Diagnostics of Rotor and Stator

Problems in Industrial Induction Motors

by

Fang Duan
B.E. (Telecommunication Engineering),
Southwest Jiaotong University, China, 2005

Thesis submitted for the degree of

Masters of Engineering Science

in

School of Electrical and Electronic Engineering

The University of Adelaide, Australia

August 2010
 © 2010
Fang Duan
All Rights Reserved
Contents

Contents iii

Abstract vii

Statement of Originality ix

Acknowledgments xi

Abbreviations xiii

Symbols xv

Publications xvii

List of Figures xix

List of Tables xxiii

Chapter 1. Introduction 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Construction of Three-phase Induction Machines . . . . . . . . . . . . . . 2
1.2.1 Types of Stator Winding Construction . . . . . . . . . . . . . . . . 4
1.2.2 Stator Winding Insulation System Features . . . . . . . . . . . . . 5
1.3 On-line Motor Condition Monitoring . . . . . . . . . . . . . . . . . . . . . 8
1.4 Literature Review of Motor Faults . . . . . . . . . . . . . . . . . . . . . . 9
1.4.1 Broken Rotor Bar Faults . . . . . . . . . . . . . . . . . . . . . . . . 11
1.4.2 Stator Related Faults . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.5 Research Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Chapter 2. Induction Motor Faults Diagnosis Methods 17

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Page iii
Contents

2.2 Motor Current Signature Analysis Method . . . . . . . . . . . . . . . . . 18

2.3 Zero Crossing Time (ZCT) Method . . . . . . . . . . . . . . . . . . . . . . 19
2.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Chapter 3. Baseline Analysis of Induction Motor 23

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.2 Baseline Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Chapter 4. Alignment Measurement Analysis 35

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.2 Alignment Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.3 Comparison of the On-line Condition Monitoring Testing Results . . . . 40
4.3.1 Healthy Motor Testing . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.3.2 The Misalignment Influence on Stator Short Circuit Detection . . 42
4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

Chapter 5. Stator Short Circuit Fault 47

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
5.2 Detection of Stator Short Circuit . . . . . . . . . . . . . . . . . . . . . . . . 49
5.3 Experiment Setup for Stator Short Circuit Faults Diagnosis . . . . . . . . 50
5.4 Off-line Test of Stator Short Circuit Faults . . . . . . . . . . . . . . . . . . 51
5.4.1 Analysis of Turn-to-turn Fault of Motor I . . . . . . . . . . . . . . 51
5.4.2 Analysis of Phase-to-phase Fault in Motor I . . . . . . . . . . . . . 58
5.5 On-line Test of Stator Short Circuit Faults . . . . . . . . . . . . . . . . . . 64
5.5.1 Analysis of Turn-to-turn Fault of Motor II . . . . . . . . . . . . . . 64
5.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

Chapter 6. Broken Rotor Bar Fault 69

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.2 Broken Rotor Bar Fault Detection Techniques . . . . . . . . . . . . . . . . 70
6.3 Analysis of Broken Rotor Bars . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

Page iv
 Contents

Chapter 7. Thesis Summary and Recommendation for Future Work 81

7.1 Thesis Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

References 83

Appendix A. MATLAB Code 89

A.1 Data Analysis Algorithm for Broken Rotor Bar Detection (umc100current.m) 89
A.2 Single-phase Zero Crossing Times Calculation (umc100zc.m) . . . . . . . 92
A.3 Three-phase Zero Crossing Time Calculation (zcrinterp3phase.m) . . . . 93

Appendix B. Test Results of Stator Short Faults before Misalignment Adjust 95

Appendix C. Test Results of Stator Short Faults after Misalignment Adjust 111

Appendix D. Induction motor parameters 127

D.1 Data of the Motor Type SZJKe 14a . . . . . . . . . . . . . . . . . . . . . . 127
D.2 Data of the Motor Type Sg112M-4 . . . . . . . . . . . . . . . . . . . . . . . 128

Appendix E. Test Results of Induction Motor Broken Rotor Bar Fault 131

Page v
Page vi
Abstract
In this project, two kinds of induction motor faults, stator short circuit fault and bro-
ken rotor bar fault, are investigated by using motor current signature analysis (MCSA)
and zero crossing time (ZCT) method. These methods are based on the detection of
sidebands around the supply frequency in the stator current signal.

The thesis starts by a review of these two common faults and two commonly used
diagnostic methods. Before the motor stator short circuit faults experiments, baseline
analysis is carried out on two same types of healthy motors. Meanwhile, signal pro-
cessing programs, composed in MATLAB and LABVIEW, are verified to ensure the
accurate diagnosis of motor faults. Through a control box, artificial turn to turn fault
and phase to phase fault are structured in each test. MCSA and ZCT are utilized to
extract broken rotor bar information from recorded stator current signal.

Although an induction motor is highly symmetrical, it may still have a detectable sig-
nal component at the fault frequencies due to imperfect manufacture, improper motor
installation and so on. The misalignment experiments reveal that improper motor in-
stallation could lead to an unexpected frequency peak, which will affect motor fault
diagnosis. Furthermore, manufacture tolerance and working environment could also
result in disturbing the motor fault diagnosis.

Through both online and offline experiments, MCSA and ZCT methods could detect
particular abnormal harmonics related to stator short circuit fault and broken rotor bar
fault. Compared with the conventional MCSA method, the ZCT method has the ad-
vantage of reduced computational burden.

Page vii
Page viii
Statement of Originality
This work contains no material that has been accepted for the award of any other de-
gree or diploma in any university or other tertiary institution and, to the best of my
knowledge and belief, contains no material previously published or written by an-
other person, except where due reference has been made in the text.

I give consent to this copy of the thesis, when deposited in the University Library,
being available for loan, photocopying, and dissemination through the library digital
thesis collection, subject to the provisions of the Copyright Act 1968.

I also give permission for the digital version of my thesis to be made available on the
web, via the University’s digital research repository, the Library catalogue, the Aus-
tralasian Digital Thesis Program (ADTP) and also through web search engines, unless
permission has been granted by the University to restrict access for a period of time.

Signed Date

Page ix
Page x
Acknowledgments
I would like to gratefully acknowledge my supervisor, Dr Rastko Zivanovic, whose
patience and kindness, as well as his academic experience, have been invaluable to
me. I also wish to express my appreciation to co-supervisor, Dr Said Al-Sarawi, for his
valuable suggestions and constructive advice.

In particular many thanks go to my industrial co-supervisors, Dr Abhisek Ukil and Dr

Andrea Andenna in ABB Corporate Research Center in Dättwil, Baden Switzerland for
their time commitment and valuated knowledge that allowed me complete this work
with the best possible results. Also, I would like to extend my thanks to the scientists
and intern students at Integrated Sensor System Group in ABB for their help and sup-
port.

Last, but not least, my parents for their continues support emotionally and financially
during my postgraduate years, without their support this thesis would not be possible.

Page xi
Page xii
Abbreviations

AI artificial intelligence
BRB broken rotor bars
DFT discrete Fourier Transform
DSP digital signal processor
FFT fast Fourier Transform
FT Fourier Transform
ICA Independent component analysis
LV low voltage
MCSA motor current signature analysis
MMF magnetic motive force
PCA Principal component analysis
RPS reconstructed phase space
ZCT zero crossing times

Page xiii
Page xiv
Symbols

f0 fundamental frequency
f comp component frequency
fr rotor frequency
fs supply frequency
f short the frequency component that is related to the short circuit
Ip currents in the primary
Is currents in the secondary
np number of turns in the primary
ns number of turns in the secondary
N number of samples
p number of pole
P number of pole pairs
s slip
Sr rotor speed
T (n) the time when the current is equal to zero
TZC ZCT signal
ωe angular speed of the stator magnetic motive force in electrical radians per second
ωr the angular frequency of rotation of the rotor shaft
ωrm rotor rotating speed
ωsm the synchronous speed in mechanical radians

Page xv
Page xvi
Publications
Referred Conference Publications (Full Paper)

[1] F. Duan and R. Zivanovic, “Estimation of DC Offset Parameters based on

Global Optimization,” AUPEC’08 - 18th Australasian Universities Power Engineering
Conference, Sydney, Australia, December 14–17 2008.

[2] F. Duan and R. Zivanovic, “Induction motor fault diagnostics using global
optimization algorithm,” AUPEC’09 - 19th Australasian Universities Power Engi-
neering Conference: Sustainable Energy Technologies and Systems, Adelaide, Australia,
September 27–30 2009.

Page xvii
Page xviii
List of Figures

1.1 Squirrel cage motor and rotor winding. . . . . . . . . . . . . . . . . . . . 3

1.2 Random-wound stator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Form-wound stator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Cross section of a random stator winding slot. . . . . . . . . . . . . . . . 7
1.5 Cross section of a form-wound stator winding slot. . . . . . . . . . . . . 7
1.6 Major causes of 100 large accidents in petrochemical plants. . . . . . . . 8
1.7 The on-line condition monitoring process. . . . . . . . . . . . . . . . . . . 8
1.8 Motor faults diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.9 Flowchart of the stator short circuit detection algorithm using the ZCT
signal. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.1 Ideal stator current of a three phase motor. . . . . . . . . . . . . . . . . . 20

2.2 Method of calculating approximate zero points. . . . . . . . . . . . . . . 21

3.1 Healthy induction Motor I without load. . . . . . . . . . . . . . . . . . . . 26

3.2 Healthy induction Motor II without load. . . . . . . . . . . . . . . . . . . 27
3.3 Healthy induction Motor I with 80% load. . . . . . . . . . . . . . . . . . . 27
3.4 Healthy induction Motor II with 80% load. . . . . . . . . . . . . . . . . . 28
3.5 Healthy induction Motor I with 90% load. . . . . . . . . . . . . . . . . . . 28
3.6 Healthy induction Motor II with 90% load. . . . . . . . . . . . . . . . . . 29
3.7 Healthy induction Motor I with 110% load. . . . . . . . . . . . . . . . . . 29
3.8 Healthy induction Motor II with 110% load. . . . . . . . . . . . . . . . . . 30
3.9 Healthy induction Motor I with 160% load. . . . . . . . . . . . . . . . . . 30
3.10 Healthy induction Motor II with 160% load. . . . . . . . . . . . . . . . . . 31
3.11 Computed slip vs. I/Inom for Motor I. . . . . . . . . . . . . . . . . . . . . 31
3.12 Computed slip vs. I/Inom for Motor II. . . . . . . . . . . . . . . . . . . . . 32
3.13 Amplitude of rotor frequency vs. I/Inom . . . . . . . . . . . . . . . . . . . 33

4.1 Motor foots misalignment. . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Page xix
List of Figures

4.2 Fixturlaser. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.3 Measurement unit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.4 Measurement process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.5 Measurement results of the Motor II. . . . . . . . . . . . . . . . . . . . . . 39
4.6 Measurement results of the Motor III. . . . . . . . . . . . . . . . . . . . . 39
4.7 ZCT signal of Motor III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.8 Coupling adjustment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.9 ZCT signal of Motor III (after coupling adjustment). . . . . . . . . . . . . 41
4.10 The test results of Motor III. . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.11 Misalignment effect for turn-to-turn stator short circuit directly. . . . . . 43
4.12 Misalignment effect for turn-to-turn stator short circuit with resistance. . 44
4.13 Misalignment effect for phase-to-phase stator short circuit directly. . . . 44
4.14 Misalignment influence for phase-to-phase stator short circuit with re-
sistance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.15 Comparison of misalignment measurement testing results. . . . . . . . . 45

5.1 Wye connected stator showing possible short circuit faults. . . . . . . . . 48

5.2 Off-line and on-line detection structure diagram. . . . . . . . . . . . . . . 52
5.3 The laboratory setup for induction motor stator short circuit experiments. 53
5.4 Healthy motor with 80% load level. . . . . . . . . . . . . . . . . . . . . . . 54
5.5 Healthy motor with 140% load level. . . . . . . . . . . . . . . . . . . . . . 55
5.6 Turn-to-turn short circuit with 80% loading level. . . . . . . . . . . . . . 55
5.7 Turn-to-turn short circuit with 140% loading level. . . . . . . . . . . . . . 56
5.8 Slip with two and five turns faults. . . . . . . . . . . . . . . . . . . . . . . 57
5.9 The amplitude of rotor frequency under different load levels. . . . . . . . 58
5.10 Phase-to-phase fault with 80% loading level. . . . . . . . . . . . . . . . . 59
5.11 Phase-to-phase fault with 140% loading level. . . . . . . . . . . . . . . . . 60
5.12 Slip of phase-to-phase short circuit and turn-to-turn short circuit under
different load level. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.13 The amplitude of rotor frequency under different load levels. . . . . . . . 62
5.14 The amplitude of rotor frequency of phase-to-phase fault and turn-to-
turn fault under different load levels. . . . . . . . . . . . . . . . . . . . . . 63

Page xx
 List of Figures

5.15 The amplitude of rotor frequency of serious faults under different load
level. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.16 Comparison of testing results of turn-to-turn stator short circuit without
resistance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.17 Comparison of testing results of turn-to-turn stator short circuit with
resistance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.18 Comparison of testing results of phase-to-phase stator short circuit with-
out resistance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.19 Comparison of testing results of phase-to-phase stator short circuit with
resistance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.20 Temperature of turn-to-turn short circuit fault. . . . . . . . . . . . . . . . 67

6.1 Current and ZCT signal from healthy motor (81.36% load level). . . . . . 72
6.2 Direct recorded ZCT signal from healthy motor (81.36% load level). . . . 73
6.3 Current and ZCT signal from healthy motor (101.36% load level). . . . . 73
6.4 Direct recorded ZCT signal from healthy motor (101.36% load level). . . 74
6.5 Current and ZCT signal from motor with one BRB. . . . . . . . . . . . . 75
6.6 Direct recorded ZCT signal from motor with one BRB. . . . . . . . . . . 75
6.7 Current and ZCT signal from motor with two BRB. . . . . . . . . . . . . 76
6.8 Direct recorded ZCT signal from motor with two BRB. . . . . . . . . . . 76
6.9 The relationship between slip and load level. . . . . . . . . . . . . . . . . 79
6.10 The relationship between the amplitude of rotor frequency and load level. 79

Page xxi
Page xxii
List of Tables

1.1 Thermal classification of rotating machine insulation material. . . . . . . 8

3.1 ABB induction motor parameters. . . . . . . . . . . . . . . . . . . . . . . 25

3.2 Difference between Motor I and Motor II. . . . . . . . . . . . . . . . . . . 31
3.3 Online testing results for Motor I (recording directly by LabVIEW). . . . 32
3.4 Online testing results for Motor II (recording directly by LabVIEW). . . . 32

4.1 Torque levels corresponding current load (Motor III). . . . . . . . . . . . 41

4.2 Torque levels corresponding current load (Motor II). . . . . . . . . . . . . 43

5.1 Turn-to-turn fault (two turns). . . . . . . . . . . . . . . . . . . . . . . . . . 56

5.2 Turn-to-turn fault (five turns). . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.3 Phase-to-phase fault. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.4 Phase-to-phase fault plus turn-to-turn fault. . . . . . . . . . . . . . . . . . 61

6.1 Motor parameters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

6.2 Symmetrical motor measurement results. . . . . . . . . . . . . . . . . . . 77
6.3 One broken rotor bar fault measurement results. . . . . . . . . . . . . . . 77
6.4 Two broken rotor bar fault measurement results. . . . . . . . . . . . . . . 78
6.5 Three broken rotor bar fault measurement results. . . . . . . . . . . . . . 78

B.1 Healthy motor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

B.2 1,2 short directly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
B.3 1,2 short with resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
B.4 1,3 short directly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
B.5 1,3 short with resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
B.6 2,3 short directly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
B.7 2,3 short with resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
B.8 1,2,3 short directly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

Page xxiii
List of Tables

B.9 1,2,3 short with resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

B.10 B1-1,B2-1 short directly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
B.11 B1-1,B2-1 short with resistance . . . . . . . . . . . . . . . . . . . . . . . . 106
B.12 B1-2,B2-1 short directly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
B.13 B1-2,B2-1 short with resistance . . . . . . . . . . . . . . . . . . . . . . . . 108
B.14 B1-3,B2-1 short directly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

C.1 1,2 short directly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112

C.2 1,2 short with resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
C.3 1,3 short directly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
C.4 1,3 short with resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
C.5 2,3 short directly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
C.6 2,3 short with resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
C.7 B1-1,B2-1 short directly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
C.8 B1-1,B2-1 short with resistance . . . . . . . . . . . . . . . . . . . . . . . . 119
C.9 B1-2,B2-1 short directly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
C.10 B1-2,B2-1 short with resistance . . . . . . . . . . . . . . . . . . . . . . . . 121
C.11 1,2,3 short directly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
C.12 1,2,3 short with resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
C.13 On line test with coupling gap . . . . . . . . . . . . . . . . . . . . . . . . . 124
C.14 On line test without coupling gap . . . . . . . . . . . . . . . . . . . . . . . 125

D.1 Ratings of motor SZJKe 14a . . . . . . . . . . . . . . . . . . . . . . . . . . 127

D.2 Equivalent circuit parameters . . . . . . . . . . . . . . . . . . . . . . . . . 128
D.3 Data of bearing SKF type 6304 ZZ CXSQ . . . . . . . . . . . . . . . . . . . 128
D.4 Ratings of motor Sg112M-4 . . . . . . . . . . . . . . . . . . . . . . . . . . 129
D.5 Data of bearing 6306-2Z . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
D.6 Equivalent circuit parameters . . . . . . . . . . . . . . . . . . . . . . . . . 129

E.1 Symmetrical Motor I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

E.2 1 broken rotor bar in Motor I . . . . . . . . . . . . . . . . . . . . . . . . . 133
E.3 2 broken rotor bars in Motor I . . . . . . . . . . . . . . . . . . . . . . . . . 134

Page xxiv
 List of Tables

E.4 3 broken rotor bars in Motor I . . . . . . . . . . . . . . . . . . . . . . . . . 135

E.5 Symmetrical Motor II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
E.6 1 broken rotor bar in Motor II . . . . . . . . . . . . . . . . . . . . . . . . . 137
E.7 2 broken rotor bars in Motor II . . . . . . . . . . . . . . . . . . . . . . . . 138
E.8 3 broken rotor bars in Motor II . . . . . . . . . . . . . . . . . . . . . . . . 139

Page xxv
Page xxvi
Chapter 1

Introduction

T
HIS chapter presents background knowledge of three-phase in-
duction motors. Also review two types of common faults that
are related to diagnostic methods. There types are broken rotor
bar and stator short circuit fault. In addition, the objectives of the research
project are summarized in the chapter.

Page 1
1.1 Introduction

1.1 Introduction

Induction motors, also named asynchronous motor, are widely used in manufacturing,
transportation, petrochemical, mining, power systems and so on due to their reliability
and simplicity of construction, high overload capability and high efficiency. The range
size of induction motors are from tiny to over 100,000 horsepower. Compared with
direct-current motors, induction motors are more rugged, less expensive and require
less maintenance. Therefore, they are the preferred choice for industrial motors.

The induction motors were invented by Nikola Tesla in 1888. Their rotating parts do
not require connecting electricity because electromagnetic induction will provide the
transfer of energy from stationary parts to the rotating rotating counterpart. The sta-
tionary winding named stator can produce a rotating magnetic field, which induces
an alternating electromotive force and current in the rotor. The induced rotor-current
and the rotating field of the stationary winding interact and produce a motor torque.
The characteristic of induction motor torque-speed is related to the resistance and reac-
tance of the rotor. Therefore, with different ratios of rotor resistance to rotor reactance
in rotor circuits, it is possible to achieve different torque-speed characteristics [1].

1.2 Construction of Three-phase Induction Machines

For three phase induction machines, the stator consists of three wire coils fixed inside
a metal frame and spaced 120 electrical degrees apart. The end shields and termi-
nal box are mounted onto the metal frame in order to bolt the frame and for making
connections to the power cable. The center of the end shields are the bearing and it
can support the shaft of the rotor. The magnetic core of the stator, that accommo-
dates the three distributed stator windings, is formed by stacking thin electrical steel
laminations with uniformly spaced slots stamped in the inner circumference [1]. For
example, the steel lamination is about 0.5 mm thick for 60 Hz machines. The stator
windings are composed by connecting copper or aluminum conductor coils. The coils
are insulated from the slot walls. The axes of the stator windings of a P-pole machine
are spaced (2/P)(2π/3) mechanical radians apart, with each phase belt occupying the
same number of slots. The terminals of the three stator phase windings can be wye (Y)

Page 2
Chapter 1 Introduction

or delta-connected (∆) [2].

The rotor consists of a cylindrical laminated iron core with uniformly spaced periph-
eral slots for accommodating the rotor winding. The rotor windings of a three-phase
induction machine can take two types: wound and squirrel-cage rotor [1]. For a wound
rotor machine, rotor windings are wound on a cylindrical laminated core with uni-
formly spaced slots on the outer periphery and they are similar as the distributed
windings on the stator. The terminals of the rotor windings can be brought out via
slip rings and brushes. The slip rings are insulated from the shaft and others. In order
to acquire higher starting torque or limited control of speed under synchronous speed,
the terminals of the rotor windings can be connected to varying external resistance
through slip rings and brushes.

NOTE:
This figure is included on page 3 of the print copy of
the thesis held in the University of Adelaide Library.

Figure 1.1. Squirrel cage motor and rotor winding.

(Ref: http://www.ieee-kc.org/library/motors/motorslip.htm)

Figure 1.1 shows the features of a squirrel-cage motor and the rotor winding. The mo-
tor consists of copper or aluminum axial bars with uniform spaces that are soldered
onto end short-circuit rings at both ends. The rotor core consists of an iron laminations
mold with molten aluminum. There is no insulation between the iron and the con-
ductors, and none is needed. In addition, the current, which is induced in the rotor, is

Page 3
1.2 Construction of Three-phase Induction Machines

contained within the circuit formed by the conductors and end-ring, also called end-
connections. Skewing the rotor slots can help avoid crawling and reduces vibration [2].
Due to simplicity of construction, high overload capability, high efficiency, low main-
tenance expense and low cost, the squirrel cage motor is widely used in industry.

The transfer of energy from the stator to the rotor, either squirrel cage or wound rotor,
is by means of electromagnetic induction. The stator is often referred to as the primary
and the rotor as the secondary. Energy is transferred electromagnetically across the
air gap between the stator and the rotor windings. A small air gap can help to reduce
reluctance. However, the width of the air gap is constrained by allowable core losses
as well as manufacturing tolerances and costs. The air gap is typically between 0.35 to
0.50 mm for motors up to 10kW, and between 0.5 to 0.8 mm for motor in the range of
10 to 100 kW [1].

1.2.1 Types of Stator Winding Construction

There are two basic types of stator winding structures namely random-wound stators
and form-wound stators using multi-turn coils and form-wound stators using Roebel
bars. The random-wound stators are usually utilized for no more than several hundred
kW of machines. Furthermore, the form-wound stators are used for larger motors and
some generators rated up to between 50 MVA and 100MVA. Roebel bar windings are
usually used for large generators [3].

Random-wound Stators

Random-wound stators are comprised of insulated copper conductors, winding wire

or magnet wire, which is round and the wound is continuously through slots in the
stator core to form a coil by a winding machine or by hand [1]. In inserting random-
wound coils, each turn or loop of winding wire could be placed against any other turn
and independent of the voltage level of the turn, as shown in the Figure 1.2. Thus, the
windings are called ”random”. During stator winding, the coils are separated using an
insulating material that is positioned to clear the inside diameter of the stator. Random
windings are used on lower kW machines.

Form-wound Stator

Page 4
Chapter 1 Introduction

NOTE:
This figure is included on page 5 of the print copy of
the thesis held in the University of Adelaide Library.

Figure 1.2. Random-wound stator.

(Ref: http://www.schulzelectric.com/)

Form-wound stators windings usually consist of an insulated coil that have been pre-
formed, such as square or rectangular magnet wire, before they are put in the slots, as
shown in the Figure 1.3. The coil winding process begins with looping of the magnet
wire. Due to the difficulty of forming wire of high width-to-thickness ratio, several
wires in parallel may make one turn [3]. Through careful design and manufacturing
process, an individual turn in a coil is arranged in precise location and adjacent to an-
other turn with the smallest possible voltage difference. Usually, each coil includes
from two to twelve turns. The insulation of the wire is designed to handle operating
turn-to-turn voltages as well as maximum surge or impulse voltages. They are usually
operated at 1000V.

1.2.2 Stator Winding Insulation System Features

The stator winding insulation system plays an important role in induction motors. A
well designed stator winding insulation system can prevent the electrical short [3].
There are several components and features in a stator winding insulation system, such
as strand (or subconductor) insulation, turn insulation and groundwall (or ground or
earth) insulation. Figures 1.4 and 1.5 show a cross section of a random wound and
form-wound coils in a stator slot. The figures illustrate the location of the insulation

Page 5
1.2 Construction of Three-phase Induction Machines

NOTE:
This figure is included on page 6 of the print copy of
the thesis held in the University of Adelaide Library.

Figure 1.3. Form-wound stator.

(Ref: http://www.schulzelectric.com/)

components in the different motors. For example, the turn insulation is used for pre-
venting shorts among the turns in a coil. When a turn short circuit happens, the shorted
turn will appear as the secondary winding of an autotransformer [3]. The transformer
law is explained as

n p I p = ns Is (1.1)

where n p and ns are number of turns in the primary and secondary, respectively, I p
and Is are the currents in the primary and secondary, respectively. Therefore, the over-
heating will appear when a big value of current flow appears in the shorted turn that
will affect the insulation life of other healthy turns and stator winding.

The electrical insulation material decides the lifetime of insulation. For different pur-
poses of insulation in electric equipment, there are several designs and combinations to
perform designed functions. The definitions of the thermal identification and classifi-
cation of insulating materials and insulation system are published by IEEE and IEC [3].
Table 1.1 shows the accepted temperature indices of materials according to IEC 60085.
The motors used in the thesis belong to the class F.

Page 6
Chapter 1 Introduction

NOTE:
This figure is included on page 7 of the print copy of
the thesis held in the University of Adelaide Library.

Figure 1.4. Cross section of a random stator winding slot [3].

NOTE:
This figure is included on page 7 of the print copy of
the thesis held in the University of Adelaide Library.

Figure 1.5. Cross section of a form-wound stator winding slot [3].

Page 7
1.3 On-line Motor Condition Monitoring

Table 1.1. Thermal classification of rotating machine insulation material.

Numerical Classification Letter Classification Temperature ( )‰

105 A 105
130 B 130
155 F 155
180 H 180

1.3 On-line Motor Condition Monitoring

On-line motor condition monitoring detects and monitors the operating characteristics
of a motor while it is running. The major advantage of on-line condition monitoring
is that the device’s potential problems could be detected early before a serious dete-
rioration or breakdown occurs. Taking petrochemical plants as an example, Figure
1.6 illustrates that mechanical fault is a major part of the various faults [4]. Thus, in-
creasing research has been developed in on-line condition monitoring techniques for
mechanical equipment, especially electrical devices, such as induction motors, gener-
ators and transformers in power plants.

NOTE:
This figure is included on page 8 of the print copy of
the thesis held in the University of Adelaide Library.

Figure 1.6. Major causes of 100 large accidents in petrochemical plants [4].

An on-line condition monitoring system includes four main parts, which are sensors,
data acquisition, fault detection, and fault diagnosis [5]. A typical on-line condition
monitoring system process is shown in Figure 1.7.

Sensors Data Acquisition Fault Detection Fault Diagnosis

Figure 1.7. The on-line condition monitoring process.

Page 8
Chapter 1 Introduction

Sensors measure physical quantities of the device and convert them into an electrical
response to time and to be read by an oscilloscope or other measuring instruments.
There are many types of sensors that are utilized in on-line condition monitoring, for
example, current sensors, Hall Effect sensors, voltage sensors, flux sensors, vibration
sensors, speed sensors and temperature sensors. The selection of a specific sensor type
depends on the method of monitoring. The common used monitoring methods are
voltage, current, temperature, deviation of magnetic potential vector and asymmetri-
cal magnetic flux distribution, gyration radius, torque and speed fluctuation and so on.

The function of data acquisition is sampling and amplifying the physical signal data
acquired from the sensors. Then, the data is converted into feature values that re-
flect the device’s operating condition. A data acquisition system includes three main
components, which are physical parameters to electrical signal converter, a signal con-
ditioning circuit and analog to digital signal converter.

The purpose of fault detection is finding out the incipient fault appearing in the ma-
chine through processing the acquired data. The signal data which is sensitive of the
machine fault is extracted and used for further analysis. The fault detection system
includes two different methods, named model-referenced method and feature extrac-
tion [5]. The model-referenced method is based on mathematical simulation models
or artificial intelligence to compare and analyze measurement results. The feature ex-
traction method extracts information from the processed signal data, such as exceeded
thresholds, amplitudes and frequency, and decides whether the machine’s operating
state is faulty or healthy.

The common machine faults can be broadly divided into two types, stator faults and
rotor faults. Figure 1.8 illustrates the faults and related elements [6]. The diagnosis
results can clearly indicate the direction of maintenance.

1.4 Literature Review of Motor Faults

Motor faults may have only minor symptoms but result in lower efficiency and hence
higher energy consumption, poorer performance and long term degradation. Even

Page 9
1.4 Literature Review of Motor Faults

Unbalanced Magnetic Pull

Coil Movement Defective Installation
Unbalanced Supply Voltage Incorrect Loading
Over Loading Unbalanced Magnetic Pull
Frane Vibration Bearing Failure
Defective Installation Over Heating Loss of Lubrications
Rotor Strike Contamination Oil

Turn to Body Fault Thermal Cycling

Abrasion of Insulation Stator Earth Unbalanced Magnetic Pull
Rotor Broken
Thermal Cycling Faults Common Large Transient
Bars
Slack Core Laminations Machine Faults Negative Sequence in Supply
Slack Slot Wedges
Shock

Damage During Installation or Service Damage of Rotor Body Defective Rotor Forgoing or Fabrication
Frequent Starting Insulation Failure Thermal Cycling
High or Low Humidity and Temperature Over Torque

Excessive Heating Incorrect Coupling

Surge in Supply Stator Large Bearing Clearance
Bearing
High Humidity Turn-Turn Stator Fault Rotor Fault Over Loading
Misalignment
Vibration Faults Defective Installation
Contamination by Oil

Failure in Installation Defective Installation

Stator Phase- Rotor
High Temperature Bearing Failure
Phase Faults Misalignment
Unbalanced Supply Unbalanced Magnetic Pull
Slacking of Coils

Shock Due to Faults Over Heating

Displacement of Bearing Loss of
Repeated Starting Oil Seal Failure
Conductor lubrication
End Winding Vibration Contamination Oil

Rotor
Slack Joints Failure of
Mechanical or End Ring Movement
Contamination Electrical
Thermal Asymmetric Blocking
Excessive Vibrations Connections
Unbalanced

Figure 1.8. Motor faults diagram [6].

small faults can cause increased losses such as reducing efficiency, increasing tempera-
ture, which will reduce insulation lifetime, and increasing vibration, which may reduce
bearing lifetime [7]. They are caused by the operating environment condition and the
machines internal factors. Therefore, the diagnosis of induction motor problems is
important and prevent expensive maintenance cost. According to published surveys,
induction motor failures include bearing related faults (which are responsible for 40%
of all faults), short circuits in stator windings (which represent 38% of the reported
faults), rotor related faults (which represent 10% of the induction motor faults) and
other faults such as end ring faults (which occupy 12% of the reported faults) [8].

Numerous methods of induction motor fault diagnosis were developed in the last
decades and many techniques have been proposed [9,10]. The most common approach
is motor current signature analysis (MCSA) [11]. Several of induction motor faults de-
tection and identification techniques are based on Fast Fourier Transform spectral sig-
nature analysis [12–14]. Other techniques include vibration analysis, temperature mea-
surements, harmonic analysis of speed fluctuations [15, 16], vibration monitoring [17],

Page 10
Chapter 1 Introduction

state and parameters estimation [18], either axial flux or air-gap torque analysis [19],
acoustic noise measurement, and magnetic field analysis [20–23]. Currently, more and
more new techniques based on artificial intelligence (AI) have been utilized for diag-
nostic induction motor faults, such as fuzzy logic [24, 25], genetic algorithms [22, 26],
neural network [27], Bayesian classifiers [28], and envelope of the three-phase stator
current with AI based on Gaussian mixture models and reconstructed phase spaces
(RPSs) [29].

The methods used to diagnose broken bars can be broadly categorized into two types,
invasive and non-invasive fault diagnosis techniques [30]. The invasive methods di-
agnose broken bars by monitoring the deviation of the magnetic potential vector and
asymmetrical magnetic flux distribution, gyration radius, asymmetrical magnetic flux
distribution, torque and speed fluctuation and so on. In these methods, sensors and
costly measurement equipments are needed, which will increase the complexity and
expense of the diagnostic method. Thus, in this literature review we are only concern
with non-invasive fault diagnosis techniques.

1.4.1 Broken Rotor Bar Faults

K. R. Cho proposed the use of state and parameter estimation for diagnosing broken
rotor bars in induction motors in 1989 [18]. This detection method is based on the hy-
pothesis that rotor resistance of an induction motor will increase when a rotor broken
bar occurs. In Cho’s experiments, the induction motor working condition is that the
apparent rotor resistance should be in the balanced single-phase electrical model when
motor operates at constant velocity. During the detection process, the stator voltage,
current, excitation frequency, and rotor velocity are measured, which are taken over a
small range of velocity. The near-least-square-error estimator was used for the mea-
surements in order to estimate motor states and parameters.

In 2001, Thomson and Fenger diagnosed induction motor faults by using current sig-
nature analysis [31]. The authors used a current transformer to measure stator current
and a sensor to collect the transformed signal. Then, the signal is transformed to the
frequency domain for spectral analysis by using a signal processing box with Discrete
Fourier Transform (DFT), low-pass filter and analogue/digital converter. The stator

Page 11
1.4 Literature Review of Motor Faults

current spectrum showed abnormal harmonics, which were related to induction mo-
tor faults. However, the load of the machine is dynamic, which may lead to changes in
current signal frequency. If the frequency is very close to the fundamental frequency,
the broken bar fault can not be detected exactly. Therefore, the application of some
algorithms to process the spectral signal is required.

In order to overcome the limitations, Çalış and Çaklır in 2007 [32] applied the zero
crossing times (ZCT) technique to diagnose the rotor’s broken bars. This method only
requires the use of a current sensor, which has already been installed in most of mo-
tors. The method diagnose broken bars by observing changes in the motor current
zero crossing instants. If motor runs at constant speed under balanced supply and sta-
tor winding, the spacing between successive zero crossings is equal. The fluctuations
in the motor ZCT will appear when broken bar failure occurs. Through an analysis of
ZCT data, motor slip and motor status information are estimated. The results show
that the method is both sensitive and low cost. The method is based on the assump-
tion that the motor runs at constant speed under balanced supply. However, in real
working environment, the conditions are not always satisfied, which may limit its ap-
plication.

1.4.2 Stator Related Faults

In 1978, Bonnett identified that five failure modes in Wye connected stator, as

A) turn-to-turn short circuit fault

B) coil-to-coil short circuit fault

C) open circuit fault

D) phase-to-phase short circuit fault

E) phase-to-ground short circuit fault

In addition, he identified that one of these modes or any combination of them can cause
a fault. The turn-to-turn fault is the initial fault, and other faults can be regarded as the

Page 12
Chapter 1 Introduction

results of it. In the early stage, turn-to-turn fault does not affect motor normal oper-
ation because the voltage across the turn insulation is relatively low. However, high
substantial transient voltage might cross the insulation when motor is switched on or
off. Thus, the destruction of the turn insulation caused by the short circuit may cause
a permanent damage. Stator faults typically have the highest rate (38%) amongst the
common type of faults in industrial induction motors [31].

In the last decade, several studies of stator winding faults in synchronous and asyn-
chronous motors are developed. The common methods include vibration monitoring,
current monitoring, and axial leakage flux monitoring [33]. Recently, new techniques
are presented for online motor protection and diagnostics for stator winding faults,
such as integrated on-line motor protection system [34] or sensorless online motor di-
agnostics [35] and so on.

Direct measurement of ZCT or

calculation from measured stator
current

Frequency spectrum of ZCT signal

using discrete Fourier transform
(DFT)

Locating stator short indicative

frequency peaks in the ZCT spectrum

NO
Are there peaks? Healthy

YES

Stator short
circuit exists

Figure 1.9. Flowchart of the stator short circuit detection algorithm using the ZCT signal [36].

Page 13
1.5 Research Objectives

Ukil et al. used zero crossing time (ZCT) signal of the stator current to diagnose the
stator short circuit faults in 2009 [36]. The algorithm of detecting the stator short circuit
is illustrated in the Figure 1.9. Firstly, ZCT signal is acquired form the the measured
stator current. Then, the frequency spectrum of ZCT signal is calculated by using FFT.
Through the analysis of characteristic fault frequencies, the stator short circuit faults
are diagnosed. The work diagnosed stator short circuit without considering the effect
of any other problem or mechanical drive characteristic.

Compared with the conventional MCSA method, the method has advantages of low
computational requirements by using small numbers of data points without the need
for additional sensors. In addition, ZCT can be considered an effective alternative ap-
proach for smaller embedded devices for motor fault diagnosis.

1.5 Research Objectives

In this thesis, the ZCT signal is acquired by both ABB intelligent sensor and calculating
from the measured stator current in order to test ABB intelligent sensor. Furthermore,
the motor faults are diagnosed by considering potential effects, such as, improper mo-
tor installation, manufacture tolerance and working environment.

The first aim of this project is to build up the background knowledge of induction
motor fault diagnose methods and supplementary computer programs. And demon-
strate if analysis techniques such as MCSA and ZCT can be used to diagnose stator
short circuit faults and broken rotor bar faults. In addition, if it is possible to use such
techniques for either off-line and on-line for faults identification.

In order to ensure the accurate diagnosis of motor faults, a comprehensive database

of sensor measurements corresponding to a healthy motor will be developed in the
baseline study. This is the second aim of this project. Although an induction motor
is highly symmetrical, it may still have a detectable signal component at the fault fre-
quencies due to manufacture tolerance, assembling tolerance, working environment

Page 14
Chapter 1 Introduction

and so on. Therefore, it is important to study motor frequency spectrum characteris-

tics and set threshold levels to distinguish between healthy and faulty motors.

Apart from manufacture defects, improper motor installation can also reduce effi-
ciency, increase current vibration and noise. The current induced by vibration will
affect the motor faults diagnosis. Therefore, the third aim is alignment measurement
analysis. In the project, the Laser Alignment system, provided by Fixturlaser Com-
pany [37], is used to detect and adjust shaft misalignment. Furthermore, the effect of
misalignment on motor stator short circuit fault diagnosis will be studied in the exper-
iment.

Fourthly, based on previous work, the motor stator short circuit fault will be experi-
mentally studied by using both off-line and on-line tests. Through a control box, artifi-
cial turn-to-turn faults and phase-to-phase faults are simulated in each test. MCSA and
ZCT methods are employed to detect the abnormal harmonics in motor current signals.

Lastly, we will investigate whether MCSA and ZCT methods can be utilized to diag-
nose broken rotor bars fault. The recorded data is provided by the machine laboratory
of the Cracow University of Technology. The data includes stator current of a symmet-
rical motor and motor with broken rotor bars fault.

Page 15
Page 16
Chapter 2

Induction Motor Faults

Diagnosis Methods

I
N this chapter, motor current signature analysis (MCSA) method
and the zero cross time (ZCT) method are reviewed. Compared
with MCSA, ZCT has advantages of using less sampling points and
reducing computational burden and storage space.

Page 17
2.1 Introduction

2.1 Introduction
Induction motor faults often generate particular frequency components in the current
spectrum. The abnormal harmonics contain potential information of motor faults.
Therefore, the frequency analysis approach is the most commonly used method to di-
agnose induction motor faults.

2.2 Motor Current Signature Analysis Method

Motor current signature analysis (MCSA) is one of the common methods for diagnos-
ing motor faults due to its low cost and simplicity. The method is based on the detec-
tion of sidebands around the supply frequency in the stator current signal [11]. Firstly,
the motor faults are simulated to observe the spectrum of stator currents for each fault
condition. From the simulation, abnormal harmonics of stator currents are obtained as
reference signals. Secondly, the recorded stator current signal is transfered from time
domain to frequency domain by using Fourier Transform (FT). Finally, the motor faults
could be diagnosed through the comparison of recorded stator current signals and the
reference signals in the frequency domain.

In the real applications, the stator current signal is sampled at frequency f . So, the
Discrete Fourier Transform (DFT) is utilized to transfer the data from the time domain
to the frequency domain. The customary formulas exhibiting the DFT of a signal f [n]
and its inverse DFT are [38]
N −1
2πkn
F (k) = ∑ f [n] exp(−i
N
), k = 0, 1, 2, · · ·, N − 1, (2.1)
n =0

N −1
1 2πkn
f [ n] =
N ∑ F(k) exp (i
N
), n = 0, 1, 2, · · ·, N − 1, (2.2)
k=0
where N is the number of samples of f [n].

According to equation (2.1), a calculation of an N-point DFT requires ( N − 1)2 mul-

tiplications and N ( N − 1) additions, i.e. O( N 2 ) operations. Numerous Fast Fourier
transform (FFT) algorithms with O( N log N ) operations, such as Cooley-Tukey FFT al-
gorithm, Prime-factor FFT algorithm, Bruun’s FFT algorithm, Rader’s FFT algorithm

Page 18
Chapter 2 Induction Motor Faults Diagnosis Methods

and Bluestein’s FFT algorithm [38], are put forward to compute the FFT and its inverse.
In our study, the stator current signal is transferred from the time domain to frequency
domain by using FFT functions (based on Cooley-Tukey FFT algorithm) in MATLAB.

Thomson and Fenger reviewed the application of MCSA in diagnosing broken rotor
bars, shorted turns in low voltage (LV) stator windings and airgap eccentricity in three-
phase induction motor drives [31]. However, there are some handicaps which may de-
grade performance and accuracy of this fault diagnosis. The first limitation comes from
the stator current sampling and FFT, for example, noisy sensor signals, finite number
of samples and finite frequency interval in the FFT. Secondly, the current could be sam-
pled only after the motor reaches steady state. In other words, the variation of the mo-
tor may result in fluctuation in stator current frequency spectrum. Last but not least,
other unexpected factors could induce ambiguous harmonic frequencies, which will
affect the judgment of abnormal harmonics. In light of the shortcomings of MCSA, a
considerable amount of research work has been based on advanced signal-processing
techniques [39, 40].

2.3 Zero Crossing Time (ZCT) Method

The zero crossing time (ZCT) method was firstly put forward by Wang in his PhD
thesis [35]. Through monitoring successive zero crossing times, the method has been
successfully applied in speed measurement and fault detection. Recently, Çalış and
Çaklır [32] applied the ZCT technique to diagnose rotor broken bars. In their experi-
ments, only a current sensor is required, which has already been installed in most of
the motors. Thus, it reduced the experimental cost and complexity.

The method diagnoses broken bars by observing the changes in the motor current zero
crossing instants. The ZCT signal is defined as the time interval between two succes-
sive zero crossing points of the stator current, given by [35]

TZC (n) = T (n) − T (n − 1), (2.3)

where TZC is the ZCT signal, T (n) is the time when the current sample is equal to
zero. If the motor runs at constant speed under balanced supply and stator winding,

Page 19
2.3 Zero Crossing Time (ZCT) Method

the spacing between successive zero crossings is equal. The ideal stator current of a
three phase motor is plotted in the Figure 2.1. The supply frequency is 50 Hz and the
sampling frequency of the current signal is 1 kHz. In a real application, it is impossible
to find the time that the current is exactly equal to zero due to the discrete sampling.
Thus, Fig. 2.2 illustrate the algorithm for calculating approximate zero points [36].
Assume that the current is linear in a small time interval (i.e. sampling time). At
the sampling time n − 1 and n, I (n − 1) × I (n) < 0. Based on this assumption, the
following equation is given by

I (n) − 0 t(n) − T (k)

= . (2.4)
I ( n ) − I ( n − 1) t ( n ) − t ( n − 1)

Then, the approximate time T (k) can be calculated by

I (n)[t(n) − t(n − 1)]

T (k) = t(n) − . (2.5)
I ( n ) − I ( n − 1)

1.5

0.5
Amplitude (a.u.)

−0.5

−1

−1.5
0 5 10 15 20 25 30
Time (ms)

Figure 2.1. Ideal stator current of a three phase motor.

Rotor unbalance, such as broken bars, results in positive and negative sequence rotor
currents. The positive sequence rotor current produces positive sequence emfs and
positive sequence currents on the stator of angular frequency

ωr + sω = ω (1 − s) + sω = ω. (2.6)

On the contrary, negative sequence rotor current results in emfs and currents of fre-
quency in the stator,

ωr − sω = ω (1 − s) − sω = (1 − 2s)ω, (2.7)

Page 20
Chapter 2 Induction Motor Faults Diagnosis Methods

I(n)

Amplitude (a.u.)
0

I(n−1)

t(n−1) T(k) t(n)

Time (ms)

Figure 2.2. Method of calculating approximate zero points.

where s is the slip, ωr (rad/s) is the angular frequency of rotation of the rotor shaft [32].

The spacing between successive zero crossings is unequal when the motor runs under
abnormal conditions. The spectrum of the ZCT signal contains a 2s f frequency com-
ponent, which is influenced only by the negative sequence current in the rotor and its
harmonics at 2ksω, k = 2, 3, ... as frequencies of the broken bar [32]. The amplitude
of the harmonic component decreases as the value of k increases. Furthermore, the
amplitude of this spectral component rises when the the degree of rotor asymmetry
becomes serious. The motor running condition can be inspected by monitoring the 2s f
component in the ZCT spectrum.

In this thesis, both off-line and on-line tests are utilized to diagnose induction faults. In
the off-line test, the time domain current signal is recorded by an oscilloscope. Based
on equation (2.5), the ZCT signal is calculated by using MATLAB. Then, the fast Fourier
transform (FFT) algorithm is employed to analyze the ZCT signal frequency compo-
nents with sampling frequency 1000 Hz and window length 1024. In the on-line test,
the time domain ZCT signal is acquired by an ABB Intelligent Sensor. Then, time do-
main and frequency domain ZCT signals are recorded and analyzed using LABVIEW.

Compared with conventional MCSA, the ZCT method uses less sampling points. For
instance, if the supply frequency is 50Hz and the sampling frequency of the current
signal is 1 kHz, there are 201 current sampling points in one period. On the contrary,
the ZCT method gets rid of many unnecessary points. Thus, ZCT could reduce the

Page 21
2.4 Summary

computational burden in the FFT. Furthermore, ZCT needs small data storage room.
This advantage is obvious in portable devices, which always is limited by less pow-
erful digital signal processors (DSPs) and limited amount of memory and storage. In
addition, the ZCT method does not need high a sampling rate [36].

2.4 Summary
Two induction motor fault detection methods, conventional MCSA and ZCT, are de-
scribed in this chapter. Although the MCSA technique is the commonly used method
for diagnosing motor faults due to sensorless, low cost and simple experimental setup,
there are some shorting comings that limit the application of this method. The ZCT
method diagnoses motor faults by monitoring successive zero crossing times. The
method uses less sampling points and hence reduces computational burden and stor-
age space.

In our experiments, both conventional MCSA and ZCT were employed to diagnose
broken rotor bar faults and stator short circuit faults. The detailed analysis of particular
abnormal harmonics related to motor faults and the experimental setup are discussed
in chapter 3, 4, 5 and 6.

Page 22
Chapter 3

Baseline Analysis of
Induction Motor

T
HE aim of this chapter is to present the baseline study for the sta-
tor short circuit fault diagnosis. The baseline data of the stator
short circuit fault were acquired from the ABB Dättwill motor lab.
In addition, the behaviours of baseline under different load conditions are
analyzed.

Page 23
3.1 Introduction

3.1 Introduction
Baseline analysis plays an important role in the induction motor study. Through the
baseline analysis, problematic characteristics of a healthy induction motor can be de-
tected. Although an induction motor is highly symmetrical, there is likelihood of faults
in the motor and in the fault diagnosis system [41]. For example, there may be a prob-
lem of motor manufacturing or motor installation. Either of them will disturb symme-
try and influence the induction motor fault diagnosis. These problems will introduce
various frequency components or influence the amplitude of the motor fault frequency.
Consequently, determination and classification of a fault becomes difficult. Therefore,
analysis of the baseline signal can find a potential fault early and ensure the condition
monitoring reliability.

3.2 Baseline Study

In the laboratory, two new ABB motors are used and the motor parameters are tabu-
lated in Table 3.1. There are no any artificial faults created for them. In this experiment,
the device that does the ZCT signal conversion is used directly to record and analyze
the spectrum of the ZCT signal from the induction motor. Although the two motors are
of the same type, there are slight differences in the testing results. The rotor frequency
as an example is used to illustrate the difference between the Motor I and Motor II
testing results.

The rotor frequencies of two motors both with and without load were studied. In
the following figures, the time domain ZCT signals are acquired directly from ZCT
signal converter and the frequency domain ZCT signals are converted by using FFT in
MATLAB. Figure 3.1 shows the operating behavior of motor I without load. For the
the reduction of computation complexity, equation (2.3) was modified as [36]

TZC (n) = T (n) − T (n − 1) − Tre f , (3.1)

where Tre f is the natural reference time of the ZCT signal. The Y-axis of Figure 3.1
(i) represents the difference between recorded zero crossing time and natural reference
time. The rotor frequency is 24.9 Hz and the amplitude is -35.57 dB. As shown in Figure
3.2, when Motor II operates without load, the rotor frequency is 24.9 Hz and amplitude

Page 24
Chapter 3 Baseline Analysis of Induction Motor

Table 3.1. ABB induction motor parameters.

Parameter Value
Active power [kW] 2.2
Nominal voltage [V] 380-420
Nominal current [A] 4.9
Nominal frequency [Hz] 50
Nominal power factor cos ϕ N [-] 0.81
Rotor speed [rpm] 1430
Winding connection Y
Number of poles per phase winding p [-] 4
Insulation Classification F∗

is -33.09 dB. In theory, the peak values of the two motors should not have much dif-
ference in light of the same detection system and the same type of motor. However,
the amplitude value of Motor I is smaller than that of Motor II. Therefore, there may
be some potential problems in Motor II. Figure 3.3 and 3.4 illustrate the performance
of Motor I and II with low load level, respectively. Although the two motors have the
same rotor frequencies, the amplitude of Motor I is still smaller than that of Motor II.
In addition, the rotor frequency of both Motor I and II decreases from 24.9 Hz to 24.46
Hz when the load increases to 80% current load level.

Motor I and Motor II operating at high load level were also tested. The two motors
running under different load levels are compared in Figure 3.5 to 3.10. From these fig-
ures, we can conclude that both the rotor frequency and the amplitude of Motor I were
smaller than Motor II at the same load level. The difference between the two motors
is tabulated in Table 3.2. In addition, for Motors I and II, the amplitude of rotation
frequency decreases when the load level increases. In all cases, the load level shown
was measured as a percentage of the nominal current.

According to the rotor frequency values which were detected directly by the ABB de-
tection system, the rotor speed Sr and slip s can be computed by

Sr = f r × 60, (3.2)

s = (1500 − Sr )/1500, (3.3)

Page 25
3.2 Baseline Study

−3
x 10 (i) Interpolated zero−crossing
1.5
x 10 intervals (no load)
-3

1.5
Mean zero−cross interval (s)

1 0.5
Zoom In 0
0.5 -0.5

-1
0
-1.5
10 15 20 25 30 35 40 45 50
−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase (no load)

0
X:24.9
Magnitude response (dB)

Y:−35.57
−20

−40

−60

−80
Sampling frequency: 1000 Hz
Window length: 1024
−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 3.1. Healthy induction Motor I without load.

From Table 3.3, the rotor frequency of Motor I with 83.57% load is 24.47 Hz and the
Sr = 24.47 × 60 = 1468.2 rpm. The motor has 4 poles and supplied by 50 Hz. There-
fore, the s = 0.0212 and 2 f s = 2.12. Tables 3.3 and 3.4 show online testing results
for Motor I and Motor II, respectively. In addition, Figures 3.11 and 3.12 illustrate the
evolution of slip with the increasing current value.

The amplitude variation of rotor frequency is plotted in Figure 3.13. As the load in-
creases to 80%, the amplitude of the two motors’ rotor frequencies increases. How-
ever, the amplitudes decline sharply after a load level higher than 80%. The difference
is more than 15 dB between zero load and 185% load.

Page 26
Chapter 3 Baseline Analysis of Induction Motor

x 10
−3 (i) Interpolated zero−crossing intervals (no load)
1.5

Mean zero−cross interval (s)

1

0.5

−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase (no load)

0
Magnitude response (dB)

X:24.9
Y:−33.09
−20

−40

−60

−80
Sampling frequency: 1000 Hz
Window length: 1024
−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 3.2. Healthy induction Motor II without load.

−3
x 10 (i) Interpolated zero−crossing intervals (80% load)
1.5
Mean zero−cross interval (s)

0.5

−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase (80% load)

0
Magnitude response (dB)

X:24.46
−20 Y:−31.15

−40

−60

−80 Sampling frequency: 1000 Hz

Window length: 1024
−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 3.3. Healthy induction Motor I with 80% load.

Page 27
3.2 Baseline Study

−3
x 10 (i) Interpolated zero−crossing intervals (80% load)
Mean zero−cross interval (s) 1.5

0.5

−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase(80% load)

0
Magnitude response (dB)

X:24.46
Y:−30.54
−20

−40

−60

−80
Sampling frequency: 1000 Hz
Window length: 1024
−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 3.4. Healthy induction Motor II with 80% load.

−3
x 10 (i) Interpolated zero−crossing intervals (90% load)
1.5
Mean zero−cross interval (s)

0.5

−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase (90% load)

0
Magnitude response ( dB)

X:24.17
−20 Y:−39.42

−40

−60

−80
Sampling frequency: 1000 Hz
Window length: 1024
−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 3.5. Healthy induction Motor I with 90% load.

Page 28
Chapter 3 Baseline Analysis of Induction Motor

−3
x 10 (i) Interpolated zero−crossing intervals (90% load)
1.5

Mean zero−cross interval (s)

1

0.5

−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase (90% load)

−20
X:24.32
Magnitude response (dB)

Y:−36.69

−40

−60

−80
Sampling frequency: 1000 Hz
Window length: 1024
−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 3.6. Healthy induction Motor II with 90% load.

−3
x 10 (i) Interpolated zero−crossing intervals (110% load)
1.5
Mean zero−cross interval (s)

0.5

−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase (110% load)

0
Magnitude response (dB)

Sampling frequency: 1000 Hz

X:23.88 Window length: 1024
−20 Y:−40.58

−40

−60

−80

−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 3.7. Healthy induction Motor I with 110% load.

Page 29
3.2 Baseline Study

−3
x 10 (i) Interpolated zero−crossing intervals (110% load)
Mean zero−cross interval (s) 1.5

0.5

−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase (110% load)

−20 X:24.02
Magnitude response (dB)

Sampling frequency: 1000 Hz

Y:−38.5 Window length: 1024

−40

−60

−80

−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 3.8. Healthy induction Motor II with 110% load.

−3
x 10 (i) Interpolated zero−crossing intervals (160% load)
1.5
Mean zero−cross interval (s)

0.5

−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase (160% load)

0
Magnitude response (dB)

X:22.85
−20 Y:−49.32

−40

−60

−80
Sampling frequency: 1000 Hz
−100 Window length: 1024

−120
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 3.9. Healthy induction Motor I with 160% load.

Page 30
Chapter 3 Baseline Analysis of Induction Motor

−4
x 10 (i) Interpolated zero−crossing intervals (160% load)
1.5

Mean zero−cross interval (s)

1

0.5

−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase (160% load)

−20 X:23.29
Magnitude response (dB)

Y:−46.32

−40

−60

−80

−100 Sampling frequency: 1000 Hz

Window length: 1024
−120
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 3.10. Healthy induction Motor II with 160% load.

Table 3.2. Difference between Motor I and Motor II.

Difference of component frequency Difference of amplitude between

Load level
between Motor I and II (Hz) Motor I and II (dB)
80% 0.15 2.73
90% 0.14 2.08
110% 0.44 3

Figure 3.11. Computed slip vs. I/Inom for Motor I.

Page 31
3.2 Baseline Study

Table 3.3. Online testing results for Motor I (recording directly by LabVIEW).

Current Rotor
Amplitude Computed speed Slip
(A) frequency Load level
(dB) (rpm) (%)
(measured by (Hz) (%)
(20-30) (from rotor peak) (from computed speed)
device)
(20-30)

24.95 -27.84 1496.86 0.2092

4.090 24.47 -25.06 1468.20 2.1012 83.57
4.685 24.22 -28.60 1453.20 3.1216 95.61
5.224 23.94 -35.00 1436.40 4.2260 106.61
6.459 23.50 -40.00 1410.00 6.0024 131.82
8.243 22.96 -40.40 1377.60 8.1800 168.22
9.174 22.55 -41.30 1353.00 9.8000 187.22

Table 3.4. Online testing results for Motor II (recording directly by LabVIEW).

Current Rotor
Amplitude Computed speed Slip
(A) frequency Load level
(dB) (rpm) (%)
(measured by (Hz) (%)
(20-30) (from rotor peak) (from computed speed)
device)
(20-30)

24.967 -27.637 1498.0 0.13

4.11 24.496 -24.308 1469.7 2.02 83.878
4.87 24.248 -27.964 1454.9 3.01 99.388
5.82 23.969 -31.106 1438.1 4.12 118.78
6.95 23.653 -38.666 1419.2 5.39 141.84
8.16 23.239 -37.436 1394.4 7.04 166.53
9.65 22.680 -40.991 1360.8 9.28 196.94

Figure 3.12. Computed slip vs. I/Inom for Motor II.

Page 32
Chapter 3 Baseline Analysis of Induction Motor

−22

−24 motor 1
motor 2
−26

Amplitude of rotor frequency (dB)

−28

−30

−32

−34

−36

−38

−40

−42

−44
0 20 40 60 80 100 120 140 160 180 200
I/Inom(%)

Figure 3.13. Amplitude of rotor frequency vs. I/Inom .

3.3 Summary
In this chapter, two healthy motors were utilized for the baseline analysis. Although
a healthy induction motor is highly symmetrical, any potential fault will alter its sym-
metrical property. These potential faults, which may result from manufacturing or
installation, will produce characteristic fault frequencies. Such fault frequencies can
be used for diagnosing broken rotor bar and stator short circuit faults. Thus, base-
line analysis is important to eliminate other potential faults and understand frequency
components of nominally healthy motors. Furthermore, the experiments show the be-
havior of the characteristic fault frequency components of two healthy motors as a
function of load levels.

In summary, through the comparison, even though there is not any artificial fault set-
ting for the two motors, the peak values of Motor I rotor frequency are less than that
of Motor II at same load level. Therefore, might a possible misalignment or another
potential faults in Motor II.

Page 33
Page 34
Chapter 4

Alignment Measurement
Analysis

I
N this chapter, the possibility of motor alignment is studied and
if effect on the motor characteristics are investigated through ex-
perimentation. The motor alignment is adjusted by using a laser
alignment system. The experiments show that motor misalignment influ-
ences the fault frequency of a stator short circuit fault at low load levels.

Page 35
4.1 Introduction

4.1 Introduction

A correct alignment of an industrial motor can increase the operating lifetime of rotat-
ing machinery and ensure smooth and efficient power transfer from the motor to the
driven equipment. A misalignment will cause shaft fracture, bearing failures and vi-
brations due to the increased stress. These faults will lead to plant breakdowns, energy
loss and other potential faults [34].

Misalignment may occur at the motor or shaft [34]. A motor misalignment results in
an imbalance among the motor feet. For example, Figure 4.1 shows a motor feet with a
vertical misalignment. The motor misalignment is considered as one of the main rea-
sons that the devices waste energy because it can result in an increased friction on the
shaft. This kind of friction will result in unnecessary increase of energy consumption.

NOTE:
This figure is included on page 36 of the print copy of
the thesis held in the University of Adelaide Library.

Figure 4.1. Motor feet misalignment [42].

On the other hand, shaft misalignment is an imbalance between the coupled rotating
plant machinery, such as motors, generators and pumps [43]. The aim of the shaft
alignment is to ensure the coupled rotating machine works at a perfect rotational cen-
ter line. The misalignment between two machines will lead to two shafts rotating the
deviation of the rotational centre line, which increases the stresses on the shaft and
coupling. Then, the friction on bearings will increase the chance of a plant breakdown.

Motor misalignment can be broadly categorized into three types [37]. Firstly, angular
misalignment occurs when there is different angles of the horizontal plane between

Page 36
Chapter 4 Alignment Measurement Analysis

two parallel shafts [44]. Some driven equipment and motor faults are caused by angu-
lar misalignment. The second type is parallel misalignment. The parallel misalignment
occurs when the two parallel shafts’ centerlines are not in the same line. The offset can
be catalogued as either a horizontal or vertical offset. The parallel misalignment can
cause an imbalance of the current signal and bring several potential problems for on-
line condition monitoring. Finally, combined misalignment is the most common when
the motor’s shaft suffers from angular misalignment and parallel misalignment at the
same time.

The Laser Alignment system, provided by Fixturlaser Company [37], was utilized to
detect motor and shaft misalignment in our lab. The specialist technology ensures
a high quality measurement. The Laser Alignment system consists of a display unit
(Figure 4.2), measurement unit and shaft bracket, including rods and chain (Figure
4.3).

NOTE:
This figure is included on page 37 of the print copy of
the thesis held in the University of Adelaide Library.

Figure 4.2. Fixturlaser [37].

The measurement process of the Laser Alignment system is described as follows. Two
of the measurement units are fixed on the two shafts. When the detectors receive a
laser beam from the transmitters, the two lights on the screen turn to green. Then, the
distances between each component need to be measured and entered manually into
the system, as shown in Figure 4.4. The measurement will be implemented at three

Page 37
4.2 Alignment Measurement

Figure 4.3. Measurement unit.

different points. The distance between these points should be at least 60 degrees to
improve the accuracy. Finally, the measurement data are analyzed and results are dis-
played on the screen [37].

Figure 4.4. Measurement process.

4.2 Alignment Measurement

Three motors have been installed in our laboratory. The motor parameters are listed
in Table 3.1. The alignment measurements of Motor II and Motor III were illustrated
as examples. The measurement results of Motor II are shown in Figure 4.5(a). The

Page 38
Chapter 4 Alignment Measurement Analysis

(a) Measurement result (b) Measurement result after adjust-

ment

Figure 4.5. Measurement results of the Motor II.

(a) Measurement result (b) Measurement result after adjust-

ment

Figure 4.6. Measurement results of the Motor III.

misalignments are shown in both number and color codes. For example, the red color
indicates out of tolerance, the orange means near the tolerance and the green means
acceptable. The results denote that the motor suffers from combined misalignment,
i.e. both angular and parallel misalignment. Figure 4.5(b) shows the alignment of the
adjusted motor. The initial and adjusted alignments of Motor III are shown in Figure
4.6(a) and 4.6(b), respectively.

Page 39
4.3 Comparison of the On-line Condition Monitoring Testing Results

4.3 Comparison of the On-line Condition Monitoring Test-

ing Results

4.3.1 Healthy Motor Testing

Through the analysis of motor current signal in the frequency domain, the condition of
under-tested motor can be diagnosed to find potential faults, which might result from
manufacturing defects and installation problems. Figure 4.7 shows the ZCT signal of
Motor III in both time and frequency domains. In frequency domain, there is a fre-
quency peak (-29.7 dB) at 24.46 Hz, which indicated the existence of other faults. After
a through inspection, a gap (Figure 4.8) between two the shafts was found. After the
gap was reduced, the frequency component decreased by 3.7 dB, as shown in Figure
4.9.

−3 (i) Interpolated zero−crossing intervals

x 10
1
Mean zero−cross interval (s)

0.5

−0.5

−1
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase

Magnitude response (dB)

−20 X:24.46
Y:−29.7
−40

−60

−80
Sampling frequency: 1000 Hz
Window length: 1024
−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 4.7. ZCT signal of Motor III.

The motor was tested with different torque levels resulting from changing the load
level. The torque level corresponding to the current load level is shown in the table
4.1. Figure 4.10 reveals that when there is a misalignment, the rotation frequency am-
plitude of the motor with a coupling gap is higher than without a gap at lower load
level. However, the amplitude will be influenced by the noise produced by the high
load level. Compared with other graphs which are shown in Figure 4.10, the graph of

Page 40
Chapter 4 Alignment Measurement Analysis

(a) Coupling with gap (b) Coupling state after adjustment

Figure 4.8. Coupling adjustment.

x 10
−3 (i) Interpolated zero−crossing intervals
1
Mean zero−cross interval (s)

0.5

−0.5

−1
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase

0
Magnitude response (dB)

−20 X:24.46
Y:−33.4
−40

−60

−80 Sampling frequency: 1000 Hz

Window length: 1024
−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 4.9. ZCT signal of Motor III (after coupling adjustment).

Table 4.1. Torque levels corresponding current load (Motor III).

Torque Level Load Level

10% > 70%
20% > 80%
30% > 98%
40% > 117%
50% > 140%
60% > 165%
70% > 196%

Page 41
4.3 Comparison of the On-line Condition Monitoring Testing Results

the motor with only a misalignment trend is different than the others. The amplitude
reduces when the torque level is 90%. When the motor runs without load, the ampli-
tude of the motor without misalignment is the lowest.

−20

−22

−24
Amplitude of the rotor frequency (dB)

−26

−28

−30

−32

−34

−36 Motor 3
Motor 3 with coupling gap+ misalignment
Motor with misalignment
−38

−40
0 20 40 60 80 100 120 140 160 180 200
Load level (%)

Figure 4.10. The test results of Motor III.

4.3.2 The Misalignment Influence on Stator Short Circuit Detection

The effect of misalignment for stator short circuit testing is also studied in our lab.
There are two types of the testing according to the two kinds of stator short circuits,
which are stator short circuit directly and stator short circuit with resistance. The test-
ing was also carried out under different torque levels. The torque levels corresponding
to the current load level is shown in the table 4.2.

Figure 4.11 shows the testing results of a motor with three different conditions, a mo-
tor with misalignment, a motor with misalignment and turn-to-turn stator short circuit
without resistance and a motor without misalignment and turn-to-turn stator short
circuit without resistance. The figure clearly shows that all peaks of red line, which
represents the motor running with misalignment and turn-to-turn stator short circuit,
is higher than the blue line’s, which represents healthy motor running with misalign-
ment. The reason is that the value of the fault frequency amplitude increases instantly

Page 42
Chapter 4 Alignment Measurement Analysis

Table 4.2. Torque levels corresponding current load (Motor II).

Torque Level Load Level

10% > 72%
20% > 83%
30% > 98%
40% > 117%
50% > 140%
60% > 165%
70% > 193%

when the stator short circuit fault happens. The influence of the amplitude by mis-
alignment is also shown in Figure 4.11. When the motor runs at the low torque level,
the peaks of the black line, which denotes the motor short circuit without resistance,
are lower than the red line’s peak values. Therefore, the misalignment can increase
the amplitude of the fault frequency. However, a high load level will produce large
amounts of noise and the peaks of the black line is slightly higher than that of the
red’s. The test results of the turn-to-turn stator short circuit fault with resistance are
shown in Figure 4.12, which shows that misalignment can lead to an increase in am-
plitude of fault frequency as well.

−15

−20
Amplitude of the rotor frequency (dB)

−25

−30

−35

Misalignment
−40
Misalignment stator short directly
Stator short directly

−45
0 20 40 60 80 100 120 140 160 180 200
Load level (%)

Figure 4.11. Misalignment effect for turn-to-turn stator short circuit directly.

Page 43
4.3 Comparison of the On-line Condition Monitoring Testing Results

−20

−25

Amplitude of the rotor frequency (dB)

−30

−35

Misalignment
−40 Misalignment stator short with resistance
Stator short with resistance

−45
0 20 40 60 80 100 120 140 160 180 200
Load level (%)

Figure 4.12. Misalignment effect for turn-to-turn stator short circuit with resistance.

−26

−28
Amplitude of the rotor frequency (dB)

−30

−32

−34

−36

−38

−40
Misalignment
Misalignment phase1,phase2 short directly
−42 Phase1,phase2 short directly

−44
0 20 40 60 80 100 120 140 160 180 200
Load level (%)

Figure 4.13. Misalignment effect for phase-to-phase stator short circuit directly.

The misalignment influence for a phase-to-phase stator short circuit with and without
resistance is demonstrated in Figure 4.13 and 4.14. When the motor operates under
misalignment and the stator short circuit without resistance, the amplitude value of
the fault frequency fluctuates. Compared with Figures 4.13 and 4.14, the resistance
will not affect the amplitude of rotor frequency of phase-to-phase stator short circuit
at higher load levels (more than 140%). On the contrary, the amplitude of rotor fre-
quency of phase-to-phase stator short circuit without resistance is lower than that of

Page 44
Chapter 4 Alignment Measurement Analysis

stator short circuit with resistance.

−15

Amplitude of the rotor frequency (dB) −20

−25

−30

−35

Misalignment
−40 Misalignment phase1,phase2 short with resistance
Phase1,phase2 short with resistance

−45
0 20 40 60 80 100 120 140 160 180 200
Load level (%)

Figure 4.14. Misalignment influence for phase-to-phase stator short circuit with resistance.

−15

−20
Amplitude of the rotor frequency (dB)

−25

−30

−35

−40 Motor 2 with misalignment

Motor 2 turn−to−turn stator short directly with misalignment
Motor 2 turn−to−turn stator short directly alignment
Motor 3 with misalignment and coupling gap
−45 Motor 3 with misalignment
Motor 3 alignment

−50
0 20 40 60 80 100 120 140 160 180 200
Load level (%)

Figure 4.15. Comparison of misalignment measurement testing results.

In order to compare Motor II and Motor III, the test results are plotted in the Figure
4.15. Although the type of motor is the same, there is a distinction between the testing
results. The amplitude of Motor III with misalignment is higher than Motor II. There
may be some potential problems and they will be diagnosed in our future research. In
addition, Figure 4.15 demonstrates that the stator short circuit is the primary reason for

Page 45
4.4 Summary

increasing the fault frequency amplitude. When the motor operates at low load levels
with a stator short circuit, the amplitude of the stator short circuit is higher than that of
the other fault amplitudes even if there is a gap in the coupling. This could be because
of the whole assembly of shaft bending, and/or an existing manufacturing defect.

4.4 Summary
In this chapter, motor misalignment is varied and adjusted by using the Laser Align-
ment system. Based on the measurements, we also found that a coupling gap can lead
to an unexpected frequency peak. At a low load level, misalignment will increase the
amplitude of the fault frequency in a stator short circuit fault. While, misalignment has
little effect on the amplitude when running under high load conditions.

Page 46
Chapter 5

Stator Short Circuit Fault

I
N this chapter, stator short circuit faults are studied by using
both off-line and on-line tests. Through a control box, artificial
turn-to-turn faults and phase-to-phase faults are simulated. In
addition, the experiments consider different load levels.

Page 47
5.1 Introduction

5.1 Introduction

A stator short circuit in winding is considered as one of the main faults in an induc-
tion motor. There are four types of stator short circuit faults, turn-to-turn short circuits
within a coil, short circuits between coils of the same phase, phase-to-phase shorts and
phase to earth short circuits [45]. Figure 5.1 shows the pattern of short circuit faults in
a stator winding. Although a stator short circuit fault does not affect the normal opera-
tion in the early stage, it will reduce the efficiency of a motor and even damage it when
the fault becomes severe. Therefore, early diagnosis of the stator short circuit fault in
operation is important before it becomes a catastrophic failure to induction motors. In
addition, it will help to eliminate costly damages in terms of lost production time and
avoid expensive cost for maintenance and repair [46].

NOTE:
This figure is included on page 48 of the print copy of
the thesis held in the University of Adelaide Library.

Figure 5.1. Wye connected stator showing possible short circuit faults [45].

A stator short circuit in winding is influenced by the quality of the insulation between
the turns and phases of the coils inside a motor [3]. Mechanical stress is generated by
stator coils moving during the operation of a motor. When the motor starts, the current
in the coils will reach its highest value and it will result in a high magnetic force. This
magnetic force will cause the coils to vibrate at two times than line frequency. This

Page 48
Chapter 5 Stator Short Circuit Fault

vibration could damage the stator and other motor components. Another fault conse-
quence is electrical stress [3]. Electrical failures are common failures in the induction
motor windings. Most of them are caused by over-voltage, current overload and so
on. It can cause some faults in the circuit during testing or operation. In addition,
thermal stress is also one of the major reasons for a stator short circuit. It results from
the insulation deterioration due to the motor working temperature being higher than
the designed temperature. The insulation life will be reduced by half when the tem-
perature increases every 10◦ C above designed temperature [3]. Thermal stress can also
reduce the life of a motor when operating at higher than rated voltage and overloaded.
Furthermore, the environmental stress cannot be neglected. Contamination, such as
dust, oil and moisture, will enter into the motor and cause environmental stress. The
contaminations influence the stator insulation lifetime and result in winding faults in
the motor.

5.2 Detection of Stator Short Circuit

The diagnosis of shorted turns in stator windings of an induction motor using the mo-
tor current signature analysis (MCSA) is based on detecting the frequency components
given by equation (5.1) [31]
h n i
f short = f s k ± (1 − s ) , (5.1)
P

where,
f short is the frequency component that is a function of the short circuit,
f s is the supply frequency,
P is the number of pole pairs,
n and k are two parameters, n = 1, 2, 3... and k = 1, 2, 3...

The slip speed s is defined as

ωsm − ωrm
s= , (5.2)
ωsm
where ωsm = 2ωe /P is the synchronous speed in mechanical radians, ωe is angular
speed of the stator Magnetic Motive Force (MMF) in electrical radians per second and
ωrm is the rotor rotating speed.

Page 49
5.3 Experiment Setup for Stator Short Circuit Faults Diagnosis

When k = 1 in equation (5.1), the component frequency, induced in the spectrum of

the stator current, superimposed on the fundamental current, will be illustrated by the
following formula
hn i
f comp = f s (1 − s ) . (5.3)
P
For two pole motors ( p = 2) in first instance, the equation (5.3) will be a whole series
of components:

 f s (1 − s ) n=1


f comp = 2 f s (1 − s ) n = 2

 3 f (1 − s ) n = 3

s

For four pole motors ( p = 4), the equation (5.3) will be revealed the following series
of components:

 (1/2) f s (1 − s) n = 1


f comp = f s (1 − s ) n=2

 (3/2) f (1 − s) n = 3

s

These components are visible in the frequency spectrum of the stator current (as side-
bands to the fundamental peak). This will affect the normal ZCT spacing, causing the
indicative frequency components to appear at locations given by equation (5.3) in the
spectrum of the ZCT signal as well [31]. Since this stator fault component in the ZCT
spectrum is independent of the motor’s inertia, it can be used as a good indicator of
the stator short circuit fault. However, putting n = 1 in equation (5.3), we can get

1−s
 
f comp = f s . (5.4)
p
Equation (5.4) expresses the relationship between the component frequency and rotor
frequency. When n = 1, it is difficult to judge the frequency peak due to stator fault or
rotor frequency in the ZCT spectrum. Therefore, we utilize values of n > 1 .

5.3 Experiment Setup for Stator Short Circuit Faults Di-

agnosis
The off-line and on-line test system sketches are shown in Figures 5.2(a) and 5.2(b).
Two of the same type of motors, Motor I and Motor II, are employed in off-line and

Page 50
Chapter 5 Stator Short Circuit Fault

on-line test, respectively. The motor parameters are listed in Table 3.1. For the off-line
test, an oscilloscope is used to show and capture the current signal. Then, the signal
is processed by using MATLAB. On the other hand, during the on-line test, the ZCT
signal is acquired by the device that does the ZCT signal conversion and LabVIEW is
employed to plot and analyze the signal.

Another motor wass utilized as a load and the load level was varied through the AC
drives to change the load level. The setup of the experiment is illustrated in Figure 5.3.
Some turns from one slot were selected and scraped off their insulation layer and sol-
dered them with different wires labeled with numbers. There are several combinations
of the wires for the stator short circuit. In addition, in order to test the phase-to-phase
stator short circuit, Some turns from another slot were selected and also soldered with
a different wires labels. Finally, the wires are connected to a switch box to control the
stator short circuit.

5.4 Off-line Test of Stator Short Circuit Faults

A turn-to-turn fault is the initial fault in the induction motor. It will cause other faults
such as phase-to-phase, phase to earth fault or even more serious faults like motor
damage. Therefore, turn-to-turn faults detection is important to prevent potential
catastrophic faults.

In this section, turn-to-turn faults are studies in both off-line and on-line tests. The
off-line testing was utilized to analyze the motor’s current signal when the motor is
operated at different load levels. The magnitude variation of the fault frequency am-
plitude was estimated. Compared with off-line testing, on-line testing is using the
current ZCT signal and it can monitor and diagnose motor faults immediately when
the motor is running.

5.4.1 Analysis of Turn-to-turn Fault of Motor I

In this section, the frequency components in the current signal, when n = 1, P = 2

and k = 1 in equation (5.1), are investigated and analyzed. In addition, two types of

Page 51
5.4 Off-line Test of Stator Short Circuit Faults

Data Acquisition and Using

MATLAB Analysis

Oscilloscope Temperature
Monitoring

Motor Motor

Connector
Box
ABB AC Drives Voltage
Meters

(a) Off-line detection structure diagram

PC with
Labview
USB
ZCT Signal Acquisition

DC Power ZCT Signal

Supply Converter Temperature
Monitoring

Motor Motor

Connector
Box
ABB AC Drives Voltage
Meters

(b) On-line detection structure diagram

Figure 5.2. Off-line and on-line detection structure diagram.

Page 52
Chapter 5 Stator Short Circuit Fault

(a) (b)

Figure 5.3. The laboratory setup for induction motor stator short circuit experiments.

turn-to-turn faults are tested: two turns short circuit and five turns short circuit.

When a healthy motor is under 80% load, the measured rotor speed is 1471 rpm. There-
fore, the slip, s = (1500 − 1471)/1500 ≈ 0.019, and the fault component frequency can
be computed f comp = 1/2 × 50 × (1 − s) ≈ 24.5 Hz. When k = 1, according to equa-
tion (5.1), the fault frequency is f short = |24.5 ± 50| Hz. Thus, the left and right hand
short frequency are f short L = |24.5 − 50| = 25.5 Hz and f short R = |24.5 + 50| = 74.5
Hz, respectively. Figure 5.4 illustrates the testing current data and ZCT signal. In
Figure 5.4, the fault frequencies can be found and the corresponding amplitudes are
-8.613 dB (25.5 Hz) and -6.947 dB (74.5 Hz). In addition, the healthy motor with high
load levels was also tested. The load level with 140%, as an example, can verify the
reliability of the experiment setup. When the healthy motor operates under 140%
load level, the measured rotor speed is 1421 rpm. Thus, the slip can be calculated
as s = (1500 − 1421)/1500 ≈ 0.053 and the component frequency can be computed
f comp = 1/2 × 50 × (1 − s) = 23.675 Hz. The the left and right hand short frequency are
f short = |23.675 ± 50| Hz, i.e. 26.325 Hz and 73.675 Hz. However, there are slight dif-
ferences between the computed fault frequencies and the fault frequencies, which are
showed in Figure 5.5. The reason could be due to error in the rotor speed measurement
or a computation error in MATLAB. In spite of the small error, the experiment setup
is considered as reliable. In addition, Figure 5.5 shows the corresponding amplitude
of the fault frequencies as -11.63 dB (26.35 Hz), -10.88 dB (73.65 Hz). Compared with
Figure 5.5 and 5.6, the fault frequencies are close to the rotor fundamental frequency

Page 53
5.4 Off-line Test of Stator Short Circuit Faults

and the amplitudes of the fault frequencies decrease when the load level increases.

Current in phase 1,2,3 Magnitude plot of current in phase 1,2,3

60

Magnitude response (dB)

0.02 40
X:25.5 X:74.5
Y:−8.613 Y:−6.947
0.01 20
Current (A)

0 0
−20
−0.01
−40
−0.02
−60 Sampling frequency: 1000 Hz
−0.03 Window length: 1024
−80
2950 3000 3050 3100 3150 0 20 40 60 80 100
Sample no. Frequency (Hz)
Interpolated zero−crossing intervals Magnitude plot of zero−cross interval in phase 1,2,3
−4
x 10
4 0
Mean zero−cross interval (s)

Magnitude response (dB)

−20
2 X:24.61
Y:−57.18
−40
0
−60

−2 −80
Sampling frequency: 1000 Hz
−100 Window length: 1024
−4
0 200 400 600 800 1000 0 50 100 150
Data blocks Frequency (Hz)

Figure 5.4. Healthy motor with 80% load level.

When a turn-to-turn fault happens, the variations in the current signal are revealed in
Figures 5.6 and 5.7. They are the test results for two turns short circuit in the same slot.
Figure 5.6 shows that the amplitudes of frequencies are higher than the amplitudes
which are presented in Figure 5.4. Although there is a slight difference with the fre-
quencies presented in Figure 5.4, the error is within tolerance because the difference is
produced by the noise of the load level. Comparing Figure 5.5 with 5.7, the magnitude
at fault frequency (23.44 Hz) increases as high as 4 dB when a turn-to-turn short circuit
occurs. Therefore, we can conclude that when the motor operates with a turn-to-turn
fault, the amplitude of specific fault frequencies will increase as we expected.

By using the switch box, different turns shorts in the same slot are realized. The test
results for two and five turn-to-turn faults are listed in Tables 5.1 and 5.2, respectively.
In each test, the harmonic frequencies and the corresponding amplitudes are recorded
at different loading conditions. In addition, the slip values are computed and the vari-
ations are plotted in Figure 5.8.

Page 54
Chapter 5 Stator Short Circuit Fault

Current in phase 1,2,3 Magnitude plot of current in phase 1,2,3

60
0.04

Magnitude response (dB)

40
X:26.35 X:73.65
0.02 20 Y:−11.63 Y:−10.88
Current (A)

0
0
−20

−0.02 −40

−60 Sampling frequency: 1000 Hz

Window length: 1024
−0.04
−80
1420 1440 1460 1480 1500 1520 0 20 40 60 80 100
Sample no. Frequency (Hz)
Interpolated zero−crossing intervals Magnitude plot of zero−cross interval in phase 1,2,3
−4
x 10
3 0
Mean zero−cross interval (s)

X:50.1

Magnitude response (dB)

2 Y:−31.54
−20
X:23.73
1 −40 Y:−59.93
0
−60
−1
−80
−2 Sampling frequency: 1000 Hz
−100 Window length: 1024
−3
0 200 400 600 800 1000 0 50 100 150
Data blocks Frequency (Hz)

Figure 5.5. Healthy motor with 140% load level.

Current in phase 1,2,3 Magnitude plot of current in phase 1,2,3

60
Magnitude response (dB)

0.01 40 X:25.6 X:74.4

Y:−6.19 Y:−8.765
20
Current (A)

0.005
0
0
−20
−0.005 −40

−0.01 −60 Sampling frequency: 1000 Hz

Window length: 1024
−80
1860 1870 1880 1890 0 20 40 60 80 100
Sample no. Frequency (Hz)
Interpolated zero−crossing intervals Magnitude plot of zero−cross interval in phase 1,2,3
−3
x 10
1 0
Mean zero−cross interval (s)

X:24.32
Magnitude response (dB)

Y:−38.95
−20
0.5
−40
0
−60

−0.5 −80
Sampling frequency: 1000 Hz
−100 Window length: 1024
−1
0 200 400 600 800 1000 0 50 100 150
Data blocks Frequency (Hz)

Figure 5.6. Turn-to-turn short circuit with 80% loading level.

Page 55
5.4 Off-line Test of Stator Short Circuit Faults

Current in phase 1,2,3 Magnitude plot of current in phase 1,2,3

60

Magnitude response (dB)

0.015 40 X:26.6 X:73.4
0.01 Y:−7.372 Y:−13.26
20
Current (A)

0.005
0
0
−20
−0.005
−40
−0.01
−0.015 −60 Sampling frequency: 1000 Hz
Window length: 1024
−80
1200 1220 1240 1260 1280 1300 1320 0 20 40 60 80 100
Sample no. Frequency (Hz)
Interpolated zero−crossing intervals Magnitude plot of zero−cross interval in phase 1,2,3
−4
x 10
4 0
Mean zero−cross interval (s)

Magnitude response (dB)

X:23.44
2 −20 Y:−44.09

0 −40

−60
−2
−80
−4
Sampling frequency: 1000 Hz
−100 Window length: 1024
−6
0 200 400 600 800 1000 0 50 100 150
Data blocks Frequency (Hz)

Figure 5.7. Turn-to-turn short circuit with 140% loading level.

Table 5.1. Turn-to-turn fault (two turns).

Amplitude of Amplitude of
Load Fault Component Fault component
frequency frequency Slip
Level Frequency (n=1 frequency (n=1
component (n=1 component (n=1 (%)
(%) k=+1) k=-1)
k=+1) k=-1)
0 25 -21.5 75 -18.4
80 25.6 -6.19 74.4 -8.765 2.4
90 25.9 -6.315 74.1 -9.391 3.6
110 26.2 -5.135 73.8 -10.03 4.8
140 26.6 -7.372 73.4 -13.26 6.4

Page 56
Chapter 5 Stator Short Circuit Fault

Table 5.2. Turn-to-turn fault (five turns).

Amplitude of Amplitude of
Load Fault Component Fault component
frequency frequency Slip
Level Frequency (n=1 frequency (n=1
component (n=1 component (n=1 (%)
(%) k=+1) k=-1)
k=+1) k=-1)
0 25 -17.81 75 -13.3
80 25.5 -18.34 74.5 18.6 2
90 25.8 -18.34 74.2 -18.6 3.2
110 26.2 -20.07 73.9 -14.3 4.4
140 26.4 -17.8 73.6 -17.7 5.6

5
Slip (%)

2
Turn−to−turn fault (two turns)
Turn−to−turn fault (five turns)

1
80 90 100 110 120 130 140
Load level (%)

Figure 5.8. Slip with two and five turns faults.

Figure 5.8 illustrates the variation of the motor slip value for different turn-to-turn
short circuits. The slip value rises while the load level increases. At the same load
level, the slip value for five turn fault is smaller than the value of two turn fault. In ad-
dition, the difference in slip values between the different levels also increases as load
level increase.

Figure 5.9 shows the fault frequencies amplitudes of a healthy motor, a slight turn-to-
turn fault (2 turns fault) and a motor with a serious turn-to-turn fault (5 turns fault).
When the motors operate without load, the fault frequency amplitude of the healthy

Page 57
5.4 Off-line Test of Stator Short Circuit Faults

motor is the lowest, as expected, and the amplitude of the serious turn-to-turn fault
frequency is the highest. The amplitudes of the frequencies of the two turns fault are
always higher than the values of the healthy motor under different load levels. How-
ever, the load levels affect the variation of rotor frequency amplitude when there is a
serious fault. The rotor frequency amplitude of a serious fault has small changes under
the different load levels.

−5

−10
The amplitude of rotor frequency (dB)

−15

−20

−25
Turn−to−turn fault (two turns fault)
Turn−to−turn fault (five turns fault)
Healthy motor

−30
0 20 40 60 80 100 120 140
Load level (%)

Figure 5.9. The amplitude of rotor frequency under different load levels.

5.4.2 Analysis of Phase-to-phase Fault in Motor I

The phase-to-phase fault is the short circuit between two turns in different slots. It is
expected to have a similar influence as the turn-to-turn faults on the amplitude of the
rotor frequency. Under this assumption, the turn-to-turn fault detection method can
also detect phase-to-phase faults. Like turn-to-turn fault detection, when n = 1, p = 2
and k = ±1 in equation (5.1).

The phase-to-phase short circuit is more serious than turn-to-turn fault because a higher
voltage difference between two phases make this fault accelerate very quickly. If the
fault is left unattended, it can lead to rapid deterioration of the winding, and poten-
tially destroy the motor. The magnitude of the fault frequency amplitude variation is

Page 58
Chapter 5 Stator Short Circuit Fault

higher than that for a turn-to-turn fault. Figures 5.10 and 5.11 show the detection re-
sults of a phase-to-phase short circuit when the machine operates under 80% load and
140% load. Comparing Figure 5.6 with 5.10, we can conclude that the amplitude of
the fault frequencies of the phase-to-phase short circuit is much higher than that of the
turn-to-turn short circuit. While the load level grows up to 140%, the amplitude of the
fault frequency increases to 0.2273 dB (when k = −1).

Current in phase 1,2,3 Magnitude plot of current in phase 1,2,3

60

Magnitude response (dB)

X:25.5 X:74.5
0.02 40
Y:−1.739 Y:−0.8316
20
Current (A)

0.01
0
0
−20
−0.01
−40
−0.02 −60 Sampling frequency: 1000 Hz
Window length: 1024
−80
1050 1100 1150 0 20 40 60 80 100
Sample no. Frequency (Hz)
Interpolated zero−crossing intervals Magnitude plot of zero−cross interval in phase 1,2,3
−4
x 10
4 0
Mean zero−cross interval (s)

Magnitude response (dB)

X:24.32
−20 Y:−42.09
2
−40
0
−60

−2 −80
Sampling frequency: 1000 Hz
−100 Window length: 1024
−4
0 200 400 600 800 1000 0 50 100 150
Data blocks Frequency (Hz)

Figure 5.10. Phase-to-phase fault with 80% loading level.

The FFT result for the phase-to-phase fault is summarized in Table 5.3. A more seri-
ous condition, both a phase-to-phase fault and a turn-to-turn short circuit fault, is also
tested and listed in Table 5.4. The tables state the frequencies and corresponding am-
plitudes when the motor was tested with different loading levels. The motor slip is
computed and listed in the tables too.

The slip variations for a phase-to-phase short circuit and phase-to-phase short circuit
combined with a turn-to-turn fault are illustrated in Figure 5.12. The slip values of
both faults rise quickly while the load level increases. In addition, under different load
levels, the slip value of a phase-to-phase short circuit fault is always higher than that

Page 59
5.4 Off-line Test of Stator Short Circuit Faults

Current in phase 1,2,3 Magnitude plot of current in phase 1,2,3

60
0.04

Magnitude response (dB)

X:26.4 X.73.6
40
Y:0.2273 Y:−3.001
0.02 20
Current (A)

0
0
−20

−0.02 −40
−60 Sampling frequency: 1000 Hz
−0.04 Window length: 1024
−80
2150 2200 2250 2300 0 20 40 60 80 100
Sample no. Frequency (Hz)
Interpolated zero−crossing intervals Magnitude plot of zero−cross interval in phase 1,2,3
−4
x 10
1.5 0
Mean zero−cross interval (s)

Magnitude response (dB)

1 −20 X:23.73
Y:−48.3
0.5 −40
0
−60
−0.5
−80
−1
Sampling frequency: 1000 Hz
−100 Window length: 1024
−1.5
0 200 400 600 800 1000 0 50 100 150
Data blocks Frequency (Hz)

Figure 5.11. Phase-to-phase fault with 140% loading level.

Table 5.3. Phase-to-phase fault.

Amplitude of Amplitude of
Load Fault Component Fault component
frequency frequency Slip
Level Frequency (n=1 frequency (n=1
component (n=1 component (n=1 (%)
(%) k=+1) k=-1)
k=+1) k=-1)
0 25 -13.11 75 -16.63
80 25.5 -1.739 74.5 -0.8316 2
90 25.8 -0.3755 74.2 -3.423 3.2
110 26.1 -0.2722 74 -6.349 4
140 26.4 0.2273 73.6 -3.001 5.6

Page 60
Chapter 5 Stator Short Circuit Fault

Table 5.4. Phase-to-phase fault plus turn-to-turn fault.

Amplitude of Amplitude of
Load Fault Component Fault component
frequency frequency Slip
Level Frequency (n=1 frequency (n=1
component (n=1 component (n=1 (%)
(%) k=+1) k=-1)
k=+1) k=-1)
0 25.1 -14.75 74.9 -8.727
80 25.6 -11.05 74.3 -20.47 2.8
90 25.9 -5.132 74.1 -12.94 3.6
110 26.2 -11.37 73.7 -16.47 5.2
140 26.6 -14.57 73.4 -18.37 6.4

for the motor with a phase-to-phase fault plus a turn-to-turn fault.

5
Slip (%)

2
Phase−to−phase short +turn−to−turn short
Phase−to−phase short

1
80 90 100 110 120 130 140
Load level (%)

Figure 5.12. Slip of phase-to-phase short circuit and turn-to-turn short circuit under different load
level.

The slip variations of the healthy motor, the motor with a phase-to-phase fault and
the motor with a serious fault (phase-to-phase fault combined with turn-to-turn fault),
are plotted in Figure 5.13. In general, the slip values of a serious fault are higher than
the slip for only a phase-to-phase fault. However, this does not appear as expected in
Figure 5.13. On the contrary, the slip values of a serious fault are lower than those of a
phase-to-phase fault. The possible reason is that there are some reverse voltages due to
the addition of a turn-to-turn fault. Especially, when the motor runs at 80% load level,

Page 61
5.4 Off-line Test of Stator Short Circuit Faults

the slip for a serious fault is lower than that of a healthy motor.

0
The amplitude of rotor frequency (dB)

−5

−10

−15

−20

Phase−to−phase short
−25
Phase−to−phase short +turn−to−turn short
Healthy motor

−30
0 20 40 60 80 100 120 140
Load level (%)

Figure 5.13. The amplitude of rotor frequency under different load levels.

The comparison of a turn-to-turn fault and a phase-to-phase fault test results are pre-
sented in Figure 5.14. As expected in theory, the amplitude of the fault frequency
(when k = 1 in equation (5.1)) of a phase-to-phase short circuit is higher than that for a
turn-to-turn fault even when the motor runs with high load level. In addition, the am-
plitudes of the fault frequencies climb significantly in these three conditions. However,
there are slight fluctuations when the healthy motor and the motor with a turn-to-turn
short circuit fault run between 90% and 110% load level.

The two kinds of serious faults’ testing results are also compared in Figure 5.15. The
amplitudes of the fault frequencies (when k = 1 in equation (5.1)) of a serious phase-
to-phase short circuit fault are higher than those of a serious turn-to-turn short circuit
fault. When the motor operates without load, the amplitude of the fault frequency of
the healthy motor is the lowest. The amplitude of frequency of the serious phase-to-
phase short circuit is higher than that for the healthy motor when the motor operates
below 70% of the full load level.

Page 62
Chapter 5 Stator Short Circuit Fault

The amplitude of rotor frequency (dB)

−5

−10

−15

−20

Phase−to−phase fault
Turn−to−turn fault
−25 Healthy motor

−30
0 20 40 60 80 100 120 140
Load level (%)

Figure 5.14. The amplitude of rotor frequency of phase-to-phase fault and turn-to-turn fault under
different load levels.

−5

−10
The amplitude of rotor frequency (dB)

−15

−20

−25
Phase−to−phase short + turn−to−turn fault
Turn−to−turn fault (five turns)
Healthy motor

−30
0 20 40 60 80 100 120 140
Load level (%)

Figure 5.15. The amplitude of rotor frequency of serious faults under different load level.

Page 63
5.5 On-line Test of Stator Short Circuit Faults

5.5 On-line Test of Stator Short Circuit Faults

The on-line testing system is described in the section 5.2. During the test, the motor
operating parameters are recorded by the ZCT signal converter device and processed
by Labview.

5.5.1 Analysis of Turn-to-turn Fault of Motor II

For the turn-to-turn short circuit experiment, the insulation of three points in two ad-
jacent turns of the same coil is scratched. Then, three wires are soldered to these three
points. Through a control box, artificial turn-to-turn faults are simulated. Figure 5.16
and 5.17 show the results for wires shorted directly and wires connected via a 5.88 kΩ
resistance. Figure 5.16 reveals that when the torque level is equal to zero, the ampli-
tude of the no fault motor is the lowest and the amplitude of wire 1 and wire 3 short
circuit without resistance is the highest. Due to the noise caused by a high load level,
there is obvious fluctuation for the amplitude of the stator short circuit without resis-
tance except or the wire 1 and wire 2 short circuit and wire 1 and wire 3 short circuit.
Furthermore, for all fault condition the amplitude values go down after fluctuation as
the load level increases. In addition, when the load level is zero, the resistance, added
in the stator short circuit, decreases the amplitude of the wire 1 and wire 3 short circuit
by nearly 10 dB. However, the amplitude of the wire 2 and wire 3 short circuit is still
higher than the wire 1 and wire 2 short circuit.

The amplitude fluctuation is obvious in the phase-phase stator short circuit without
resistance, as shown in Figure 5.18. In Figure 5.19, the amplitude of the no fault motor
is lower than the two conditions of stator short circuit even at high torque level. The
amplitude of the phase 1-2 and phase 2-1 short circuit is lower than that for the phase
1-1 and phase 2-1 short circuit except when the torque level is at zero. Comparing
Figure 5.18 with 5.19, we can conclude that the resistance effect on the phase-to-phase
stator short circuit is more significant in Motor II.

Temperature monitoring is another effective method to monitor the motor operating

state. Increased temperature could be related to overload, high environment temper-
ature or motor faults, such as a stator short circuit. The high temperature in a motor

Page 64
Chapter 5 Stator Short Circuit Fault

−15
Misalignment + no stator short
Misalignment + turn−to−turn short(1,2)directly
Misalignment + turn−to−turn short(1,3)directly
Misalignment + turn−to−turn short(2,3)directly
−20 Misalignment + turn−to−turn short(1,2,3)directly

The amplitude of totation frequency(dB)

−25

−30

−35

−40

−45
0 20 40 60 80 100 120 140 160 180 200
Load level (%)

Figure 5.16. Comparison of testing results of turn-to-turn stator short circuit without resistance.
−15

−20
The amplitude of roto frequency (dB)

−25

−30

−35

Misalignment+no stator short

Misalignment + turn−to−turn short(1,2) with resistance
−40 Misalignment + turn−to−turn short(1,3)with resistance
Misalignment + turn−to−turn short(2,3)with resistance
Misalignment + turn−to−turn short(1,2,3)with resistance

−45
0 20 40 60 80 100 120 140 160 180 200
Load level (%)

Figure 5.17. Comparison of testing results of turn-to-turn stator short circuit with resistance.

could also reduce the insulation life and result in a permanent failure.

Figure 5.20 illustrates the motor temperature during a stator turn-to-turn short circuit
fault without resistance. The temperature with the 82.24% load increases more sharply
compared to the motor without load. The temperature with 82.84% load arrived at
95◦ C in 13 min. For the motor in our lab, the temperature tolerance is 150◦ C. If the
temperature exceeds 150◦ C, the insulation could be damaged.

Page 65
5.6 Summary

−24
Misalignment+no stator short
Misalignment + phase 1−1 to phase 2−1 short directly
−26
Misalignment + phase 1−2 to phase 2−1 short directly

−28

The amplitude of rotor frequency (dB)

−30

−32

−34

−36

−38

−40

−42

−44
0 20 40 60 80 100 120 140 160 180 200
Load level (%)

Figure 5.18. Comparison of testing results of phase-to-phase stator short circuit without resistance.

−15

−20
The amplitude of rotor frequency (dB)

−25

−30

−35

−40 Misalignment+no stator short

Misalignment + phase 1−1 to phase 2−1 short with resistance
Misalignment + phase 1−2 to phase 2−1 short with resistance

−45
0 20 40 60 80 100 120 140 160 180 200
Load level (%)

Figure 5.19. Comparison of testing results of phase-to-phase stator short circuit with resistance.

5.6 Summary
The experimental results are in good agreement with the theoretical calculations. Through
the analysis of the current or the ZCT signal, a stator short circuit fault can be detected
in both the off-line and on-line test. In addition, temperature monitoring can also be
considered as a supplementary method to monitor the motor operating condition.

Page 66
Chapter 5 Stator Short Circuit Fault

100

90

80

70
Temperature ( o C)

60
short with noload
50
short with 82.24% load
40

30

20

10

0
0 200 400 600 800 1000
Time (Second)

Figure 5.20. Temperature of turn-to-turn short circuit fault.

Page 67
Page 68
Chapter 6

Broken Rotor Bar Fault

T
WO techniques, MCSA and ZCT are implemented to diagnose
induction motor broken rotor bar faults. Through comparison of
their results, the ZCT method is verified as an effective approach.

Page 69
6.1 Introduction

6.1 Introduction
Rotor bar faults occupy about 10% of overall fault conditions in an induction motor.
It is usually caused by the poor operating conditions such as high temperatures, high
mechanical loading especially in the starting time [32]. In addition, the problems of
the manufacturing process, like defective casting, or poor jointing can also cause rotor
broken bars. Initially, these problems were presented as an increase of resistance and
temperature. Then they progressed as cracking or small holes in the rotor bars. The
rotor bars which are near the cage end rings are vulnerable.

In this section, the data of a symmetrical motor and a motor with a broken rotor bars
fault were provided by laboratory of the Cracow University of Technology. The motor
parameters and the test data are shown in Appendices D and E, respectively.

6.2 Broken Rotor Bar Fault Detection Techniques

The frequency components that are related to a broken rotor bar ( BRB) condition for
slip s [10] are
f BRB = (1 ± 2ks) f 0 (6.1)

where k = 1, 2, 3, 4, ....

Dominant BRB frequencies (k = 1) are very close to the fundamental frequency f 0 ,

and this combined with low signal to noise ratio makes the task of detecting a BRB
condition difficult. If the load is time-varying the spectral estimation should be done
using a very short data window. Methods based on the Discrete Fourier Transform
(DFT) are not able to detect frequencies close to the fundamental when using a very
short data window. By monitoring spacing between successive zero crossings, the ZCT
method could increase the detection accuracy.

6.3 Analysis of Broken Rotor Bars

The detection for the following cases with different load levels are carried out:

Page 70
Chapter 6 Broken Rotor Bar Fault

Table 6.1. Motor parameters.

Parameter Value
Active power PN [kW ] 0.8
Nominal voltage UN [V ] 380
Nominal current [ A] 2.2
Nominal power factor cos ϕ 0.74
Rotor speed [rpm] 1400
No load speed [rpm] 1497
Winding connection Y
Number of poles per phase winding p 2
Nominal frequency [ Hz] 50
Number of rotor bars 22
Number of stator slots 24
Rotor inertia [kg ∗ m2 ] 0.0025
Coil number per phase 4
Turn number of coil 90

A) symmetrical motor

B) one broken rotor bar

C) two broken rotor bars

D) three broken rotor bars

The parameters of the motor, which is used in the laboratory are listed in the Table 6.1.

Current and ZCT signals are recorded for the symmetrical motor with a 81.36% load
level. The FFT results are shown in Figures 6.1 and 6.2. The current signal and its
spectra are presented on the top of Figure 6.1. The ZCT signals of the current signals
are computed by MATLAB and plotted in the bottom of Figure 6.1. The synchronous
speed is 1500 rpm. For the rotor measurement speed of 1450 rpm, the slip can be cal-
culated as s = (1500 − 1450)/1500 = 0.033. The peak is presented in ZCT spectra of
Figure 6.1 and the amplitude of this peak is very low. Figure 6.2 shows the ZCT signal,
which is acquired directly in the monitoring process. According to Figures 6.1 and 6.2,
there are no obvious peaks appearing at sideband frequencies (1 ± 2s) f 0 . Therefore,

Page 71
6.3 Analysis of Broken Rotor Bars

we can regard this motor is running under normal conditions.

Current in phase 1,2,3 Magnitude plot of current in phase 1,2,3

3 120
X:50

Magnitude response (dB)

Y:89.52
2 100
Current (A)

1
80
0
60
−1

−2 40
Sampling frequency: 1000 Hz
Window length: 1024
−3 20
4750 4800 4850 4900 45 50 55
Sample no. Frequency (Hz)
Interpolated zero−crossing intervals Magnitude plot of zero−cross interval in phase 1,2,3
−4
x 10
2 0
Mean zero−cross interval (s)

Magnitude response (dB)

Sampling frequency: 1000 Hz

1.5 −20 Window length: 1024

X:3.223
1 −40 Y:−67.75
0.5 −60

0 −80

−0.5 −100

−1 −120
0 200 400 600 800 1000 0 10 20 30 40 50
Data blocks Frequency (Hz)

Figure 6.1. Current and ZCT signal from healthy motor (81.36% load level).

Figure 6.3 shows the symmetrical motor detection results when the load current is at
2.23 A (101.36% load level). The current data and ZCT signal spectra are computed by
MATLAB. When the load current of the symmetrical motor is 2.23 A, the rotor speed is
measured as 1419 rpm. The slip can be calculated as s = (1500 − 1419)/1500 = 0.054.
The 2s f peak is revealed in the ZCT spectra of Figure 6.3. Figure 6.4 shows the analysis
from the ZCT signal directly acquired during the test. Although there is no broken bar
fault, there are small peaks appearing at sideband frequencies (1 ± 2s) f in the current
spectra of Figure 6.3. The possible reason is that there is a slight crack in the rotor bar.

For the case shown in Figure 6.5, there is one artificial broken rotor bar in the test motor,
the measured rotor speed is 1394 rpm. Then, the slip value is s = (1500 − 1394)/1500 =
0.07067 and the 2s f = 100s = 7.067 Hz. Figure 6.5 illustrates that peaks appear at side-
band frequencies (1 ± 2s) f in the current spectra and the 2s f peak is showed in the
ZCT spectra, corresponding to the computed 2s f . The 2s f peak appears in the directly
recorded ZCT signal in Figure 6.6 the same as in the computed ZCT in Figure 6.5. The

Page 72
Chapter 6 Broken Rotor Bar Fault

x 10
−3 (i) Interpolated zero−crossing intervals
1.5

Mean zero−cross interval (s) 1

0.5

−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase

−20
Magnitude response (dB)

Sampling frequency: 1000 Hz

Window length: 1024
−40 X:3.223
Y:−62.57

−60

−80

−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 6.2. Direct recorded ZCT signal from healthy motor (81.36% load level).

Current in phase 1,2,3 Magnitude plot of current in phase 1,2,3

120
X:50
Magnitude response (dB)

3 Y:92
100
2
Current (A)

1 80 X:44.6 X:55.44
0 Y:47.87 Y:45.95

−1 60

−2
40
−3 Sampling frequency: 1000 Hz
Window length: 1024
20
3900 3950 4000 4050 45 50 55
Sample no. Frequency (Hz)
Interpolated zero−crossing intervals Magnitude plot of zero−cross interval in phase 1,2,3
−4
x 10
1.5 0
Mean zero−cross interval (s)

Magnitude response (dB)

Sampling frequency: 1000 Hz

1 −20 X:5.566 Window length: 1024
Y:−62.69
0.5 −40

0 −60

−0.5 −80

−1 −100

−1.5 −120
0 200 400 600 800 1000 0 10 20 30 40 50
Data blocks Frequency (Hz)

Figure 6.3. Current and ZCT signal from healthy motor (101.36% load level).

Page 73
6.3 Analysis of Broken Rotor Bars

x 10
−3 (i) Interpolated zero−crossing intervals
Mean zero−cross interval (s) 1.5

0.5

−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase

−20
Magnitude response (dB)

X:5.42 Sampling frequency: 1000 Hz

Window length: 1024
−40 Y:−54

−60

−80

−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 6.4. Direct recorded ZCT signal from healthy motor (101.36% load level).

magnitude of 2s f of the directly recorded ZCT signal is 4 dB higher than that of the cal-
culated ZCT signal under the same sampling frequency (1000 Hz) and window length
(1024). The second harmonic 2s f could also be observed in Figure 6.6. There are a few
glitches in the ZCT measurement in the top plot of Figure 6.6. This may result from the
small interruption during the measurement. However, these do not affect the 2s f peak
detection.

The test results of the two artificial broken rotor bars fault are illustrated in Figure 6.7
and 6.8. The rotor speed is 1415 rpm. According to equation ( 6.1), the computed slip
is 0.0567. Thus, 2s f is equal to 5.67 Hz. In Figure 6.7, the two peaks of the sideband
frequencies (1 ± 2s) f are showed in the current spectra and the 2s f peak is showed in
the ZCT spectra. The 2s f peak and second harmonic peak of the directly recorded ZCT
signal are also observed in Figure 6.8.

The signals acquired directly from the ZCT signal converter are summarized in the fol-
lowing tables. Table 6.2 illustrates the measurement results of the symmetrical motor
with different load levels. Tables 6.3, 6.4 and 6.5 present the measurement results of the

Page 74
Chapter 6 Broken Rotor Bar Fault

Current in phase 1,2,3 Magnitude plot of current in phase 1,2,3

4 120

Magnitude response (dB)

X:50
Y:89.77
100
2
Current (A)

X:42.96
80
Y:57.31 X:56.96
0
Y:49.05
60
−2
40
Sampling frequency: 1000 Hz
Window length: 1024
−4
20
5900 5950 6000 40 45 50 55 60
Sample no. Frequency (Hz)
Interpolated zero−crossing intervals Magnitude plot of zero−cross interval in phase 1,2,3
−4
x 10
2 0
Mean zero−cross interval (s)

X:7.031

Magnitude response (dB)

−20 Y:−45.12 X:13.77
1 X:−48.66
−40

0 −60

−80
−1
−100 Sampling frequency: 1000 Hz
Window length: 1024
−2 −120
0 200 400 600 800 1000 0 10 20 30 40 50
Data blocks Frequency (Hz)

Figure 6.5. Current and ZCT signal from motor with one BRB.

x 10
−3 (i) Interpolated zero−crossing intervals
1.5
Mean zero−cross interval (s)

0.5

−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase

Magnitude response (dB)

−20 X:6.885
Y:−40.78 X:13.92
Y:−40.66
−40

−60

−80 Sampling frequency: 1000 Hz

Window length: 1024
−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 6.6. Direct recorded ZCT signal from motor with one BRB.

Page 75
6.3 Analysis of Broken Rotor Bars

Current in phase 1,2,3 Magnitude plot of current in phase 1,2,3

120

Magnitude response (dB)

3 X:50
100 Y:86.8
2
Current (A)

1 X:44.36 X:55.64
80
Y:58 Y:56.58
0
−1 60

−2
40
−3 Sampling frequency: 1000 Hz
Window length: 1024
20
5150 5200 5250 40 45 50 55
Sample no. Frequency (Hz)
Interpolated zero−crossing intervals Magnitude plot of zero−cross interval in phase 1,2,3
−4
x 10
1.5 0
Mean zero−cross interval (s)

X:5.566

Magnitude response (dB)

1 −20 Y:−40.94 X:11.13
Y:−54.51
0.5 −40

0 −60

−0.5 −80

−1 −100 Sampling frequency: 1000 Hz

Window length: 1024
−1.5 −120
0 200 400 600 800 1000 0 10 20 30 40 50
Data blocks Frequency (Hz)

Figure 6.7. Current and ZCT signal from motor with two BRB.

(i) Interpolated zero−crossing intervals

−3
x 10
1.5
Mean zero−cross interval (s)

0.5

−0.5

−1

−1.5
0 100 200 300 400 500 600 700 800 900 1000
Data blocks

(ii) Magnitude plot of zero−cross interval in phase

Magnitude response (dB)

−20 X:5.566
Y:−36.43 X:11.28
Y:−51.54
−40

−60

−80
Sampling frequency: 1000 Hz
Window length: 1024
−100
0 10 20 30 40 50 60 70
Frequency (Hz)

Figure 6.8. Direct recorded ZCT signal from motor with two BRB.

Page 76
Chapter 6 Broken Rotor Bar Fault

Table 6.2. Symmetrical motor measurement results.

Peak Peak Aver-

Peak
Peak
Measured value value age
Loading 2sf value value
data file Slip (dB) (Hz) (dB) rotor
(%) (Hz) (Hz)
(ZCT (ZCT (ZCT
Name (ZCT ex- speed
extraction) acquired acquired
traction)
directly) directly) (rpm)
S1 002.dat 0.03333 81.3636 3.333 3.2227 -67.7536 3.2227 -62.5701 1450
S1 003.dat 0.04333 90.9091 4.333 4.3945 -66.5511 4.248 -58.8159 1435
S1 004.dat 0.054 101.3636 5.4 5.5664 -62.6907 5.4199 -54.0002 1419
S1 005.dat 0.066 112.7273 6.6 6.4453 -60.4214 6.5918 -51.6884 1401

Table 6.3. One broken rotor bar fault measurement results.

Peak Peak Aver-

Peak
Peak
Measured value value age
Loading 2sf value value
data file Slip (dB) (Hz) (dB) rotor
(%) (Hz) (Hz)
(ZCT (ZCT (ZCT
Name (ZCT ex- speed
extraction) acquired acquired
traction)
directly) directly) (rpm)
S1 007.dat 0.03267 82.2727 3.267 3.2227 -56.4234 6.4453 -49.9055 1451
S1 008.dat 0.048 95.4545 4.8 4.6875 -49.8731 4.6875 -44.4785 1428
S1 009.dat 0.058 105 5.8 5.859 -47.61 5.8594 -40.6925 1413
S1 010.dat 0.07067 115.9091 7.067 7.031 -39.32 6.8848 -40.7818 1394

motor with one BRB, two BRBs and three BRBs at different load levels, respectively.
From the tables, we can find that there is not a big difference between the extracted
ZCT signal from current signal by MATLAB and the directly acquired ZCT signal. The
amplitude of 2s f increases in these four states when the load level rises.

The slip variation of the symmetric motor and the motor with one, two and three BRB
faults is represented in Figure 6.9. Under these four conditions, the value of the slip
increases with an increase in the load level. In theory, the more broken bars appear,
the larger slip will be observed. However, the slip of one broken rotor bar fault is
bigger than that of two broken rotor bars when the load level is bigger than 85% load
level. Furthermore, the slip of one broken rotor bar fault should be bigger than that of
a healthy motor. On the contrary, the smallest slip value of one broken rotor bar fault

Page 77
6.3 Analysis of Broken Rotor Bars

Table 6.4. Two broken rotor bar fault measurement results.

Peak Peak Aver-

Peak Peak
Measured value value age
Loading 2sf value value
data file Slip (dB) (Hz) (dB) rotor
(%) (Hz) (Hz)
(ZCT (ZCT (ZCT
Name (ZCT ex- speed
extraction) acquired acquired
traction)
directly) directly) (rpm)
S1 012.dat 0.03467 85 3.467 3.5156 -53.234 3.3692 -44.4412 1448
S1 013.dat 0.046 95 4.6 4.3945 -48.6896 4.541 -38.3987 1431
S1 014.dat 0.05667 105 5.667 5.5664 -40.937 5.5664 -36.4333 1415
S1 015.dat 0.068 115.9091 6.8 6.7383 -38.0139 6.7383 -33.2382 1398

Table 6.5. Three broken rotor bar fault measurement results.

Peak Peak Aver-

Peak Peak
Measured value value age
Loading 2sf value value
data file Slip (dB) (Hz) (dB) rotor
(%) (Hz) (Hz)
(ZCT (ZCT (ZCT
Name (ZCT ex- speed
extraction) acquired acquired
traction)
directly) directly) (rpm)
S1 017.dat 0.036 82.2727 3.6 3.5156 -54.5292 3.51565 -50.3325 1446
S1 018.dat 0.05067 95 5.067 4.9805 -44.55 4.98045 -41.9625 1424
S1 019.dat 0.06 103.1818 6 5.8594 -46.6981 6.00585 -40.86 1410
S1 020.dat 0.07267 114.0909 7.267 7.3242 -45.502 7.17775 -36.7028 1391

is obtained under 80% load level. The abnormal behavior of one broken rotor bar slip
illustrates there might be some problems in the rotor speed measurement.

Figure 6.10 represents a comparison of the variation of the 2s f peaks magnitude when
the motor runs in the different fault conditions at several load levels. The increasing
load level leads to an increase in frequency amplitude at 2s f peaks under all motor
running conditions. According to Figure 6.10, the rotor broken bars fault could be
diagnosed by monitoring the 2s f peaks amplitude [47].

Page 78
Chapter 6 Broken Rotor Bar Fault

7.5

Symmetrical motor
7 One broken rotor bar
Two broken rotor bars
Three broken rotor bars
6.5

Slip (%)
5.5

4.5

3.5

3
80 85 90 95 100 105 110
Load level (%)

Figure 6.9. The relationship between slip and load level.

−30
Symmetrical motor
One bar broken fault
−35 Two bars broken fault
Three bars broken fault

−40
Amplitude of 2sf frequency (dB)

−45

Broken rotor bar fault

condition
−50

Healthy condition
−55

−60

−65
85 90 95 100 105 110 115
Load level (%)

Figure 6.10. The relationship between the amplitude of rotor frequency and load level.

6.4 Summary

Different numbers of rotor broken bars are studied with several load levels. The faults
are analyzed and diagnosed by using the current signal and two ZCT signals, one com-
puted by MATLAB from the current signal and another acquired directly from the ZCT
signal converter device. The peaks of the (1 ± 2s) f sideband frequencies are very close
to the fundamental frequency and difficult to be represented in the current spectra. In

Page 79
6.4 Summary

contrast, the ZCT method, which diagnoses BRB by monitoring the 2s f peak, is an ef-
fective approach. Even if there is a small fault, the 2s f peaks are represented in the zero
crossing time signal. In our experiments, the magnitude of 2s f of the directly acquired
ZCT signal was higher than that of the calculated ZCT signal.

Page 80
Chapter 7

Thesis Summary and

Recommendation for
Future Work

7.1 Thesis Summary

Two steady-state fault detection methods, MCSA and ZCT, have been utilized to di-
agnose stator short circuit faults and broken rotor bar faults. After a literature review
of these two common faults and two commonly used diagnosis methods, the baseline
analysis was performed using two healthy motors. Through the baseline study, signal
processing programs, using MATLAB and LabVIEW, the motor healthy states are ver-
ified to ensure the accurate diagnosis of motor faults.

Motor misalignment is tested and adjusted by using the Laser Alignment system. The
experimental results show that improper motor installation will affect motor fault di-
agnosis. For example, a coupling gap can lead to an unexpected frequency peak; mis-
alignment will increase the amplitude of the fault frequency in a stator short circuit
fault at low load levels. Improper motor installation, combined with manufacture tol-
erance and working environment, may lead to an unexpected difference in motors of
the same type. This will decrease the accuracy of motor faults diagnosis.

Based on these preparations, motor stator short circuit faults are comprehensively
studied by using both off-line and on-line tests. Through a control box, artificial turn-
to-turn faults and phase-to-phase faults are simulated. Furthermore, the experiments
are implemented under different load levels. The experimental results are in good
agreement with the theoretical calculations. Compared with the conventional MCSA

Page 81
7.2 Future Work

method, the ZCT method has the advantages of lower sampling rate and lower com-
putational burden. In the experiments, the ZCT method shows frequency components,
which are related to motor faults. In others words, the accuracy of the ZCT method is
comparable to the conventional MCSA.

In addition, these two method are also used to analysis recorded stator current data
provided by the machine laboratory of the Cracow University of Technology. The data
includes stator current of a symmetrical motor and motor with broken rotor bars fault.
The results show that the ZCT method can be used reliably for detecting broken rotor
bars faults.

7.2 Future Work

Most work in this thesis is from the experimental point of view. The recorded stator
circuit data is analyzed using MCSA and ZCT methods to diagnose stator short circuit
faults and broken rotor bar faults. FFT is the main and only type of analysis method
used in this work. The study can be expanded further by employing similar or more
advanced signal processing techniques such as wavelet transforms. Some parametric
spectrum analysis methods as well as that of modern non-parametric techniques might
reveal more useful information and enable easier fault diction. These methods might
automatically detect faults and allow for better diagnostic of faults. Furthermore, some
pattern recognition techniques, such as principal component analysis (PCA) and inde-
pendent component analysis (ICA), could reveal ”hidden” components of the ZCT sig-
nals, which can improve the analysis of motor current signals. Moreover, the investi-
gation can be undertaken different types of induction motors with different ratings. In
addition, the further simulation work could enhance the understanding of significant
frequencies related to the induction motor faults and improve the overall effectiveness
of the fault detection.

Page 82
References
[1] C. I. Hubert, Electric machines: theory, operation, applications, adjustment and control.
Merrill, 1991, ch. 2.

[2] C. Ong, Dynamic Simulation of Electric Machinery: Using MATLAB/SIMULINK.

Prentice Hall PTR, 1998, ch. 6.

[3] G. C. Stone, E. A. Boulter, I. Culbert, and H. Dhirani, Electrical Insulation For Rotat-
ing Machines-Design, Evaluation, Aging, Testing and Repair. IEEE Press and A John
Wiley & Sons, INC., 2004, ch. 4.

[4] K. Wang, Intelligent Condition Monitoring And Diagnosis Systems A Computational

Intelligence Approach. IOS Press, 2003, ch. 3.

[5] Y. Han and Y. Song, “Condition monitoring techniques for electrical equipment -
a literature survey,” Power Delivery, IEEE Transactions on, vol. 18, no. 1, pp. 4–13,
Jan 2003.

[6] G. Singh and S. A. Kazzaz, “Induction machine drive condition monitoring and
diagnostic research - a survey,” Electric Power Systems Research, vol. 64, pp. 145–
158, 2003.

[7] J. Faiz and M. Ojaghi, “Different indexes for eccentricity faults diagnosis in three-
phase squirrel-cage induction motors: A review,” Mechatronics, vol. 19, no. 1, pp.
2–13, 2009.

[8] S. Nandi, H. A. Toliyat, and X. Li, “Condition monitoring and fault diagnosis of
electrical motors - a review,” IEEE Transactions on Energy Conversion, vol. 20, no. 4,
pp. 719 – 729, 2005.

[9] W. Deleroi, “Squirrel cage motor with broken bar in the rotor -physical phenom-
ena and their experimental assessment,” in International Conference on Emergency
Medicine (ICEM 1982), Budapest, Hungary, 1982, pp. 767–770.

[10] N. Elkasabgy, A. Eastham, and G. Dawson, “Detection of broken bars in the cage
rotor on an induction machine,” Industry Applications, IEEE Transactions on, vol. 28,
no. 1, pp. 165 –171, Jan. 1992.

Page 83
REFERENCES

[11] J.-H. Jung, J.-J. Lee, and B.-H. Kwon, “Online Diagnosis of Induction Motors Using
MCSA,” Industrial Electronics, IEEE Transactions on, vol. 53, no. 6, pp. 1842 –1852,
Dec. 2006.

[12] M. Benbouzid, “A review of induction motors signature analysis as a medium for

faults detection,” in Proceedings of the 24th Annual Conference of the IEEE Industrial
Electronics Society, 1998. IECON ’98, vol. 4, 1998, pp. 1950 –1955.

[13] G. Kliman, R. Koegl, J. Stein, R. Endicott, and M. Madden, “Noninvasive detection

of broken rotor bars in operating induction motors,” IEEE Transactions on Energy
Conversion, vol. 3, no. 4, pp. 873 –879, Dec. 1988.

[14] A. Bellini, F. Filippetti, G. Franceschini, C. Tassoni, and G. Kliman, “Quantitative

evaluation of induction motor broken bars by means of electrical signature analy-
sis,” IEEE Transactions on Industry Applications, vol. 37, pp. 1248 –1255, Sep. 2001.

[15] B. Gaydon, “Instrument to detect induction motor rotor circuit defects by speed
fluctuation measurements.” in International Conference on Advances in Medical Sig-
nal and Information Processing, no. 174, 1979, pp. 5 – 8.

[16] F. Filippetti, G. Franceschini, C. Tassoni, and P. Vas, “Ai techniques in induction

machines diagnosis including the speed ripple effect,” IEEE Transactions on Indus-
try Applications, vol. 34, no. 1, pp. 98 –108, Jan. 1998.

[17] A. Sadoughi, M. Ebrahimi, and E. Razaei, “A new approach for induction motor
broken bar diagnosis by using vibration spectrum,” in International Joint Confer-
ence SICE-ICASE, 2006., Oct. 2006, pp. 4715 –4720.

[18] K. Cho, J. Lang, and S. Umans, “Detection of broken rotor bars in induction motors
using state and parameter estimation,” IEEE Transactions on Industry Applications,
vol. 28, no. 3, pp. 702 –709, May 1992.

[19] M. G. Melero, M. F. Cabanas, F. R. Faya, C. H. Rojas, and J. Solares, “Electromag-

netic torque harmonics for on-line inter turn short circuits detection in squirrel
cage induction motors,” in 8th European Conf. on Power Electronics and Appl. EPE,
Lausanne, Switzerland, Sep 1999, p. 9.

Page 84
 REFERENCES

[20] D. Dorrell, W. Thomson, and S. Roach, “Analysis of airgap flux, current, and vi-
bration signals as a function of the combination of static and dynamic airgap ec-
centricity in 3-phase induction motors,” IEEE Transactions on Industry Applications,
vol. 33, no. 1, pp. 24 –34, Jan 1997.

[21] S. Fruechtenicht, E. Pittius, and H. Seinsch, “Diagnostic system for three-phase

asynchronous machines,” in International Conference on Advances in Medical Signal
and Information Processing, no. 310, 1989, pp. 163 – 171.

[22] A. Siddique, G. Yadava, and B. Singh, “A review of stator fault monitoring tech-
niques of induction motors,” IEEE Transactions on Energy Conversion, vol. 20, no. 1,
pp. 106–114, March 2005.

[23] H. Su and K. T. Chong, “Induction machine condition monitoring using neural

network modeling,” IEEE Transactions on Industrial Electronics, vol. 54, no. 1, pp.
241–249, Feb. 2007.

[24] M. Benbouzid and H. Nejjari, “A simple fuzzy logic approach for induction mo-
tors stator condition monitoring,” in IEEE International Electric Machines and Drives
Conference, IEMDC, 2001, pp. 634–639.

[25] B. Ayhan, M.-Y. Chow, and M.-H. Song, “Multiple discriminant analysis and
neural-network-based monolith and partition fault-detection schemes for broken
rotor bar in induction motors,” IEEE Transactions on Industrial Electronics, vol. 53,
no. 4, pp. 1298 –1308, June 2006.

[26] L. Cristaldi, M. Lazzaroni, A. Monti, F. Ponci, and F. Zocchi, “A genetic algorithm

for fault identification in electrical drives: a comparison with neuro-fuzzy compu-
tation,” in Proceedings of the 21st IEEE Instrumentation and Measurement Technology
Conference, 2004. IMTC 04., vol. 2, May 2004, pp. 1454–1459.

[27] F. Filippetti, G. Franceschini, and T. Carla, “Neural networks aided on-line diag-
nostics of induction motor rotor faults,” in Conference Record - IAS Annual Meeting
(IEEE Industry Applications Society), vol. 1, 1993, pp. 316 – 323.

[28] M. Haji and H. Toliyat, “Pattern recognition - a technique for induction machines
rotor broken bar detection,” IEEE Transactions on Energy Conversion, vol. 16, no. 4,
pp. 312 –317, Dec. 2001.

Page 85
REFERENCES

[29] A. da Silva, R. Povinelli, and N. Demerdash, “Induction machine broken bar

and stator short-circuit fault diagnostics based on three-phase stator current en-
velopes,” IEEE Transactions on Industrial Electronics, vol. 55, no. 3, pp. 1310 –1318,
March 2008.

[30] P. Vas, Parameter Estimation, Condition Monitoring, and Diagnosis of Electrical Ma-
chines. Clarendron Press Oxfords, 1993, ch. 3.

[31] W. Thomson and M. Fenger, “Current signature analysis to detect induction motor
faults,” IEEE Industry Applications Magazine, vol. 7, no. 4, pp. 26 –34, Jul/Aug 2001.

[32] H. Çalış and A. Çaklır, “Rotor bar fault diagnosis in three phase induction motors
by monitoring fluctuations of motor current zero crossing instants,” Electric Power
Systems Research, vol. 77, no. 5-6, pp. 385–392, 2007.

[33] D. Dorrell, W. Thomson, and S. Roach, “Analysis of airgap flux, current, and vi-
bration signals as a function of the combination of static and dynamic airgap ec-
centricity in 3-phase induction motors,” IEEE Transactions on Industry Applications,
vol. 33, no. 1, pp. 24–34, Jan. 1997.

[34] W. Thomson, “A review of on-line condition monitoring techniques for three

phase squirrel cage induction motors-past, present and future,” in SDEMPED’99,
Gijon, Spain,, vol. 33, no. 1, Sep. 1999, pp. 3–18.

[35] Y. Wang, “The ZCT method of induction motor failure prediction and speed mon-
itoring,” Ph.D. dissertation, University of Sussex, England, 1997.

[36] A. Ukil, S. Chen, and A. Andenn, “Detection of stator short circuit faults in three-
phase induction motors using motor current zero crossing instants,” Electric Power
System Research, 2009, to be published.

[37] Fixturlaser Company. [Online]. Available: www.fixturlaser.com/products-

services/fixturlaser-product-range/fixturlaser-xa-2.aspx

[38] R. N. Brancewell, The Fourier Transform and Its Applications, 3rd ed. McGraw-Hill,
2000, ch. 2.

[39] H. Razik, M. B. R. Correa, and E. R. C. Da Silva, “An application of genetic algo-

rithm and fuzzy logic for the induction motor diagnosis,” in 34th Annual Confer-
ence of IEEE Industrial Electronics, 2008. IECON 2008., Nov. 2008, pp. 3067–3072.

Page 86
Appendix REFERENCES

[40] R. Puche-Panadero, M. Pineda-Sanchez, M. Riera-Guasp, J. Roger-Folch,

E. Hurtado-Perez, and J. Perez-Cruz, “Improved resolution of the MCSA method
via Hilbert transform, enabling the diagnosis of rotor asymmetries at very low
slip,” IEEE Transactions on Energy Conversion, vol. 24, no. 1, pp. 52 – 59, 2009.

[41] K. Ohyama, G. Asher, and M. Sumner, “Comparative analysis of experimental

performance and stability of sensorless induction motor drives,” IEEE Transactions
on Industrial Electronics, vol. 53, no. 1, pp. 178 – 186, Feb. 2005.

[42] Manual for low Voltage Motors, ABB, 2004.

[43] S. Chaudhury and S. Gupta, “Online Identification Of AC Motor Misalignment

Using Current Signature Analysis and Modified K-Mean Clustering Technique,”
in Industrial Technology, 2006. ICIT 2006. IEEE International Conference on, Dec. 2006,
pp. 2331 –2336.

[44] M. E. Brumbach and J. A. Clade, Industrial Machines. Thomson Delmar Learning,

Inc., 2003, ch. 4.

[45] A. Bonnett and G. Soukup, “Cause and analysis of stator and rotor failures in
three-phase squirrel-cage induction motors,” IEEE Transactions on Industry Appli-
cations, vol. 28, no. 4, pp. 921 –937, Jul. 1992.

[46] G. Joksimovic and J. Penman, “The detection of inter-turn short circuits in the
stator windings of operating motors,” IEEE Transactions on Industrial Electronics,
vol. 47, no. 5, pp. 1078 –1084, Oct. 2000.

[47] S. Chen, “Induction motor broken rotor bar diagnostics,” ABB, Tech. Rep., Sep.
2008.

Page 87
Page 88
Appendix A

MATLAB Code

A.1 Data Analysis Algorithm for Broken Rotor Bar De-

tection (umc100current.m)

% ZCT spectral analysis of Uni Krakow new data for broken bar analysis
% with plot, for real data
% can choose 1 or 3 phase analysis. Use this one!!
% data: S1 001 to S1 020 (change last digit 1,2,3,...,20)
% S2 001 to S2 020 (change last digit 1,2,3,...,20)

clear all

%data=load('C:\test\Krakow data\S1\S1 043.dat');

data=csvread('04−11−09−50−20.csv');
downsample=10; % downsampling factor

N ph = 3; %number of phases for analysis

F rated = 50; %fundamental frequency

Fs=1/(data(2,1)−data(1,1)); %Sampling frequency
Fs=Fs/downsample;
disp(['Sampling frequency =' ' ' num2str(Fs) ' ' 'Hz'])

%resample properly
if downsample>1
filterorder=5; %vary between 2−15
[b,a] = butter(filterorder,1/downsample);
%[b,a] =cheby1(filterorder,0.5,1/downsample);
data=filter(b,a,data);

Page 89
A.1 Data Analysis Algorithm for Broken Rotor Bar Detection (umc100current.m)

end

data=data(1:downsample:end,:);

if N ph==3
phase=[1 2 3]; % for three phase ZCT analyiss
elseif N ph==1
phase = 1; %should be 1, 2 or 3
end
index=1:size(data,1);
sig=data(index,phase+1); %data(:,1) is samples, so phase+1 gives phase,
%choose Fs samples for having 1 second data

if N ph==3
[zcount,zindex,zspace]=zcrinterp3phase(sig,Fs,F rated); %count=zero−cross count,
%index and space intervals between contiguous zero−crossings
elseif N ph==1
[zcount,zindex,zspace]=zcrinterp(sig,Fs,F rated); %count=zero−cross count,
%index and space intervals between contiguous zero−crossings
end

S=fft(sig(:,1)); %FFT of current signal

Smag=abs(S); % Magnitude response of current signal
Smag=20*log10(Smag); %in dB scale
Smag(1)=0; %DC correction

fftPoints = 2ˆ10; %FFT of ZCTs, 2ˆ9 uses 512 points

%we use the last portion of the data
zspace = zspace(length(zspace) − fftPoints + 1 : length(zspace));
%zspace=zspace+randn(size(zspace))*50*1e−3; % with added noise
Z=fft(zspace,fftPoints);
Zmag=abs(Z); % Magnitude response of zero−cross intervals
Zmag=20*log10(Zmag); %in dB scale
Zmag(1)=0; %DC correction

sN=numel(Smag);
sF=[0:round(sN/2)]/sN*Fs; %Frequency range for current signal
zN=numel(Zmag);
zF=[0:round(zN/2)]/zN*2* N ph * F rated;

Page 90
Appendix A MATLAB Code

%Whole frequency range for ZCT spectrum is N ph *2*50 Hz

close all
subplot(221),plot(index,sig,'*−'),title(strcat('Current in phase:',...
num2str(phase),'; ',num2str(Fs),' Hz sampling freq')),...
ylabel('Current in A'),xlabel('Sample no.')
subplot(223),plot(zspace,'*−'),title('Interpolated zero−crossing intervals'),...
ylabel('Mean zero−cross interval (no. of samples)'),xlabel('Data blocks')
subplot(222),plot(sF,Smag(1:numel(sF)),'*−'),title(strcat('Magnitude plot of ...
current in phase:',num2str(phase),'; ',num2str(numel(S)),' samples')),...
ylabel('Magnitude response in dB'),xlabel('Frequency in Hz'),xlim([0 100])
subplot(224),plot(zF,Zmag(1:numel(zF)),'*−'),title(strcat('Magnitude plot...
of zero−cross interval in phase:',num2str(phase),'; ',num2str(zcount),...
' samples')), ylabel('Magnitude response in dB'),...
xlabel('Frequency in Hz'),xlim([0 zF(end)])

%Broken Bar analysis based on zero−cross

sf2 range=10; % slip range of interest (Hz)
n=round(length(Z)*sf2 range/(2* N ph *F rated)); % 10Hz is up to n data points
% here only take the first half to calculate threshold
%beware the length of Z is the actual point of fft. Shuold not use zspace
ZZ=Zmag(2:n); %check for max Zmag in the range 0.1−10 Hz
[max value,max index]=max(ZZ);
%threshold=mean(ZZ)+2.5*std(ZZ);
threshold=−67;

disp('[MAX peak in 0−10Hz range Fixed threshold]')

disp([num2str(ZZ(max index)) ' ' num2str(threshold)])

if ZZ(max index)>threshold
brb = 1;
disp('−−−−−−−− There are broken bars −−−−−−−−−−')
else
disp('############### NO broken rotor bars ################')
sf2 = 0;
brb = 0;
end

disp('The following estimations are valid only if the peak is significant,...

i.e. above the threshold.')

Page 91
A.2 Single-phase Zero Crossing Times Calculation (umc100zc.m)

sf2=max index*(N ph *2*F rated/zN); % twice slip frequency 2sf

% length(zspace) point FFT, N ph *2*F rated Hz sampling freq for zero crossing
slip=sf2/(2*F rated); % slip
rotor speed=60*F rated*(1−slip)/2; % rotor mechanic speed
disp(['slip =' ' ' num2str(slip) ' '])
disp(['twice slip frequency =' ' ' num2str(sf2) ' ' 'Hz'])
disp(['rotor mechanic speed =' ' ' num2str(rotor speed) ' ' 'rmp'])

result = [0,max value,0,threshold,0,sf2,0,brb];

A.2 Single-phase Zero Crossing Times Calculation (umc100zc.m)

% ZCT spectral analysis of Mikko motor setup data for broken bar analysis
% with plot, for real data
% can choose 1 or 3 phase analysis. Use this one!!
% data: S1 001 to S1 020 (change last digit 1,2,3,...,20)
% S2 001 to S2 020 (change last digit 1,2,3,...,20)

clear all
clc

N ph = 3; %number of phases for analysis

F rated = 50; %fundamental frequency

%data=dlmread('C:\test\Krakow data\S1\S1 059.log',';');

data=dlmread('2009−08−04 Bunch1 NoShort Load 20.txt',';');

zspace=[];

%process columnwise to retrive zct signal

signal=[];
for i=1:size(data,1)
for j=1:3
signal(end+1)=data(i,j*2); %get col 2, 4 6
end
end

%take difference
for i=1:numel(signal)−1

Page 92
Appendix A MATLAB Code

temp zspace=signal(i+1)−signal(i);
if temp zspace<0
temp zspace=temp zspace+65536;
end
zspace(end+1)=temp zspace;
end

zspace ref=5833.33;

fftPoints = 2ˆ10; %FFT of ZCTs, 2ˆ9 uses 512 points

% zspace = zspace(length(zspace) − fftPoints + 1 : length(zspace));
%
% zspace=(zspace−zspace ref)*(8e−6/7); % make zspace to time

figure(2)
plot(zspace,'*−')

A.3 Three-phase Zero Crossing Time Calculation (zcrin-

terp3phase.m)

function [count,index,space]=zcrinterp3phase(x,f s,f rated)

% Function zero−crossing count using interpolation
% [count,index,space]=zcrinterp(x,f s,f rated)
% Counts the zero crossings of x in count
% Index of the zero values in index
% Spacings of the zero−indeces in space
%
% Analyze the 3−phase ZCT

count=0;
if size(x,1)<size(x,2) % if rows > columns
x=x';
end

%If shifted signal adjust by mean centring

for clom=[1:3]
x(:,clom)=x(:,clom)−mean(x(:,clom));
end

Page 93
A.3 Three-phase Zero Crossing Time Calculation (zcrinterp3phase.m)

index=[]; % unit times at each ZC point

% /f foundamental is the real ZC times
id=size(x); % id = [rows columns]

for i=1:id(1)−1 % rows

for k=1:id(2) % columns
if x(i,k)==0
index(end+1)=i; % exact zero
count=count+1;
elseif (x(i,k)<=0 & x(i+1,k)>=0) | (x(i,k)>0 & x(i+1,k)<=0)
index(end+1)=i+x(i,k)/(x(i,k)−x(i+1,k));
count=count+1;
end
end
end

index=index/f s;
space=[];

if ˜isempty(index)
for i=1:numel(index)−1
%space(end+1)=index(i+1)−index(i); % calculate the ZCT
%space(end+1)=index(i+1)−index(i)−f s/(6*f rated);
% subtract the fundamental component
space(end+1)=index(i+1)−index(i)−1/(6*f rated);
% subtract the fundamental component
end
end

Page 94
Appendix B

Test Results of Stator

Short Faults before
Misalignment Adjust

Page 95
 Table B.1. Healthy motor

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.97 -29.58 1494.80 1497.93 0.35 0.14 3.40 69.39

10% 24.75 -29.33 1485.10 1484.72 0.99 1.02 3.59 73.27
20% 24.47 -27.20 1472.70 1468.25 1.82 2.12 4.11 83.88
30% 24.27 -28.62 1461.80 1456.21 2.55 2.92 4.87 99.39
40% 24.01 -31.70 1446.80 1440.51 3.55 3.97 5.79 118.16
50% 23.70 -35.55 1430.10 1422.04 4.66 5.20 6.88 140.41
60% 23.34 -38.71 1408.40 1400.21 6.11 6.65 8.10 165.31
70% 23.69 -42.02 1379.80 1421.41 8.01 5.24 9.53 194.49

Page 96
 Appendix B
Table B.2. 1,2 short directly

Test Results of Stator Short Faults before Misalignment Adjust

Measured Computed
% Slip % Slip Current %
Corrected Amplitude Speed speed
Load (from (from (A) Current
(Hz) (dB) (rpm) (rpm)
Level measured computed (measured /Nominal
(20-30) (20-30) (from drives (from rotor
speed) speed) by device) (4.9 A)
panel) peak)

noload 24.96 -27.87 1495.00 1497.85 0.33 0.14 3.40 69.39

10% 24.72 -25.29 1484.50 1483.01 1.03 1.13 3.63 74.08
20% 24.51 -25.69 1474.10 1470.47 1.73 1.97 4.14 84.49
30% 24.27 -28.28 1460.90 1455.94 2.61 2.94 4.88 99.59
40% 24.00 -32.25 1447.40 1440.14 3.51 3.99 5.81 118.57
50% 23.69 -35.28 1430.40 1421.63 4.64 5.22 6.88 140.41
60% 23.34 -37.85 1409.10 1400.14 6.06 6.66 8.10 165.31
70% 22.84 -40.17 1381.50 1370.54 7.90 8.63 9.52 194.29
Page 97
 Table B.3. 1,2 short with resistance

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)
Base-
25.00 -33.61 3.35 68.37
line
noload 24.97 -27.40 3.40 69.39
10% 24.71 -21.94 1484.10 1482.58 1.06 1.16 3.57 72.86
20% 24.49 -24.99 1472.70 1469.22 1.82 2.05 4.08 83.27
30% 24.24 -29.46 1460.20 1454.42 2.65 3.04 4.85 98.98
40% 23.96 -33.51 1444.50 1437.45 3.70 4.17 5.81 118.57
50% 23.64 -34.72 1425.60 1418.51 4.96 5.43 6.91 141.02
60% 23.22 -37.47 1401.50 1394.06 6.57 7.06 8.20 167.35
70% 22.64 -40.08 1365.70 1358.54 8.95 9.43 9.74 198.78

Page 98
 Appendix B
Table B.4. 1,3 short directly

Test Results of Stator Short Faults before Misalignment Adjust

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.97 -19.00 1494.40 1498.25 0.37 0.12 3.32 67.76

10% 24.72 -21.14 1484.90 1483.26 1.01 1.12 3.59 73.27
20% 24.51 -26.88 1474.30 1470.55 1.71 1.96 4.14 84.49
30% 24.26 -27.78 1461.50 1455.61 2.57 2.96 4.94 100.82
40% 23.99 -30.66 1446.70 1439.31 3.55 4.05 5.90 120.41
50% 23.66 -34.70 1428.30 1419.84 4.78 5.34 7.01 143.06
60% 23.25 -36.73 1403.80 1394.73 6.41 7.02 8.27 168.78
70% 22.70 -40.94 1370.70 1361.71 8.62 9.22 9.79 199.80
Page 99
 Table B.5. 1,3 short with resistance

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current Voltage
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal (V)
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)
Base-
24.97 -25.35 3.35 68.37 3.9
line
noload 24.95 -22.73 1495.10 1496.77 0.33 0.22 3.37 68.78 3.8
10% 24.72 -22.45 1485.70 1483.43 0.95 1.10 3.57 72.86 3.8
20% 24.51 -26.69 1474.70 1470.73 1.69 1.95 4.10 83.67 3.8
30% 24.27 -27.89 1461.70 1456.44 2.55 2.90 4.84 98.78 3.8
40% 24.01 -31.10 1447.50 1440.69 3.50 3.95 5.77 117.76 3.8
50% 23.69 -35.06 1430.50 1421.66 4.63 5.22 6.85 139.80 3.8
60% 23.32 -37.57 1408.70 1399.25 6.09 6.72 8.08 164.90 3.8
70% 22.77 -42.30 1375.30 1366.30 8.31 8.91 9.55 194.90 3.8

Page 100
 Appendix B
Table B.6. 2,3 short directly

Test Results of Stator Short Faults before Misalignment Adjust

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.92 -26.05 1495.20 1495.09 0.32 0.33 3.55 72.45

10% 24.72 -40.26 1485.30 1483.20 0.98 1.12 3.86 78.78
20% 24.51 -32.21 1474.10 1470.86 1.73 1.94 4.38 89.39
30% 24.27 -31.95 1461.70 1456.33 2.55 2.91 5.18 105.71
40% 24.00 -33.89 1447.70 1439.87 3.49 4.01 6.40 130.61
50% 23.68 -34.14 1430.70 1420.99 4.62 5.27 7.25 147.96
60% 23.30 -37.88 1406.30 1397.74 6.25 6.82 8.48 173.06
70% 22.77 -41.03 1378.20 1366.31 8.12 8.91 9.95 203.06
Page 101
 Table B.7. 2,3 short with resistance

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current Voltage
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal (V)
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)
Base-
24.97 -26.25 3.40 69.39 10.9
line
noload 24.94 -24.77 1495.30 1496.27 0.31 0.25 3.40 69.39 10.5
10% 24.72 -23.52 1485.80 1483.43 0.95 1.10 3.62 73.88 10.5
20% 24.52 -25.88 1475.10 1471.09 1.66 1.93 4.13 84.29 10.5
30% 24.28 -26.42 1461.70 1457.00 2.55 2.87 4.88 99.59 10.5
40% 24.02 -29.87 1447.70 1441.21 3.49 3.92 5.79 118.16 10.5
50% 23.72 -32.95 1431.50 1423.07 4.57 5.13 6.86 140.00 10.5
60% 23.35 -36.52 1409.50 1400.70 6.03 6.62 8.08 164.90 10.5
70% 22.85 -38.46 1381.60 1370.97 7.89 8.60 9.47 193.27 10.5

Page 102
 Appendix B
Table B.8. 1,2,3 short directly

Test Results of Stator Short Faults before Misalignment Adjust

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.96 -28.71 1495.30 1497.53 0.31 0.16 3.49 71.22

10% 24.72 -29.96 1485.70 1483.24 0.95 1.12 3.80 77.55
20% 24.90 -35.80 1474.90 1494.04 1.67 0.40 4.37 89.18
30% 24.28 -32.47 1462.30 1456.66 2.51 2.89 5.17 105.51
40% 23.99 -37.17 1447.60 1439.66 3.49 4.02 6.16 125.71
50% 23.67 -36.53 1428.20 1420.40 4.79 5.31 7.28 148.57
60% 23.25 -39.18 1403.70 1394.87 6.42 7.01 8.56 174.69
70% 22.65 -42.83 1368.10 1358.89 8.79 9.41 10.08 205.71
Page 103
 Table B.9. 1,2,3 short with resistance

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)
Base-
24.97 -18.39 3.40
line
noload 24.93 -19.75 1494.60 1495.52 0.36 0.30 3.38 68.98
10% 24.73 -33.08 1485.50 1483.75 0.97 1.08 3.61 73.67
20% 24.53 -26.41 1475.30 1471.69 1.65 1.89 4.12 84.08
30% 24.29 -27.20 1462.80 1457.46 2.48 2.84 4.86 99.18
40% 24.03 -29.99 1448.80 1441.90 3.41 3.87 5.80 118.37
50% 23.72 -32.76 1431.70 1423.49 4.55 5.10 6.88 140.41
60% 23.34 -35.73 1409.20 1400.21 6.05 6.65 8.11 165.51
70% 22.80 -39.35 1376.30 1368.10 8.25 8.79 9.55 194.90

Page 104
 Appendix B
Table B.10. B1-1,B2-1 short directly

Test Results of Stator Short Faults before Misalignment Adjust

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.93 -28.23 1495.20 1496.07 0.32 0.26 3.61 73.67

10% 24.73 -29.88 1485.80 1483.61 0.95 1.09 3.90 79.59
20% 24.52 -33.50 1474.20 1470.92 1.72 1.94 4.46 91.02
30% 24.28 -31.69 1462.10 1456.51 2.53 2.90 5.26 107.35
40% 24.01 -39.02 1447.70 1440.57 3.49 3.96 6.20 126.53
50% 23.69 -35.21 1428.50 1421.11 4.77 5.26 7.27 148.37
60% 23.28 -38.53 1405.30 1396.75 6.31 6.88 8.51 173.67
70% 22.76 -41.43 1369.80 1365.72 8.68 8.95 9.98 203.67
Page 105
 Table B.11. B1-1,B2-1 short with resistance

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)
Base-
24.97 -18.72 3.34
line
noload 24.94 -18.49 1494.70 1496.59 0.35 0.23 3.35 68.37
10% 24.73 -20.96 1485.20 1483.84 0.99 1.08 3.61 73.67
20% 24.53 -27.14 1475.30 1471.66 1.65 1.89 4.17 85.10
30% 24.30 -27.55 1463.40 1457.97 2.44 2.80 4.94 100.82
40% 24.04 -30.66 1449.30 1442.16 3.38 3.86 5.89 120.20
50% 23.73 -33.85 1432.20 1423.95 4.52 5.07 6.97 142.24
60% 23.35 -37.62 1410.20 1400.95 5.99 6.60 8.22 167.76
70% 22.77 -40.63 1380.20 1366.25 7.99 8.92 9.68 197.55

Page 106
 Appendix B
Table B.12. B1-2,B2-1 short directly

Test Results of Stator Short Faults before Misalignment Adjust

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.95 -27.37 1494.20 1497.10 0.39 0.19 3.68 75.10

10% 24.73 -29.26 1484.90 1483.71 1.01 1.09 3.99 81.43
20% 25.49 -35.31 1473.90 1529.47 1.74 -1.96 4.60 93.88
30% 25.01 -43.18 1460.70 1500.44 2.62 -0.03 5.41 110.41
40% 23.98 -38.63 1445.30 1438.86 3.65 4.08 6.38 130.20
50% 23.64 -40.82 1426.40 1418.26 4.91 5.45 7.50 153.06
60% 23.21 -41.02 1399.50 1392.50 6.70 7.17 8.81 179.80
70% 22.52 -43.91 1360.30 1350.97 9.31 9.94 10.40 212.24
Page 107
 Table B.13. B1-2,B2-1 short with resistance

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)
Base-
24.97 -18.46 3.34 68.16
line
noload 24.93 -20.40 1495.50 1495.55 0.30 0.30 3.37 68.78
10% 24.74 -20.54 1486.70 1484.17 0.89 1.06 3.64 74.29
20% 24.53 -26.27 1476.20 1471.99 1.59 1.87 4.21 85.92
30% 24.31 -27.06 1464.50 1458.43 2.37 2.77 4.98 101.63
40% 24.05 -30.02 1451.50 1443.05 3.23 3.80 5.95 121.43
50% 23.75 -32.61 1434.10 1425.07 4.39 5.00 7.05 143.88
60% 23.37 -36.15 1411.80 1402.37 5.88 6.51 8.31 169.59
70% 22.84 -39.42 1379.30 1370.50 8.05 8.63 9.81 200.20

Page 108
 Appendix B
Table B.14. B1-3,B2-1 short directly

Test Results of Stator Short Faults before Misalignment Adjust

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.97 -25.22 1495.10 1498.20 0.33 0.12 3.51 71.63

10% 24.72 -25.41 1485.50 1483.48 0.97 1.10 3.81 77.76
20% 24.51 -29.98 1475.10 1470.56 1.66 1.96 4.34 88.57
30% 24.27 -33.11 1461.20 1456.10 2.59 2.93 5.11 104.29
40% 24.00 -34.22 1446.70 1440.11 3.55 3.99 6.09 124.29
50% 23.69 -35.75 1429.50 1421.68 4.70 5.22 7.17 146.33
60% 23.27 -36.89 1405.70 1396.30 6.29 6.91 8.46 172.65
70% 22.73 -40.76 1370.70 1363.78 8.62 9.08 9.95 203.06
Page 109
Page 110
Appendix C

Test Results of Stator

Short Faults after
Misalignment Adjust

Page 111
 Table C.1. 1,2 short directly

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.97 -25.52 1495.30 1498.23 0.31 0.12 3.31 67.55

10% 24.73 -23.36 1485.30 1483.51 0.98 1.10 3.63 74.08
20% 24.51 -28.77 1474.70 1470.80 1.69 1.95 4.20 85.71
30% 24.27 -31.13 1461.70 1456.22 2.55 2.92 5.01 102.24
40% 24.00 -30.68 1446.80 1440.14 3.55 3.99 5.99 122.24
50% 23.67 -33.45 1428.40 1420.27 4.77 5.32 7.11 145.10
60% 23.26 -35.06 1405.10 1395.58 6.33 6.96 8.38 171.02
70% 22.73 -37.77 1369.40 1363.61 8.71 9.09 9.91 202.24

Page 112
 Appendix C
Table C.2. 1,2 short with resistance

Measured Computed
Rotor % Slip % Slip Current %

Test Results of Stator Short Faults after Misalignment Adjust

Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)
Base-
24.97 -33.35 3.29
line
noload 24.98 -32.54 1495.30 1498.56 0.31 0.10 3.29 67.14
10% 24.73 -23.21 1486.20 1483.97 0.92 1.07 3.52 71.84
20% 24.53 -25.21 1475.70 1471.91 1.62 1.87 4.06 82.86
30% 24.30 -26.37 1463.70 1458.05 2.42 2.80 4.83 98.57
40% 24.04 -29.33 1449.70 1442.69 3.35 3.82 5.80 118.37
50% 23.74 -31.69 1432.50 1424.36 4.50 5.04 6.88 140.41
60% 23.36 -35.16 1410.90 1401.45 5.94 6.57 8.15 166.33
70% 22.83 -36.51 1377.80 1370.01 8.15 8.67 9.65 196.94
Page 113
 Table C.3. 1,3 short directly

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.97 -29.54 1495.50 1498.10 0.30 0.13 3.26 66.53

10% 24.73 -24.14 1486.10 1483.83 0.93 1.08 3.55 72.45
20% 24.52 -26.57 1475.50 1471.38 1.63 1.91 4.11 83.88
30% 24.29 -27.84 1463.30 1457.28 2.45 2.85 4.92 100.41
40% 24.03 -30.20 1449.30 1441.51 3.38 3.90 5.87 119.80
50% 23.71 -32.75 1430.60 1422.46 4.63 5.17 7.01 143.06
60% 23.32 -35.18 1407.80 1399.09 6.15 6.73 8.29 169.18
70% 22.76 -38.13 1372.60 1365.70 8.49 8.95 9.83 200.61

Page 114
 Appendix C
Table C.4. 1,3 short with resistance

Measured Computed
Rotor % Slip % Slip Current %

Test Results of Stator Short Faults after Misalignment Adjust

Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)
Base-
24.97 -27.33 3.41
line
noload 24.97 -27.22 1495.60 1498.28 0.29 0.11 3.42 69.80
10% 24.72 -25.00 1485.90 1483.46 0.94 1.10 3.53 72.04
20% 24.51 -25.86 1474.80 1470.76 1.68 1.95 4.03 82.24
30% 24.27 -27.83 1462.50 1456.15 2.50 2.92 4.79 97.76
40% 24.00 -30.87 1447.60 1440.10 3.49 3.99 5.76 117.55
50% 23.68 -33.19 1429.30 1420.87 4.71 5.28 6.88 140.41
60% 23.29 -35.26 1406.40 1397.15 6.24 6.86 8.14 166.12
70% 22.72 -38.04 1371.20 1362.95 8.59 9.14 9.65 196.94
Page 115
 Table C.5. 2,3 short directly

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.96 -29.83 1494.80 1497.74 0.35 0.15 3.43 70.00

10% 24.72 -32.58 1485.30 1482.99 0.98 1.13 3.74 76.33
20% 25.50 -34.54 1473.70 1529.91 1.75 -1.99 4.34 88.57
30% 24.26 -33.89 1461.70 1455.52 2.55 2.97 5.14 104.90
40% 23.98 -34.00 1445.90 1438.75 3.61 4.08 6.11 124.69
50% 23.66 -36.37 1426.80 1419.37 4.88 5.38 7.24 147.76
60% 23.25 -35.70 1403.90 1394.85 6.41 7.01 8.51 173.67
70% 22.71 -37.42 1370.40 1362.38 8.64 9.17 10.04 204.90

Page 116
 Appendix C
Table C.6. 2,3 short with resistance

Measured Computed
Rotor % Slip % Slip Current %

Test Results of Stator Short Faults after Misalignment Adjust

Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)
Base-
24.97 -26.75 3.23
line
noload 24.97 -26.60 1495.60 1498.28 0.29 0.11 3.23 65.92
10% 24.72 -24.47 1485.30 1483.04 0.98 1.13 3.48 71.02
20% 24.50 -25.65 1474.70 1470.19 1.69 1.99 4.04 82.45
30% 24.26 -28.10 1462.20 1455.66 2.52 2.96 4.81 98.16
40% 23.99 -30.53 1447.30 1439.66 3.51 4.02 5.78 117.96
50% 23.67 -34.05 1429.40 1420.04 4.71 5.33 6.88 140.41
60% 23.28 -35.68 1405.20 1396.64 6.32 6.89 8.16 166.53
70% 22.75 -37.96 1372.40 1364.96 8.51 9.00 9.63 196.53
Page 117
 Table C.7. B1-1,B2-1 short directly

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.94 -37.06 1494.80 1496.43 0.35 0.24 3.57 72.86

10% 24.71 -32.60 1484.70 1482.59 1.02 1.16 3.87 78.98
20% 24.49 -33.90 1473.20 1469.32 1.79 2.05 4.42 90.20
30% 24.24 -32.63 1460.70 1454.39 2.62 3.04 5.19 105.92
40% 23.97 -34.38 1444.80 1437.94 3.68 4.14 6.14 125.31
50% 23.63 -35.99 1426.10 1418.00 4.93 5.47 7.23 147.55
60% 23.22 -36.82 1402.40 1393.39 6.51 7.11 8.48 173.06
70% 22.68 -38.59 1368.50 1360.66 8.77 9.29 9.95 203.06

Page 118
 Appendix C
Table C.8. B1-1,B2-1 short with resistance

Measured Computed
Rotor % Slip % Slip Current %

Test Results of Stator Short Faults after Misalignment Adjust

Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)
Base-
24.97 -27.81 3.30
line
noload 24.97 -27.36 1494.50 1498.24 0.37 0.12 3.32 67.76
10% 24.72 -24.15 1484.70 1483.48 1.02 1.10 3.57 72.86
20% 24.51 -26.16 1479.60 1470.80 1.36 1.95 4.12 84.08
30% 24.27 -28.24 1462.30 1456.30 2.51 2.91 4.89 99.80
40% 24.01 -30.33 1447.50 1440.47 3.50 3.97 5.85 119.39
50% 23.69 -32.93 1430.40 1421.60 4.64 5.23 6.95 141.84
60% 23.31 -35.04 1406.30 1398.73 6.25 6.75 8.21 167.55
70% 22.80 -38.19 1376.50 1368.11 8.23 8.79 9.66 197.14
Page 119
 Table C.9. B1-2,B2-1 short directly

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.99 -35.12 1494.70 1499.14 0.35 0.06 3.66 74.69

10% 24.94 -33.63 1485.10 1496.57 0.99 0.23 4.01 81.84
20% 25.51 -34.77 1472.80 1530.43 1.81 -2.03 4.60 93.88
30% 25.75 -37.34 1460.20 1545.10 2.65 -3.01 5.36 109.39
40% 23.96 -37.95 1443.80 1437.34 3.75 4.18 6.32 128.98
50% 23.62 -36.21 1424.60 1417.07 5.03 5.53 7.42 151.43
60% 23.19 -38.68 1399.60 1391.45 6.69 7.24 8.71 177.76
70% 22.62 -38.78 1364.20 1357.06 9.05 9.53 10.22 208.57

Page 120
 Appendix C
Table C.10. B1-2,B2-1 short with resistance

Measured Computed
Rotor % Slip % Slip Current %

Test Results of Stator Short Faults after Misalignment Adjust

Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)
Base-
24.97 -27.32 3.33
line
noload 24.97 -26.85 1495.10 1498.27 0.33 0.12 3.38 68.98
10% 24.72 -24.42 1484.30 1483.24 1.05 1.12 3.65 74.49
20% 24.51 -26.13 1473.40 1470.52 1.77 1.97 4.19 85.51
30% 24.27 -28.18 1461.30 1455.95 2.58 2.94 4.95 101.02
40% 24.00 -30.23 1446.40 1439.83 3.57 4.01 5.90 120.41
50% 23.68 -33.22 1429.20 1420.78 4.72 5.28 6.98 142.45
60% 23.30 -34.89 1406.80 1398.28 6.21 6.78 8.24 168.16
70% 22.79 -38.04 1375.20 1367.53 8.32 8.83 9.71 198.16
Page 121
 Table C.11. 1,2,3 short directly

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.95 -34.92 1495.20 1497.11 0.32 0.19 3.43 70.00

10% 24.73 -36.99 1486.30 1483.76 0.91 1.08 3.76 76.73
20% 24.83 -34.11 1475.30 1490.08 1.65 0.66 4.36 88.98
30% 24.29 -31.93 1463.40 1457.36 2.44 2.84 5.19 105.92
40% 24.02 -33.14 1449.20 1441.44 3.39 3.90 6.12 124.90
50% 23.70 -32.77 1430.70 1422.04 4.62 5.20 7.24 147.76
60% 23.29 -34.76 1406.40 1397.52 6.24 6.83 8.52 173.88
70% 22.76 -37.98 1373.50 1365.53 8.43 8.96 10.04 204.90

Page 122
 Appendix C
Table C.12. 1,2,3 short with resistance

Measured Computed
Rotor % Slip % Slip Current %

Test Results of Stator Short Faults after Misalignment Adjust

Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)
Base-
24.97 -29.61 3.26
line
noload 24.97 -29.01 1494.70 1498.33 0.35 0.11 3.27 66.73
10% 24.73 -23.28 1485.70 1483.90 0.95 1.07 3.51 71.63
20% 24.52 -25.56 1475.10 1471.49 1.66 1.90 4.03 82.24
30% 24.29 -26.46 1463.50 1457.54 2.43 2.83 4.82 98.37
40% 24.02 -30.29 1448.70 1441.21 3.42 3.92 5.78 117.96
50% 23.71 -32.07 1430.40 1422.60 4.64 5.16 6.88 140.41
60% 23.31 -35.23 1408.30 1398.81 6.11 6.75 8.16 166.53
70% 22.79 -37.71 1376.10 1367.59 8.26 8.83 9.69 197.76
Page 123
 Table C.13. On line test with coupling gap

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.97 -26.04 1494.90 1497.91 0.34 0.14 3.21 65.51

10% 24.73 -23.75 1485.20 1483.64 0.99 1.09 3.47 70.82
20% 24.52 -22.30 1474.90 1471.48 1.67 1.90 4.04 82.45
30% 24.29 -25.70 1463.10 1457.42 2.46 2.84 4.82 98.37
40% 24.04 -29.18 1449.20 1442.10 3.39 3.86 5.80 118.37
50% 23.73 -31.92 1432.40 1423.77 4.51 5.08 6.91 141.02
60% 23.36 -35.33 1410.30 1401.31 5.98 6.58 8.18 166.94
70% 22.86 -39.20 1380.50 1371.59 7.97 8.56 9.65 196.94

Page 124
 Appendix C
Table C.14. On line test without coupling gap

Test Results of Stator Short Faults after Misalignment Adjust

Measured Computed
Rotor % Slip % Slip Current %
Amplitude Speed speed
Load Frequency (from (from (A) Current
(dB) (rpm) (rpm)
Level (Hz) measured computed (measured /Nominal
(20-30) (from drives (from rotor
(20-30) speed) speed) by device) (4.9 A)
panel) peak)

noload 24.96 -25.69 1494.30 1497.53 0.38 0.16 3.22 65.71

10% 24.73 -28.46 1485.60 1483.92 0.96 1.07 3.48 71.02
20% 24.53 -26.72 1475.70 1471.69 1.62 1.89 4.04 82.45
30% 24.30 -26.54 1464.10 1457.91 2.39 2.81 4.82 98.37
40% 24.04 -29.69 1449.70 1442.34 3.35 3.84 5.77 117.76
50% 23.73 -32.45 1432.20 1424.01 4.52 5.07 6.89 140.61
60% 23.35 -35.95 1410.50 1401.08 5.97 6.59 8.13 165.92
70% 22.83 -37.43 1377.50 1369.94 8.17 8.67 9.63 196.53
Page 125
Page 126
Appendix D

Induction motor
parameters

D.1 Data of the Motor Type SZJKe 14a

The following tables show the parameters of SZJKe 14a.

Table D.1. Ratings of motor SZJKe 14a

Parameter Value
Active power PN [kW] 0.8
Nominal voltage UN [V] 380
Nominal current [A] 2.2
Nominal power factor cosN [-] 0.74
Rotor speed [rpm] 1400
No load speed [rpm] 1497
Winding connection Y
Number of poles per phase winding p[-] 2
Nominal frequency [Hz] 50
Number of rotor bars [-] 22
Number of stator slots [-] 24
Rotor inertia [kg ∗ m2 ] 0.0025
Coil number per phase 4
Turn number of coil 90

Page 127
D.2 Data of the Motor Type Sg112M-4

Table D.2. Equivalent circuit parameters

Parameter Value
Stator winding resistance Rs [Ohms] 8.4
Referred rotor winding resistance Rr[Ohm] 8.2
Stator winding reactance Xs [Ohms] 10.3
Referred rotor winding reactance Xr [Ohms] 10.3
Magnetizing reactance [Ohms] 137.5
Core loss component RFe [Ohms] 2938

Table D.3. Data of bearing SKF type 6304 ZZ CXSQ

Parameter Value
Numbers of balls Nb [-] 7
Ball diameter Bd [mm] 6
Bearing pitch diameter Pd [mm] 32
Ball contact angle a [rad] 0

D.2 Data of the Motor Type Sg112M-4

The following tables show the parameters of Sg112M-4.

Page 128
Appendix D Induction motor parameters

Table D.4. Ratings of motor Sg112M-4

Parameter Value
Active power PN [kW] 1.33
Nominal voltage UN [V] 380
Nominal current [A] 2.867
Nominal power factor cosN [-] 0.84
Rotor speed [rpm] 1445
No load speed [rpm] 1492
Winding connection Y
Number of poles per phase winding p[-] 2
Nominal frequency [Hz] 50
Number of rotor bars [-] 28
Number of stator slots [-] 24
Rotor inertia [kg ∗ m2 ] 0.0197

Table D.5. Data of bearing 6306-2Z

Parameter Value
Numbers of balls Nb [-] 8
Ball diameter Bd [mm] 12
Bearing pitch diameter Pd [mm] 52
Ball contact angle a [rad] 0

Table D.6. Equivalent circuit parameters

Parameter Value
Stator winding resistance Rs [Ohms] 5.6
Referred rotor winding resistance Rr[Ohm] 2.24
Stator winding reactance Xs [Ohms] 6.45
Referred rotor winding reactance Xr [Ohms] 6.45
Magnetizing reactance [Ohms] 238
Core loss component RFe [Ohms] 2252

Page 129
Page 130
Appendix E

Test Results of Induction

Motor Broken Rotor Bar
Fault

Experiment Data Provided by the Cracow University of Technology

Page 131
 Table E.1. Symmetrical Motor I

Measure- Peak Peak Peak Peak Differ- Average

Load
Loading value value value value ence with rotor
ment file Slip 2sf current
(%) (Hz) (dB) (Hz) (dB) 50 Hz speed
name (A)
(Current) (Current) Device Device (dB) (rpm)

S1 001.dat 0.01 66.82 0.93 9.38 -69.43 9.38 -66.79 1486.00 1.47
S1 002.dat 0.03 81.36 3.33 3.22 -67.75 3.22 -62.57 34.62 1450.00 1.79
S1 003.dat 0.04 90.91 4.33 4.39 -66.55 4.25 -58.82 37.45 1435.00 2.00
S1 004.dat 0.05 101.36 5.40 5.57 -62.69 5.42 -54.00 44.13 1419.00 2.23
S1 005.dat 0.07 112.73 6.60 6.45 -60.42 6.59 -51.69 48.77 1401.00 2.48

Page 132
 Appendix E
Table E.2. 1 broken rotor bar in Motor I

Test Results of Induction Motor Broken Rotor Bar Fault

Measure- Peak Peak Peak Peak Differ- Average
Load
Loading value value value value ence with rotor
ment file Slip 2sf current
(%) (Hz) (dB) (Hz) (dB) 50 Hz speed
name (A)
(Current) (Current) Device Device (dB) (rpm)

S1 006.dat 0.01 68.64 0.93 9.38 -76.70 9.96 -49.44 1486.00 1.51
S1 007.dat 0.03 82.27 3.27 3.22 -56.42 6.45 -49.91 37.05 1451.00 1.81
S1 008.dat 0.05 95.45 4.80 4.69 -49.87 4.69 -44.48 29.81 1428.00 2.10
S1 009.dat 0.06 105.00 5.80 5.86 -47.61 5.86 -40.69 35.54 1413.00 2.31
S1 010.dat 0.07 115.91 7.07 7.03 -39.32 6.88 -40.78 32.46 1394.00 2.55
Page 133
 Table E.3. 2 broken rotor bars in Motor I

Measure- Peak Peak Peak Peak Differ- Average

Load
Loading value value value value ence with rotor
ment file Slip 2sf current
(%) (Hz) (dB) (Hz) (dB) 50 Hz speed
name (A)
(Current) (Current) Device Device (dB) (rpm)

S1 011.dat 0.01 69.55 1.07 9.38 -62.40 10.69 -58.95 1484.00 1.53
S1 012.dat 0.03 85.00 3.47 3.52 -53.23 3.37 -44.44 28.59 1448.00 1.87
S1 013.dat 0.05 95.00 4.60 4.39 -48.69 4.54 -38.40 29.17 1431.00 2.09
S1 014.dat 0.06 105.00 5.67 5.57 -40.94 5.57 -36.43 28.80 1415.00 2.31
S1 015.dat 0.07 115.91 6.80 6.74 -38.01 6.74 -33.24 36.07 1398.00 2.55

Page 134
 Appendix E
Table E.4. 3 broken rotor bars in Motor I

Test Results of Induction Motor Broken Rotor Bar Fault

Measure- Peak Peak Peak Peak Differ- Average
Load
Loading value value value value ence with rotor
ment file Slip 2sf current
(%) (Hz) (dB) (Hz) (dB) 50 Hz speed
name (A)
(Current) (Current) Device Device (dB) (rpm)

S1 016.dat 0.02 68.64 1.53 0.29 -71.68 10.55 -58.01 1477.00 1.51
S1 017.dat 0.04 82.27 3.60 3.52 -54.53 3.52 -50.33 35.53 1446.00 1.81
S1 018.dat 0.05 95.00 5.07 4.98 -44.55 4.98 -41.96 37.71 1424.00 2.09
S1 019.dat 0.06 103.18 6.00 5.86 -46.70 6.01 -40.87 40.91 1410.00 2.27
S1 020.dat 0.07 114.09 7.27 7.32 -45.50 7.18 -36.70 36.28 1391.00 2.51
Page 135
 Table E.5. Symmetrical Motor II

Measure- Peak Peak Peak Peak Differ- Average

Load
Loading value value value value ence with rotor
ment file Slip 2sf current
(%) (Hz) (dB) (Hz) (dB) 50 Hz speed
name (A)
(Current) (Current) Device Device (dB) (rpm)

S1 001.dat 0.01 51.36 0.50 0.29 -75.10 4.25 -22.47 1492.44 1.13
S1 002.dat 0.02 92.27 2.15 8.79 -72.31 10.69 -49.49 1467.69 2.03
S1 003.dat 0.04 145.45 4.04 8.20 -63.68 8.06 -56.61 1439.34 3.20
S1 004.dat 0.05 184.55 5.49 5.57 -65.71 10.84 -57.97 40.77 1417.59 4.06
S1 005.dat 0.08 240.91 7.79 7.91 -62.68 7.76 -58.88 45.20 1383.14 5.30

Page 136
 Appendix E
Table E.6. 1 broken rotor bar in Motor II

Test Results of Induction Motor Broken Rotor Bar Fault

Measure- Peak Peak Peak Peak Differ- Average
Load
Loading value value value value ence with rotor
ment file Slip 2sf current
(%) (Hz) (dB) (Hz) (dB) 50 Hz speed
name (A)
(Current) (Current) Device Device (dB) (rpm)

S2 006.dat 0.00 38.72 0.48 8.50 -71.70 9.96 -58.38 1492.74 1.11
S2 007.dat 0.02 71.85 2.31 2.34 -55.52 2.34 -49.28 1465.34 2.06
S2 008.dat 0.04 110.57 4.04 4.10 -50.03 4.10 -44.15 34.05 1439.34 3.17
S2 009.dat 0.06 142.31 5.64 5.57 -48.04 5.57 -44.44 35.00 1415.44 4.08
S2 010.dat 0.08 184.86 8.04 7.91 -47.13 8.06 -37.02 31.41 1379.34 5.30
Page 137
 Table E.7. 2 broken rotor bars in Motor II

Measure- Peak Peak Peak Peak Differ- Average

Load
Loading value value value value ence with rotor
ment file Slip 2sf current
(%) (Hz) (dB) (Hz) (dB) 50 Hz speed
name (A)
(Current) (Current) Device Device (dB) (rpm)

S2 011.dat 0.00 39.07 0.39 8.50 -68.43 7.76 -60.42 1494.14 1.12
S2 012.dat 0.02 70.46 2.15 2.05 -58.49 2.20 -54.47 1467.74 2.02
S2 013.dat 0.04 109.52 4.01 3.81 -54.37 3.96 -45.30 29.17 1439.89 3.14
S2 014.dat 0.06 147.19 5.80 5.86 -48.54 6.01 -22.44 34.66 1412.99 4.22
S2 015.dat 0.08 184.51 7.82 7.91 -45.82 7.76 -39.58 28.77 1382.64 5.29

Page 138
 Appendix E
Table E.8. 3 broken rotor bars in Motor II

Test Results of Induction Motor Broken Rotor Bar Fault

Measure- Peak Peak Peak Peak Differ- Average
Load
Loading value value value value ence with rotor
ment file Slip 2sf current
(%) (Hz) (dB) (Hz) (dB) 50 Hz speed
name (A)
(Current) (Current) Device Device (dB) (rpm)

S2 016.dat 0.01 38.72 0.51 9.67 -69.75 6.88 -66.81 1492.39 1.11
S2 017.dat 0.02 69.76 2.25 2.34 -54.94 2.20 -48.94 1466.29 2.00
S2 018.dat 0.04 109.52 4.03 4.10 -46.87 4.10 -42.10 29.75 1439.54 3.14
S2 019.dat 0.06 148.59 5.93 5.86 -44.13 6.01 -40.53 33.59 1410.99 4.26
S2 020.dat 0.07 178.93 7.42 7.32 -44.84 7.47 -34.72 26.82 1388.74 5.13
Page 139