Conservation, IEEE, 2009.
[23] Z. Liu, P. Zhang, S. He, and J. Huang, “A review of modeling and diagnostic techniques for eccentricity fault in electric machines,”
Energies (Basel), vol. 14, no. 14, p. 4296, Jul. 2021, doi: 10.3390/en14144296.
[24] I. Bouchareb, A. Lebaroud, A. J. M. Cardoso, and S. Bin Lee, “Towards advanced diagnosis recognition for eccentricities faults:
application on induction motor,” in 2019 IEEE 12th International Symposium on Diagnostics for Electrical Machines, Power
Electronics and Drives (SDEMPED), IEEE, Aug. 2019, pp. 271–282. doi: 10.1109/DEMPED.2019.8864920.
[25] A. Polat, Y. D. Ertugrul, and L. T. Ergene, “Static, dynamic and mixed eccentricity of induction motor,” in 2015 IEEE 10th
International Symposium on Diagnostics for Electrical Machines, Power Electronics and Drives (SDEMPED), IEEE, Sep. 2015,
pp. 284–288. doi: 10.1109/DEMPED.2015.7303703.
[26] H. Torkaman, E. Afjei, and P. Yadegari, “Static, dynamic, and mixed eccentricity faults diagnosis in switched reluctance motors
using transient finite element method and experiments,” IEEE Trans Magn, vol. 48, no. 8, pp. 2254–2264, Aug. 2012, doi:
10.1109/TMAG.2012.2191619.
[27] N. Lashkari, H. F. Azgomi, J. Poshtan, and M. Poshtan, “Robust stator fault detection under load variation in induction motors using
AI techniques,” in 2015 IEEE International Electric Machines & Drives Conference (IEMDC), IEEE, May 2015, pp. 1446–1451.
doi: 10.1109/IEMDC.2015.7409252.
[28] Y. Park, H. Choi, S. B. Lee, and K. Gyftakis, “Flux-based detection of non-adjacent rotor bar damage in squirrel cage induction
motors,” in 2019 IEEE Energy Conversion Congress and Exposition (ECCE), IEEE, Sep. 2019, pp. 7019–7026. doi:
10.1109/ECCE.2019.8912859.
[29] D. G. Jerkan, D. D. Reljic, and D. P. Marcetic, “Broken rotor bar fault detection of im based on the counter-current braking method,”
IEEE Transactions on Energy Conversion, vol. 32, no. 4, pp. 1356–1366, Dec. 2017, doi: 10.1109/TEC.2017.2696578.
[30] H. J. Lee, S. H. Im, D. Y. Um, and G. S. Park, “A Design of rotor bar for improving starting torque by analyzing rotor resistance
and reactance in squirrel cage induction motor,” IEEE Trans Magn, vol. 54, no. 3, pp. 1–4, Mar. 2018, doi:
10.1109/TMAG.2017.2764525.
[31] W. T. Thomson, “MCSA case histories diagnosis of cage winding defects in SCIM s fitted with end ring retaining rings,” in Current
Signature Analysis for Condition Monitoring of Cage Induction Motors, Wiley-IEEE Press, 2016, pp. 147–172. doi:
10.1002/9781119175476.ch6.
[32] L. A. Trujillo-Guajardo, J. Rodriguez-Maldonado, M. A. Moonem, and M. A. Platas-Garza, “A multiresolution taylor–kalman
approach for broken rotor bar detection in cage induction motors,” IEEE Trans Instrum Meas, vol. 67, no. 6, pp. 1317–1328, Jun.
2018, doi: 10.1109/TIM.2018.2795895.
[33] T. Bensic, T. Varga, M. Barukcic, S. Krizanic, and V. JerkovicStil, “Identification of Inertia constant from induction motor load
transient,” in 2019 International IEEE Conference and Workshop in Óbuda on Electrical and Power Engineering (CANDO-EPE),
IEEE, Nov. 2019, pp. 55–60. doi: 10.1109/CANDO-EPE47959.2019.9111041.
[34] P. A. Panagiotou, I. Arvanitakis, N. Lophitis, J. A. Antonino-Daviu, and K. N. Gyftakis, “A new approach for broken rotor bar
detection in induction motors using frequency extraction in stray flux signals,” IEEE Trans Ind Appl, vol. 55, no. 4, pp. 3501–3511,
Jul. 2019, doi: 10.1109/TIA.2019.2905803.
[35] C. Terron-Santiago, J. Martinez-Roman, R. Puche-Panadero, and A. Sapena-Bano, “Low-computational-cost hybrid FEM-
analytical induction machine model for the diagnosis of rotor eccentricity, based on sparse identification techniques and
trigonometric interpolation,” Sensors, vol. 21, no. 21, p. 6963, Oct. 2021, doi: 10.3390/s21216963.
[36] A. Naha, A. K. Samanta, A. Routray, and A. K. Deb, “A method for detecting half-broken rotor bar in lightly loaded induction
motors using current,” IEEE Trans Instrum Meas, vol. 65, no. 7, pp. 1614–1625, Jul. 2016, doi: 10.1109/TIM.2016.2540941.
[37] J. Burriel-Valencia et al., “Automatic fault diagnostic system for induction motors under transient regime optimized with expert
systems,” Electronics (Basel), vol. 8, no. 1, p. 6, Dec. 2018, doi: 10.3390/electronics8010006.
[38] J. Yun, K. Lee, K.-W. Lee, S. Bin Lee, and J.-Y. Yoo, “Detection and classification of stator turn faults and high-resistance electrical
connections for induction machines,” IEEE Trans Ind Appl, vol. 45, no. 2, pp. 666–675, 2009, doi: 10.1109/TIA.2009.2013557.
[39] M. Wolkiewicz, G. Tarchala, T. Orlowska-Kowalska, and C. T. Kowalski, “online stator interturn short circuits monitoring in the
DFOC induction-motor drive,” IEEE Transactions on Industrial Electronics, vol. 63, no. 4, pp. 2517–2528, Apr. 2016, doi:
10.1109/TIE.2016.2520902.
1986  ISSN: 2088-8694
[40] O. I. Okoro, “Steady and transient states thermal analysis of a 7.5-kW squirrel-cage induction machine at rated-load operation,”
IEEE Transactions on Energy Conversion, vol. 20, no. 4, pp. 730–736, Dec. 2005, doi: 10.1109/TEC.2005.852965.
[41] K. Laadjal, H. R. P. Antunes, M. Sahraoui, F. Bento, and A. J. M. Cardoso, “online diagnosis and discrimination of stator faults in
six-phase induction motors based on voltage symmetrical components,” IEEE Transactions on Transportation Electrification, vol.
9, no. 2, pp. 3115–3126, Jun. 2023, doi: 10.1109/TTE.2022.3219722.
[42] V. F. Victor, F. O. Quintaes, J. S. B. Lopes, L. dos S. Junior, A. S. Lock, and A. O. Salazar, “Analysis and study of a bearingless
ac motor type divided winding, based on a conventional squirrel cage induction motor,” IEEE Trans Magn, vol. 48, no. 11, pp.
3571–3574, Nov. 2012, doi: 10.1109/TMAG.2012.2201142.
[43] Z. Li, Q. Gao, and W. Zhang, “An improved current sampling scheme using three resistors for induction motor drives based on
current prediction,” in 2016 IEEE 8th International Power Electronics and Motion Control Conference (IPEMC-ECCE Asia), IEEE,
May 2016, pp. 104–108. doi: 10.1109/IPEMC.2016.7512269.
[44] T. Yang, H. Pen, Z. Wang, and C. S. Chang, “Feature knowledge-based fault detection of induction motors through the analysis of
stator current data,” IEEE Trans Instrum Meas, vol. 65, no. 3, pp. 549–558, Mar. 2016, doi: 10.1109/TIM.2015.2498978.
[45] M. Blodt, P. Granjon, B. Raison, and G. Rostaing, “Models for bearing damage detection in induction motors using stator current
monitoring,” IEEE Transactions on Industrial Electronics, vol. 55, no. 4, pp. 1813–1822, Apr. 2008, doi:
10.1109/TIE.2008.917108.
[46] M. Ojaghi, M. Sabouri, and J. Faiz, “Analytic model for induction motors under localized bearing faults,” IEEE Transactions on
Energy Conversion, vol. 33, no. 2, pp. 617–626, Jun. 2018, doi: 10.1109/TEC.2017.2758382.
[47] S. E. Pandarakone, Y. Mizuno, and H. Nakamura, “Distinct fault analysis of induction motor bearing using frequency spectrum
determination and support vector machine,” IEEE Trans Ind Appl, vol. 53, no. 3, pp. 3049–3056, May 2017, doi:
10.1109/TIA.2016.2639453.
[48] E. Navasari, D. A. Asfani, and M. Y. Negara, “Detection of induction motor bearing damage with starting current analysis using
wavelet discrete transform and artificial neural network,” in 2018 10th International Conference on Information Technology and
Electrical Engineering (ICITEE), IEEE, Jul. 2018, pp. 316–319. doi: 10.1109/ICITEED.2018.8534749.
[49] A. Bouzida, R. Abdelli, O. Touhami, and A. Aibeche, “Dynamic eccentricity fault diagnosis in induction motors using finite element
method and experimental tests,” International Journal of Industrial Electronics and Drives, vol. 3, no. 4, pp. 199–209, 2017, doi:
10.1504/IJIED.2017.10007181.
[50] A. H. Boudinar, N. Benouzza, A. Bendiabdellah, and M.-E.-A. Khodja, “Induction motor bearing fault analysis using a root-music
method,” IEEE Trans Ind Appl, vol. 52, no. 5, pp. 3851–3860, Sep. 2016, doi: 10.1109/TIA.2016.2581143.
[51] L. Frosini, C. Harlisca, and L. Szabo, “Induction machine bearing fault detection by means of statistical processing of the stray flux
measurement,” IEEE Transactions on Industrial Electronics, vol. 62, no. 3, pp. 1846–1854, 2015, doi: 10.1109/TIE.2014.2361115.
[52] I. Chernyavska and O. Vitek, “Analysis of air-gap eccentricity in inverter fed induction motor by means of motor current signature
analysis and stray flux of motor,” in 2017 IEEE 11th International Symposium on Diagnostics for Electrical Machines, Power
Electronics and Drives (SDEMPED), IEEE, Aug. 2017, pp. 72–76. doi: 10.1109/DEMPED.2017.8062336.
[53] S. Shao, W. Sun, P. Wang, R. X. Gao, and R. Yan, “Learning features from vibration signals for induction motor fault diagnosis,”
in 2016 International Symposium on Flexible Automation (ISFA), IEEE, Aug. 2016, pp. 71–76. doi: 10.1109/ISFA.2016.7790138.
[54] S. B. Jiang, P. K. Wong, R. Guan, Y. Liang, and J. Li, “An efficient fault diagnostic method for three-phase induction motors based
on incremental broad learning and non-negative matrix factorization,” IEEE Access, vol. 7, pp. 17780–17790, 2019, doi:
10.1109/ACCESS.2019.2895909.
[55] H. Samir, T. Omar, and I. Rachid, “Generalized two axes model of a squirrel-cage induction motor for rotor fault diagnosis,” Serbian
Journal of Electrical Engineering, vol. 5, no. 1, pp. 155–170, 2008, doi: 10.2298/SJEE0801155S.
[56] P. K. Wong and J. S. Biao, “Fault diagnosis of induction motors under untrained loads with a feature adaptation and improved broad
learning framework,” IEEE/ASME Transactions on Mechatronics, vol. 27, no. 5, pp. 3041–3052, Oct. 2022, doi:
10.1109/TMECH.2021.3125767.
[57] S. Chattopadhyay, S. Karmakar, M. Mitra, and S. Sengupta, “Symmetrical components and current Concordia based assessment of
single phasing of an induction motor by feature pattern extraction method and radar analysis,” International Journal of Electrical
Power & Energy Systems, vol. 37, no. 1, pp. 43–49, May 2012, doi: 10.1016/j.ijepes.2011.12.002.
[58] R. Puche-Panadero, J. Martinez-Roman, A. Sapena-Bano, and J. Burriel-Valencia, “Diagnosis of rotor asymmetries faults in
induction machines using the rectified stator current,” IEEE Transactions on Energy Conversion, vol. 35, no. 1, pp. 213–221, Mar.
2020, doi: 10.1109/TEC.2019.2951008.
[59] G. Mirzaeva, K. I. Saad, and M. G. Jahromi, “Comprehensive diagnostics of induction motor faults based on measurement of space
and time dependencies of air gap flux,” IEEE Trans Ind Appl, vol. 53, no. 3, pp. 2657–2666, May 2017, doi:
10.1109/TIA.2016.2628718.
[60] A. Widodo et al., “Fault diagnosis of low speed bearing based on relevance vector machine and support vector machine,” Expert
Syst Appl, vol. 36, no. 3, pp. 7252–7261, Apr. 2009, doi: 10.1016/j.eswa.2008.09.033.
[61] A. Abdelkefi, A. Souissi, and I. Abdhennader, “Analytical investigation of flux switching pm machines: air gap flux density
formulation,” in 2022 Second International Conference on Sustainable Mobility Applications, Renewables and Technology
(SMART), IEEE, Nov. 2022, pp. 1–7. doi: 10.1109/SMART55236.2022.9990436.
[62] A. Gonzalez-Parada, F. Trillaud, R. Guzman-Cabrera, and M. Abatal, “Torque ripple reduction in an axial flux high temperature
superconducting motor,” IEEE Transactions on Applied Superconductivity, vol. 25, no. 3, pp. 1–5, Jun. 2015, doi:
10.1109/TASC.2014.2384738.
[63] J. Liu, C. Di, and X. Bao, “An efficient air-gap flux density analysis method for the design of induction machines,” IEEE Trans
Magn, vol. 58, no. 8, pp. 1–6, Aug. 2022, doi: 10.1109/TMAG.2022.3154748.
[64] J. Herman, S. Begus, P. Mihalic, and J. Bojkovski, “Novel method for direct measurement of air gap anomalies in direct-drive
electrical motors,” IEEE Transactions on Industrial Electronics, vol. 67, no. 3, pp. 2422–2429, Mar. 2020, doi:
10.1109/TIE.2019.2905832.
[65] P. Redon, M. J. P. Rodenas, and J. Antonino-Daviu, “Development of a diagnosis tool, based on deep learning algorithms and
infrared images, applicable to condition monitoring of induction motors under transient regime,” in IECON 2020 The 46th Annual
Conference of the IEEE Industrial Electronics Society, IEEE, Oct. 2020, pp. 2505–2510. doi: 10.1109/IECON43393.2020.9254639.
[66] G. H. Bazan, P. R. Scalassara, W. Endo, A. Goedtel, R. H. C. Palacios, and W. F. Godoy, “Stator short-circuit diagnosis in induction
motors using mutual information and intelligent systems,” IEEE Transactions on Industrial Electronics, vol. 66, no. 4, pp. 3237–
3246, Apr. 2019, doi: 10.1109/TIE.2018.2840983.
[67] D. G. Dorrell, J. K. H. Shek, and M.-F. Hsieh, “the development of an indexing method for the comparison of unbalanced magnetic
pull in electrical machines,” IEEE Trans Ind Appl, vol. 52, no. 1, pp. 145–153, Jan. 2016, doi: 10.1109/TIA.2015.2466554.
[68] M. Subha, N. S. Kumar, and K. S. K. Veni, “Artificial intelligence based stator winding fault estimation in three phase induction
motor,” in 2018 Second International Conference on Electronics, Communication and Aerospace Technology (ICECA), IEEE, Mar.
2018, pp. 1929–1933. doi: 10.1109/ICECA.2018.8474562.
[69] F. Zidani, M. E. H. Benbouzid, D. Diallo, and M. S. Nait-Said, “Induction motor stator faults diagnosis by a current concordia
pattern-based fuzzy decision system,” IEEE Transactions on Energy Conversion, vol. 18, no. 4, pp. 469–475, Dec. 2003, doi:
10.1109/TEC.2003.815832.
[70] A. Lakrout, M. Bendjebbar, and N. Henini, “Electric stator faults detection in induction motor based on fuzzy logic technique,” in
2018 International Conference on Applied Smart Systems (ICASS), IEEE, Nov. 2018, pp. 1–5. doi: 10.1109/ICASS.2018.8651983.
[71] A.-I. Constantin, “Detection based on stator current signature of the single and combined short-circuit, broken bar and eccentricity
faults in induction motors,” in 2019 11th International Symposium on Advanced Topics in Electrical Engineering (ATEE), IEEE,
Mar. 2019, pp. 1–6. doi: 10.1109/ATEE.2019.8724984.
[72] A. G. de A. Cruz, R. D. Gomes, F. A. Belo, and A. C. L. Filho, “A hybrid system based on fuzzy logic to failure diagnosis in
induction motors,” IEEE Latin America Transactions, vol. 15, no. 8, pp. 1480–1489, 2017, doi: 10.1109/TLA.2017.7994796.
[73] J. Antonino-Daviu and L. Dunai, “Software tool for training on electric motors condition monitoring,” in 2022 IEEE 9th
International Conference on e-Learning in Industrial Electronics (ICELIE), IEEE, Oct. 2022, pp. 1–6. doi:
10.1109/ICELIE55228.2022.9969431.
[74] F. Babaa and O. Bennis, “An accurate inter-turn short circuit faults model dedicated to induction motors,” International Journal of
Electrical and Computer Engineering (IJECE), vol. 11, no. 1, pp. 9–16, Feb. 2021, doi: 10.11591/ijece.v11i1.pp9-16.
[75] Min Kang and Jin Huang, “Simulation and analysis of squirrel cage induction machines under rotor internal faults,” in 2005
International Conference on Electrical Machines and Systems, IEEE, 2005, pp. 2023–2027. doi: 10.1109/ICEMS.2005.202916.
[76] H. A. Toliyat, S. Nandi, S. Choi, and H. Meshgin-Kelk, Electric machines: modeling, condition monitoring, and fault diagnosis.
CRC Press Taylor & Francis Group, 2012.
[77] P. Gangsar and R. Tiwari, “Signal based condition monitoring techniques for fault detection and diagnosis of induction motors: A
state-of-the-art review,” Mech Syst Signal Process, vol. 144, p. 106908, Oct. 2020, doi: 10.1016/j.ymssp.2020.106908.
[78] Z. Yin and J. Hou, “Recent advances on SVM based fault diagnosis and process monitoring in complicated industrial processes,”
Neurocomputing, vol. 174, pp. 643–650, Jan. 2016, doi: 10.1016/j.neucom.2015.09.081.
[79] J. B. Ali, N. Fnaiech, L. Saidi, B. Chebel-Morello, and F. Fnaiech, “Application of empirical mode decomposition and artificial
neural network for automatic bearing fault diagnosis based on vibration signals,” Applied Acoustics, vol. 89, pp. 16–27, Mar. 2015,
doi: 10.1016/j.apacoust.2014.08.016.
[80] A. Kabul and A. Unsal, “An alternative approach for the detection of broken rotor bars and bearing faults of induction motor based
on vibration signals,” in 2021 8th International Conference on Electrical and Electronics Engineering (ICEEE), IEEE, Apr. 2021,
pp. 126–131. doi: 10.1109/ICEEE52452.2021.9415920.
[81] V. Devanneaux, B. Dagues, J. Faucher, and G. Barakat, “An accurate model of squirrel cage induction machines under stator faults,”
Math Comput Simul, vol. 63, no. 3–5, pp. 377–391, Nov. 2003, doi: 10.1016/S0378-4754(03)00083-1.
[82] M. Jannati, N. R. N. Idris, and Z. Salam, “A new method for modeling and vector control of unbalanced induction motors,” in 2012
IEEE Energy Conversion Congress and Exposition (ECCE), IEEE, Sep. 2012, pp. 3625–3632. doi: 10.1109/ECCE.2012.6342483.
[83] J. Tang, J. Chen, K. Dong, Y. Yang, H. Lv, and Z. Liu, “Modeling and evaluation of stator and rotor faults for induction motors,”
Energies (Basel), vol. 13, no. 1, p. 133, Dec. 2019, doi: 10.3390/en13010133.
[84] M. S. R. Krishna and K. S. Ravi, “Fault diagnosis of induction motor using Motor Current Signature Analysis,” in 2013 International
Conference on Circuits, Power and Computing Technologies (ICCPCT), IEEE, Mar. 2013, pp. 180–186. doi:
10.1109/ICCPCT.2013.6528849.
[85] A. Choudhary, D. C. Meena, and A. K. Patra, “Asynchronous motor modeling in simulink for stator and rotor fault analysis,” in
2019 International Conference on Green and Human Information Technology (ICGHIT), IEEE, Jan. 2019, pp. 82–85. doi:
10.1109/ICGHIT.2019.00026.
[86] L. S. Maraaba, S. Twaha, A. Memon, and Z. Al-Hamouz, “Recognition of stator winding inter-turn fault in interior-mount LSPMSM
using acoustic signals,” Symmetry (Basel), vol. 12, no. 8, p. 1370, Aug. 2020, doi: 10.3390/sym12081370.
[87] G. Zhang, “Hardware-in-the-loop for on-line identification of SSP driving motor,” in Measuring Technology and Mechatronics
Automation in Electrical Engineering, 2012, pp. 251–256. doi: 10.1007/978-1-4614-2185-6_31.
[88] P. Naderi, M. Rostami, and A. Ramezannezhad, “Phase-to-phase fault detection method for synchronous reluctance machine using
MEC method,” Electrical Engineering, vol. 101, no. 2, pp. 575–586, Jun. 2019, doi: 10.1007/s00202-019-00806-9.
[89] H. H. Eldeeb, A. Berzoy, and O. Mohammed, “Stator fault detection on DTC-driven IM via magnetic signatures aided by 2-D FEA
Co-simulation,” IEEE Trans Magn, vol. 55, no. 6, pp. 1–5, Jun. 2019, doi: 10.1109/TMAG.2019.2892707.
[90] V. Fireteanu, V. Leconte, and A.-I. Constantin, “Finite element analysis on early detection of the short-circuit faults in induction
machines through harmonics of the stator currents or of the neighboring magnetic field,” in 2017 International Conference on
Optimization of Electrical and Electronic Equipment (OPTIM) & 2017 Intl Aegean Conference on Electrical Machines and Power
Electronics (ACEMP), IEEE, May 2017, pp. 420–427. doi: 10.1109/OPTIM.2017.7975006.
[91] K. Prasob, N. P. Kumar, and T. B. Isha, “Inter-turn short circuit fault analysis of PWM inverter fed three-phase induction motor
using Finite Element Method,” in 2017 International Conference on Circuit, Power and Computing Technologies (ICCPCT), IEEE,
Apr. 2017, pp. 1–6. doi: 10.1109/ICCPCT.2017.8074384.
[92] J. Faiz, M. Ghasemi-Bijan, and B. Mahdi Ebrahimi, “Modeling and diagnosing eccentricity fault using three-dimensional magnetic
equivalent circuit model of three-phase squirrel-cage induction motor,” Electric Power Components and Systems, vol. 43, no. 11,
pp. 1246–1256, Jul. 2015, doi: 10.1080/15325008.2015.1029651.
[93] S. E. Zouzou, A. Ghoggal, A. Aboubou, M. Sahraoui, and H. Razik, “Modeling of induction machines with skewed rotor slots
dedicated to rotor faults,” in 2005 5th IEEE International Symposium on Diagnostics for Electric Machines, Power Electronics and
Drives, IEEE, Sep. 2005, pp. 1–6. doi: 10.1109/DEMPED.2005.4662501.
[94] M. Y. Kaikaa, M. Hadjami, and A. Khezzar, “Effects of the simultaneous presence of static eccentricity and broken rotor bars on
the stator current of induction machine,” IEEE Transactions on Industrial Electronics, vol. 61, no. 5, pp. 2452–2463, May 2014,
doi: 10.1109/TIE.2013.2270216.
[95] M. Bayrak and A. Küçüker, “Detection of rotor bar corrosion in three phase asynchronous motors using wavelet analysis,” in 2010
Modern Electric Power Systems, 2010.

1988  ISSN: 2088-8694
[96] J. F. Bangura, R. J. Povinelli, N. A. O. Demerdash, and R. H. Brown, “Diagnostics of eccentricities and bar/end-ring connector
breakages in polyphase induction motors through a combination of time-series data mining and time-stepping coupled FE-state
space techniques,” in Conference Record of the 2001 IEEE Industry Applications Conference. 36th IAS Annual Meeting (Cat.
No.01CH37248), IEEE, 2001, pp. 1579–1586. doi: 10.1109/IAS.2001.955746.
[97] P. Naderi, “Modified magnetic‐equivalent‐circuit approach for various faults studying in saturable double‐cage‐induction
machines,” IET Electr Power Appl, vol. 11, no. 7, pp. 1224–1234, Aug. 2017, doi: 10.1049/iet-epa.2016.0782.
[98] B. Jandaghi and V. Dinavahi, “Real-time HIL emulation of faulted electric machines based on nonlinear MEC model,” IEEE
Transactions on Energy Conversion, vol. 34, no. 3, pp. 1190–1199, Sep. 2019, doi: 10.1109/TEC.2019.2891560.
[99] T. J. Sobczyk, J. Tulicki, K. Weinreb, R. Mielnik, and M. Sulowicz, “Characteristic features of rotor bar current frequency spectrum
in cage induction machine with inner faults,” in 2019 IEEE 12th International Symposium on Diagnostics for Electrical Machines,
Power Electronics and Drives (SDEMPED), IEEE, Aug. 2019, pp. 115–120. doi: 10.1109/DEMPED.2019.8864907.
[100] B. Asad, L. Eensalu, T. Vaimann, A. Kallaste, A. Rassolkin, and A. Belahcen, “The FEM based modeling and corresponding test
rig preparation for broken rotor bars analysis,” in 2019 IEEE 60th International Scientific Conference on Power and Electrical
Engineering of Riga Technical University (RTUCON), IEEE, Oct. 2019, pp. 1–9. doi: 10.1109/RTUCON48111.2019.8982350.
[101] O. Touhami, L. Noureddine, R. Ibtiouen, and M. Fadel, “Modeling of the induction machine for the diagnosis of rotor defects. Part.
II. Simulation and experimental results,” in 31st Annual Conference of IEEE Industrial Electronics Society, 2005. IECON 2005.,
IEEE, 2005, p. 6. doi: 10.1109/IECON.2005.1569147.
[102] A. Iqbal, F. I. Bakhsh, and G. K. Singh, “Operation and testing of indirect field oriented control of asymmetrical six-phase open-
ended winding induction machine using hardware-in-loop (HIL) emulator,” International Transactions on Electrical Energy
Systems, vol. 2023, pp. 1–13, Jan. 2023, doi: 10.1155/2023/4623140.
[103] J. F. Bangura, R. J. Povinelli, N. A. O. Demerdash, and R. H. Brown, “Diagnostics of eccentricities and bar/end-ring connector
breakages in polyphase induction motors through a combination of time-series data mining and time-stepping coupled fe$~$state-
space techniques,” IEEE Trans Ind Appl, vol. 39, no. 4, pp. 1005–1013, Jul. 2003, doi: 10.1109/TIA.2003.814582.
[104] A. Sapena-Bano, F. Chinesta, R. Puche-Panadero, J. Martinez-Roman, and M. Pineda-Sanchez, “Model reduction based on sparse
identification techniques for induction machines: Towards the real time and accuracy-guaranteed simulation of faulty induction
machines,” International Journal of Electrical Power & Energy Systems, vol. 125, p. 106417, Feb. 2021, doi:
10.1016/j.ijepes.2020.106417.
[105] S. Bindu and V. V Thomas, “A modified direct-quadrature axis model for characterization of air-gap mixed eccentricity faults in
three-phase induction motor,” International Review on Modelling and Simulations (IREMOS), vol. 11, no. 6, p. 359, Dec. 2018,
doi: 10.15866/iremos.v11i6.15513.
[106] C. Terron-Santiago, J. Martinez-Roman, R. Puche-Panadero, A. Sapena-Bano, J. Burriel-Valencia, and M. Pineda-Sanchez,
“Analytical model of eccentric induction machines using the conformal winding tensor approach,” Sensors, vol. 22, no. 9, p. 3150,
Apr. 2022, doi: 10.3390/s22093150.
[107] R. Alipour-Sarabi, Z. Nasiri-Gheidari, and H. Oraee, “Development of a three-dimensional magnetic equivalent circuit model for
axial flux machines,” IEEE Transactions on Industrial Electronics, vol. 67, no. 7, pp. 5758–5767, Jul. 2020, doi:
10.1109/TIE.2019.2934065.
[108] J. Sobra, T. Kavalir, M. Krizek, and B. Skala, “Experimental verification of the finite element analysis of an induction machine
with implemented static eccentricity fault,” in 2018 18th International Conference on Mechatronics - Mechatronika (ME), 2019.
[109] G. M. Joksimovic, M. D. Durovic, J. Penman, and N. Arthur, “Dynamic simulation of dynamic eccentricity in induction machines-
winding function approach,” IEEE Transactions on Energy Conversion, vol. 15, no. 2, pp. 143–148, Jun. 2000, doi:
10.1109/60.866991.
[110] J. Lannoo, D. Vanoost, J. Peuteman, S. Debruyne, H. De Gersem, and D. Pissoort, “Improved air gap permeance model to
characterise the transient behaviour of electrical machines using magnetic equivalent circuit method,” International Journal of
Numerical Modelling: Electronic Networks, Devices and Fields, vol. 33, no. 5, Sep. 2020, doi: 10.1002/jnm.2749.
[111] L. Huang, N. Hu, Y. Yang, L. Chen, J. Wen, and G. Shen, “Study on electromagnetic–dynamic coupled modeling method—
detection by stator current of the induction motors with bearing faults,” Machines, vol. 10, no. 8, p. 682, Aug. 2022, doi:
10.3390/machines10080682.
[112] C. Wang, M. Wang, B. Yang, and K. Song, “A model-based method for bearing fault detection using motor current,” J Phys Conf
Ser, vol. 1650, no. 3, p. 32130, Oct. 2020, doi: 10.1088/1742-6596/1650/3/032130.
[113] X. Liang, M. Z. Ali, and H. Zhang, “Induction motors fault diagnosis using finite element method: a review,” IEEE Trans Ind Appl,
vol. 56, no. 2, pp. 1205–1217, Mar. 2020, doi: 10.1109/TIA.2019.2958908.
[114] M. Z. Ali, M. N. S. K. Shabbir, X. Liang, Y. Zhang, and T. Hu, “Machine learning-based fault diagnosis for single- and multi-faults
in induction motors using measured stator currents and vibration signals,” IEEE Trans Ind Appl, vol. 55, no. 3, pp. 2378–2391, May
2019, doi: 10.1109/TIA.2019.2895797.
[115] Y. Gao, X. Liu, and J. Xiang, “FEM simulation-based generative adversarial networks to detect bearing faults,” IEEE Trans Industr
Inform, vol. 16, no. 7, pp. 4961–4971, Jul. 2020, doi: 10.1109/TII.2020.2968370.
[116] F. Immovilli, C. Bianchini, M. Cocconcelli, A. Bellini, and R. Rubini, “bearing fault model for induction motor with externally
induced vibration,” IEEE Transactions on Industrial Electronics, vol. 60, no. 8, pp. 3408–3418, Aug. 2013, doi:
10.1109/TIE.2012.2213566.
[117] C. Vinothraj, N. Praveen Kumar, and T. B. Isha, “bearing fault analysis in induction motor drives using finite element method,”
International Journal of Engineering & Technology, vol. 7, no. 3.6, p. 30, Jul. 2018, doi: 10.14419/ijet.v7i3.6.14928.
[118] W. Abu Elhaija and Q. Abu Al-Haija, “A novel dataset and lightweight detection system for broken bars induction motors using
optimizable neural networks,” Intelligent Systems with Applications, vol. 17, p. 200167, Feb. 2023, doi:
10.1016/j.iswa.2022.200167.
[119] S. Lu, Q. He, T. Yuan, and F. Kong, “Online fault diagnosis of motor bearing via stochastic-resonance-based adaptive filter in an
embedded system,” IEEE Trans Syst Man Cybern Syst, vol. 47, no. 7, pp. 1111–1122, Jul. 2017, doi: 10.1109/TSMC.2016.2531692.
[120] H. Sabir, M. Ouassaid, and N. Ngote, “An experimental method for diagnostic of incipient broken rotor bar fault in induction
machines,” Heliyon, vol. 8, no. 3, Mar. 2022, doi: 10.1016/j.heliyon.2022.e09136.
[121] M. Valtierra-Rodriguez, J. R. Rivera-Guillen, J. J. De Santiago-Perez, G. I. Perez-Soto, and J. P. Amezquita-Sanchez, “Expert
system based on autoencoders for detection of broken rotor bars in induction motors employing start-up and steady-state regimes,”
Machines, vol. 11, no. 2, p. 156, Jan. 2023, doi: 10.3390/machines11020156.
[122] O. AlShorman et al., “Advancements in condition monitoring and fault diagnosis of rotating machinery: A comprehensive review
of image-based intelligent techniques for induction motors,” Eng Appl Artif Intell, vol. 130, p. 107724, Apr. 2024, doi:
10.1016/j.engappai.2023.107724.
[123] A. Nemat Saberi, A. Belahcen, J. Sobra, and T. Vaimann, “LightGBM-based fault diagnosis of rotating machinery under changing
working conditions using modified recursive feature elimination,” IEEE Access, vol. 10, pp. 81910–81925, 2022, doi:
10.1109/ACCESS.2022.3195939.
[124] G. Mirzaeva and K. I. Saad, “Advanced diagnosis of rotor faults and eccentricity in induction motors based on internal flux
measurement,” IEEE Trans Ind Appl, vol. 54, no. 3, pp. 2981–2991, May 2018, doi: 10.1109/TIA.2018.2805730.
[125] U. Khalique et al., “A novel detection method for diagnosis of rotor eccentricity in three-phase induction motor,” Meas Sci Technol,
vol. 32, no. 11, p. 114002, Nov. 2021, doi: 10.1088/1361-6501/ac06fe.
[126] W. T. Thomson and M. Fenger, “Current signature analysis to detect induction motor faults,” IEEE Industry Applications Magazine,
vol. 7, no. 4, pp. 26–34, 2001, doi: 10.1109/2943.930988.
BIOGRAPHIES OF AUTHORS
Bahgat Hafez Bahgat is a M.Sc. student in Electrical Engineering Department

at Zagazig University, Zagazig, Egypt. He received his B.Eng. degree in Mechatronics
Engineering from Heliopolis University, Egypt, in 2019. He is currently an automation
engineer at Elsewedy Elec. His research interests include the field of artificial intelligence
and intelligent control. He can be contacted at email: b.hafez020@eng.zu.edu.eg.
Enas A. Elhay received the B.Sc. degree (1998), M.Sc. degree (2004) and Ph.D.
degree (2010), all in Electrical Power and Machines Engineering from Zagazig University in
Zagazig, Egypt. She has been with the University of Zagazig since 1999, presently as
associate professor. Her current interests include renewable energy, energy conversion,
optimization of electrical machines performance, modelling and control of electrical
machines, applications of power electronics, and application of artificial intelligence to solve
electrical machines problems. She can be contacted at email: eaabdelhay@eng.zu.edu.eg.
Tole Sutikno is a lecturer and the head of the Master Program of Electrical
Engineering at the Faculty of Industrial Technology at Universitas Ahmad Dahlan (UAD) in
Yogyakarta, Indonesia. He received his bachelor of engineering from Universitas
Diponegoro in 1999, master of engineering from Universitas Gadjah Mada in 2004, and
doctor of philosophy in Electrical Engineering from Universiti Teknologi Malaysia in 2016.
All three degrees are in Electrical Engineering. He has been a professor at UAD in
Yogyakarta, Indonesia, since July 2023, following his tenure as an associate professor in June
2008. He is the Editor-in-Chief of TELKOMNIKA and head of the Embedded Systems and
Power Electronics Research Group (ESPERG). He is one of the top 2% of researchers
worldwide, according to Stanford University and Elsevier BV’s list of the most influential
scientists from 2021 to the present. His research interests cover digital design, industrial
applications, industrial electronics, industrial informatics, power electronics, motor drives,
renewable energy, FPGA applications, embedded systems, artificial intelligence, intelligent
control, digital libraries, and information technology. He can be contacted at email:
tole@te.uad.ac.id.
Mahmoud M. Elkholy received the B.Sc. degree (1994), M.Sc. degree (1998)
and Ph.D. degree (2001), all in Electrical Power and Machines Engineering from Zagazig
University in Zagazig, Egypt. He has been with the University of Zagazig since 1995,
presently as full professor. He is a senior member of the IEEE. His research is concerned with
renewable energy (wind, PV, and FC), the use of the artificial intelligent techniques in control
and energy saving of electrical machines and solving electric power system problems. He can
be contacted at email: melkholy71@yahoo.com.

Revolutionizing Motor Maintenance A Comp

Uploaded by

Copyright:

Available Formats

Revolutionizing Motor Maintenance A Comp

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Revolutionizing Motor Maintenance A Comp

Uploaded by

Copyright:

Available Formats

International Journal of Power Electronics and Drive Systems (IJPEDS)

Vol. 15, No. 3, September 2024, pp. 1968~1989

Revolutionizing motor maintenance: a comprehensive survey of

Bahgat Hafez Bahgat1, Enas A. Elhay1, Tole Sutikno2,3, Mahmoud M. Elkholy1

Article Info ABSTRACT

Journal homepage: http://ijpeds.iaescore.com

2. THE COMPREHENSIVE THEORETICAL BASIS OF IM FAULTS

Induction Motor Faults

Internal Faults External Faults

Electrical Mechanical Electrical Mechanical Environmental

Figure 1. Classification of the most frequent induction motor faults [16]

2.1. Eccentricity faults

Table 1. IM faults, causes, detection techniques, and effects

Revolutionizing motor maintenance: a comprehensive survey of … (Bahgat Hafez Bahgat)

Figure 2. Rotor position effect on stator

2.3. Stator or armature faults

2.4. Bearing faults

3. METHODS FOR ADVANCED FAULT DETECTION

3.1. Incremental broad learning and non-negative matrix factorization (NFM-IBL)

Figure 3. Feature extraction by IBL and NMF [53], [55]

3.1.1. Data acquisition

3.1.2. Data Processing

Revolutionizing motor maintenance: a comprehensive survey of … (Bahgat Hafez Bahgat)

3.1.3. Board learning [54]

3.1.4. Incremental broad learning [54]

3.1.5. Structure simplification [53]

3.2. Rectified stator current

3.3.1. Faulty stator winding

Figure 4. Stator winding with a turn-to-turn short

3.3.2. Eccentricity, both static, and dynamic

Figure 5. Membership functions for both input and output variables

Models featuring MCC

Models with hybrid

Figure 7. Recent developments in IM models

4.1. Models featuring coupled circuits

4.1.1. Multiple coupled circuit models

[𝑄𝑠] = [𝐿5𝑠 ][𝐼𝑠 ] + [𝐿5𝑟 ][𝐼𝑟 ] (6)

[𝜙𝑟 ] = [𝑙𝑟𝑠 ][𝐼𝑠 ] + [𝐿𝑟𝑟 ][𝐼𝑟 ] (8)

4.1.2. d-q model

4.1.3. Models based on magnetic circuits (MEC)

[𝜙] = [𝐹][𝑅] (11)

4.1.4. Models based on mathematical procedures (FEM)

4.2. Models with hybrid components

5. CHALLENGES AND FUTURE DIRECTIONS

Revolutionizing motor maintenance: a comprehensive survey of … (Bahgat Hafez Bahgat)

Bahgat Hafez Bahgat is a M.Sc. student in Electrical Engineering Department

Revolutionizing motor maintenance: a comprehensive survey of … (Bahgat Hafez Bahgat)

You might also like