Makalah Seminar Omicron 24-25 Juli 2018

Seminar for Diagnostic Measurements
and On-line Monitoring on Rotating Machines
July 24 - 25, 2018

Jakarta, Indonesia
Invitation
After the seminar, the participants will have a clear understanding of the common
electrical defects in rotating machines including the methods for detecting them reliably
and in time.
Rotating machines, such as motors and generators, are highly important components
in power generation and industrial applications. Machine reliability and availability are
therefore in high demand. Premature failure may lead to significant economic losses, due
to unexpected outages and possible damage to the asset itself and the surroundings. To
plan maintenance effectively, it is essential to have accurate condition information. This
will enable operators to take effective maintenance decisions, when required.
The seminar gives an overview about state of the art testing methods that are
characterized through high sensitivity and capability for detecting various faults. Also
their limitations will be discussed critically. Additionally, new trends and developments
are also examined. All statements are supported by scientific investigations, practical
measurements and case studies.
Organisation Information
Date: July 24 - 25, 2018
Venue: Hotel Ciputra Jakarta , Jalan Letnan Jenderal S. Parman, Jakarta 11470
Event Organizer
OMICRON electronics Asia Limited Hotel Ciputra Jakarta
Unit 2812-19, 28/F, The Metropolis Tower, Jalan Letnan Jenderal S. Parman,
10 Metropolis Drive, Hung Hom, Jakarta 11470, Indonesia
Kowloon, Hong Kong Phone: (62-21) 566 0640
Phone: +852 3767 5500 Fax: (62-21) 566 1172
Fax: +852 3767 5400
E-mail: events.asia@omicronenergy.com
Major Speakers
Fabian Öttl received his MSc degree in Electrical

Engineering in the Graz University of Technology in 2011.
With his specialisation and extensive field experiences in
electrical machines, Fabian is now working as the Product
Manager for rotating machines testing and diagnostics at
OMICRON headquarters in Klaus, Austria. Prior to joining
OMICRON, Fabian worked as the high voltage insulation
specialist in the Research and Development Department
of a manufacturer of large generators. Fabian is active
in different working groups of CIGRE and IEC. He has
written and presented numerous scientific and technical
papers at various conferences.
Seokhoon Hong is a Regional Application Specialist on

Partial Discharge for OMICRON in the Asia-Pacific Region.
He has been working at OMICRON Korea and OMICRON
Asia limited since 2011 as an Application Engineer with
special focus on testing and diagnostic solutions for high
voltage assets. He studied and received his bachelor’s
and master’s degree in electrical engineering at MyungJi
and HanYang universities, respectively. Before joining
OMICRON, he had worked as an Electrical Engineer in
the power plant engineering department.
Agenda
Day 1 - July 24, 2018 (Tuesday)
Time Topics
Introduction in electrical testing of rotating machines

09:00 - 09:45 • Common electrical tests
• Insulation buildup
High voltage AC-Tests

• Dissipation factor (tanδ) and voltage withstand test;
09:45 - 10:30 measurement principle
• Why should the test be performed?
• What can be detected
10:30 - 11:00 Coffee break
Benefits of a portable high voltage source

11:00 - 11:45 • How to establish the parallel resonance circuit
• Practical measurements and use cases
11:45 - 12:30 Customer Case Study
12:30 - 14:00 Lunch
Unconventional measurements on rotating machines

14:00 - 15:00 • Sweep Frequency Response Analysis
• Dielectric Response Analysis
Partial discharge (PD) measurement

15:30 - 16:15 • What is partial discharge
• How to perform partial discharge measurements

Day 2 - July 25, 2018 (Wednesday)
Time Topics
Advanced source separation methods

09:00 - 09:45 • 3 Phase Amplitude Ratio Diagram (3PARD)
• 3 Center Frequency Ratio Diagram (3FREQ)
Partial discharge measurements on rotating machines

• Why should PD measurements be conducted?
09:45 - 10:30
• What parameters are influencing the measurement
• Use cases
12:30 - 14:00 Lunch
Propagation of partial discharges in rotating machines

14:00 - 15:00 • Investigation on different stators
• Comparison between measurement approaches
PD Monitoring system for rotating machines

15:30 - 16:15 • Temporary monitoring vs. continuous
• Installation examples and key characteristics
Benefits from continuous PD Monitoring

16:15 - 17:00 • Examples of failure detection with online PD Monitoring
• Innovative source separation and interpretation tools
Content Page
Day 1
Morning
01 Introduction in electrical testing of rotating machines

p.2 - 26
Afternoon
02 Unconventional measurements on rotating machines
p.26 - 46
Day 2 Morning
Morning
03 Advanced source separation method

p.47 - 62
Afternoon
04 Propagation of partial discharges in rotating machines windings

Benefits from continuous monitoring of rotating machines p.62 - 83
Disclaimer
Copyright
The material made available is intended for the customers personal use only. OMICRON (or its licensor)
reserves all rights to the documents and all parts thereof, including the translation, the reprint and the
reproduction. Therefore, third parties may not – for whatever purpose – reproduce or replicate,
or process or distribute with the help of electronic systems any part of the provided documents without the
author‘s written permission. This obligation continues to apply after the end of the meeting.
Request
If you identify any information in the meeting documentation which might be incorrect or misleading,
please advice suggested amendments to OMICRON
(https://www.omicronenergy.com/en/contact/addresses/singapore/).
Terms and Conditions
Attendance at this meeting is subject to the General Sales Conditions of OMICRON including the Special Terms
for Training courses and Meetings. These are available online on
https://www.omicronenergy.com/en/legal/terms/ or upon request.
Generator principle
Electromagnetic induction principle (linear movement)

> A potential difference across a conductor is generated by exposure to a
varying magnetic field
> In this example by moving the magnet a voltage is induced in the winding
Seminar for Diagnostic Measurements and

On-line Monitoring on Rotating Machines
Introduction in Electrical Testing of Rotating Machines
OMICRON Electronics Asia Ltd. July , 2018

© OMICRON Page 2
Generator principle Generator principle
Electromagnetic induction principle (rotational movement) Winding arrangement

> Rotating the magnets is the simplest technical solution > In motors/generators 3 equal windings are
arranged in an angle of 120°
> All components rotating with the magnet are called ROTOR
> Phase voltages are shifted of 120° to each
> All stationary parts (e.g. winding, housing,...) are part of the STATOR
other
© OMICRON Page 3 © OMICRON Page 4
© OMICRON electronics Asia Limited 2018 Page 1

Construction of stator winding Roebel bar
Source: Wikipedia
Source: Partzsch Elektromotoren
Roebel bar Roebel bar

Air
> Skin effect at large diameter conductors
Primary current I magnetic field H Increased current at
outer part of conductor
Decreased primary Iron

current through center
> Unequal magnetic resistance in upper (3) and lower segments (1) of slots
Increased current at > Inhomogeneous flux distribution in the slots
flowing self-induction current outer part of conductor
Source: radartutorial.eu, Visualization of skin effect Isolated strands of Roebel bar > Circular currents in conductors with concentration at slot base (1)
> To equal current density Æ twist of single strands
> AC current produce a magnetic field along the wire cross-section Side view
> Self-induced currents increase at outside of conductor

Æ Manufacture wire of single isolated strands to avoid skin effect
Top view
Source: Wikipedia

Insulation Build up Inner potential grading (IPG)
> Stator winding strand Facts

> Inner potential grading consists of
conductive tape (fleece)
Insulated Strands
> Applied over Roebel bars
Filling material > Harmonizing the electric field at
edge of the Roebel bars causes
Inner Potential Grading lower E-field concentration With IPG Æ E ~ 0,8 x Emax
(IPG)
Main Insulation
Outer Corona Protection,

End Potential Grading
(OCP, EPG)
Reduction of dielectric stress Illustration off electric field distribution

distribut
uttio
io
ion
on
n at stator winding
g strand
sttrra
rand
a with
and without IPG
Inner potential grading (IPG) Main insulation

Cracks or delamination
Inner potential > Delamination: due to > Layers of mica tape on glass carrier
grading (IPG)
different thermal expansion > Muscovite and Phlogopite with
Voids between insulation and excellent electrical behavior (high
copper strands specific resistance, high resistance
> IPG can prevent PD against PD,...)
Nominal insulation thickness
caused by delamination > Resin fills the gaps between the
Insulated
Copper > Cracks/ Delamination: due mica-layers
Strands Effective insulation thickness to T.E.A.M.
> Voids: manufacture and Specific Resistance
Material
Main Insulation T.E.A.M. in Ω mm²/m
Copper 1,7 x 10²
Outer Corona Protection (OCP) Iron 0,1
Porcelain 1018
Delamination Paper 1015 – 1018
Mica 1015 – 1018

Source: MICAMATION

Main insulation Main insulation
> Winding insulation with mica layers Arrangement of Mica layers zig-zag
straight
> Epoxy resin has medium electrical

properties
Epoxy resin
> Propagation of electrical trees only
through epoxy resin
> Mica “flakes” act like a barrier for
Conductor strand
electrical treeing
Mica flakes
Theoretical discharge > Simulated treeing behavior under

channel
electrical, thermal and mechanical
stress
Outer corona Source: R.Vogelsang; Detection of electrical tree propagation by
protection (OCP) partial discharge measurements; ETEP; 2005
Source: R. Vogelsang, Electrical breakdown in HV Winding Insulations of

Different Manufacturing Qualities; IEEE, June 2006
Resin rich systems Vacuum Pressure Impregnation (VPI)
Resin rich technique Facts of VPI

> Pre-assembled wet mica tapes > Impregnation of single winding bars
(Wrapped with 40-50 % warm (single bar) and full stator windings
hardened resin) (Global VPI) under vacuum
> Insulated winding bars are placed in > Assist penetration of liquids into space
heated form press and shaped into (voids)
dimension of stator slot
> Better insulation system than obtained Global VPI autoclave, Source: GMR electric
> Through high pressure unused with conventional methods

epoxy resin drips out at the ends of
> Better heat dissipation
winding bars
> Longer insulation life time
> Voids will be displaced due to
the high pressure > Improvement against insulation
degradation (moisture, chemical)
Source: R. Vogelsang, Electrical breakdown in
HV Winding Insulations...; IEEE, June 2006
Source: R. Vogelsang, Electrical breakdown in HV Winding
Insulations of Different Manufacturing Qualities; IEEE, June 2006

Outer corona protection (OCP) Outer corona protection (OCP)
Arrangement > Without OCP

> Partial discharge in the spaces between insulation surface and laminated core
> Outer corona protection is a semi-
conductive tape > With OCP
> Harmonization of electric field and thereby elimination of the PD between
> Applied after the main insulation insulation and laminated core
> Controlled conduction of the current from EPG
> Layer with defined potential
> Defined resistivity of the bar surface
(ground potential – stator core)
Conductor
Insulation
End potential grading (EPG)

EPG
Laminated core Outer corona protection
(OCP) - conductive
OCP
Source: Voith
End Potential Grading (EPG) End Potential Grading (EPG)

U
Arrangement & facts Arrangement & facts UN
> Semi-conductive coating consisting > Even if a insulation is applied at the

of paint or tape end-winding area, it is on high 0
OCP EPG
voltage potential
> Applied on bars and coils where
they exit stator slot > Without EPG Æ high streamer OA
discharge at end-winding area

> Overlap with OCP and extend end-
winding area > With EPG Æ smooth high-voltage OA… Overlapping Area
potential grading at the end-winding
area
> Main component silicon carbide
(SiC) which shows voltage
dependent conductivity
Different probes with and without EPG, Source: J. Weidner,

Illustration of electric field concentration at end winding area without and with EPG ETG-Fachtagung, Grenzflächen in Isoliersystem, 2008

Insulation Build-Up Stator core lamination
> Summary Stacking of insulated iron sheets

> Insulated layers of iron sheets to reduce hysteresis and eddy current losses
> Sheets of layers stacked on top of each other
> Roebel bar with insulated strands

> Inner Potential Grading (IPG)
> Main insulation
> Outer Corona Protection (OCP)
> End Potential Grading (EPG)
Source: Partzsch Elektromotoren Source: VG Power Turbo
Know about the condition of your rotating machine Extension of expected lifetime...
Prevention is Better than Cure > ... through periodic testing and/or monitoring
Negative influences
> Mechanical impacts
e.g. vibration of stator windings
> Aging
e.g. degradation of stator insulation
ion
> Protection problems
e.g. under function of protection device
Testing and corrective measures

es
> Maintenance
> Replace parts
> Processing of insulation

Definitions of tests Temporary vs. Continuous monitoring
> OFFLINE tests: performed at de-energized and disconnected rotating > Temporary monitoring
machine from the network > Short time monitoring of parameters for minutes, hours, days or load cycles
> Insulation resistance & winding resistance > Pre-installation of setup components (e.g. coupling capacitors for PD)
> Capacitance and Loss-factor measurement (C & Tan delta) > Non-permanent installations of measurement setup
> Dielectric response measurement (FDS & PDC) > Reduced amount of measurement data
> Partial discharge measurement (temporary)
> And many more... > Continuous monitoring
> Permanent monitoring with pre-installation of measurement setup
> ONLINE tests: performed “live” and providing information about condition of > Permanent analyzation of parameters (dissipation factor, partial discharge,...)
the machine under operating conditions > Definition system configuration, setting thresholds levels for alarming
> Temperature monitoring > Huge amount of measurement data
> Vibration and analysis
> Partial discharge monitoring (continuous)
> And many more...
Test Overview Test overview
Electrical tests on rotating machines Machine part tested What should be checked Recommended electrical tests
Partial discharge (PD) ■ ■ ■
Contamination ■ ■ ■ ■ ■
> Stator Winding Insulation degradation ■ ■ ■ ■ ■
> DC – Resistance Stator winding Voltage withstand ■
> Insulation resistance Insulation integrity ■ ■ ■
> Rotor Winding > Dissipation factor Connection problems ■

Inter turn faults ■
> Pole impedance > Voltage withstand
Inter turn faults ■ ■
measurement > Partial discharge Rotor winding
High contact resistance ■
> SFRA > ...
n
tio
> Winding resistance
rp
so
or
ab
> Contact resistance
ct
fa
ric
er
ct
nt
t
> ...
is
en
w
le
ys
en
ts )/ po
ts ie
m
em
al
is
en d d
re
em
g
an
ys
ur
su
g
in
m nδ
ur
al
as
ri n
st
re a
se
ea
an
as
te
e
su ex
ito
re ta
in
on
m
)m
e
en
m
su r (
st
nd
se
ea nd
on
sp
e
te
ea t o
nc
on
ta
m
m ni
re
(P
m fac
op
nc
s
ta
sp
tio tio
PD
th
cy
e
dr
sis
ta
re
i
n
ra riza
> Stator Core rg
en
is
io
ne
le
re
ha
c
ge
es
at
qu
la
Po
tri
l i
ng
sip
sc
n-
tr
lta
po
ec
fre
O
di
> Flux test
di
ac
Vo
is
el
e,
in
p
,d
nt
al
Di
ee
nc
w
rt i
Co
ce
Sw
ta
DC
Pa
> Electromag. imperfection

an
sis
cit
re
pa
n
tio
Ca
la
su
In

Thank you

fabian.oettl@omicronenergy.com On-line Monitoring on Rotating Machines
High Voltage AC Tests on Rotating Machines

© OMICRON Page 29
Introduction
Agenda
> Why conducting electrical tests on Motors and Generators?
> Introduction
> Capacitance – Power/Dissipation Factor Measurement
Defective corona Overvoltages;
– Measurement principle protection; 3% 2%
Thermal cycling or
> Voltage Withstand Test Thermal
damage
overloading; 7%
17%
– Measurement principle Loosening of
bars in slot; 10% Ageing; 31%
> What can be identified? Insulation
Mechanical
damage
damage
> Use case 105 MW Hydro Generator 56%
24%
Internal partial
discharge; 22%
Contamination
of winding;
Bearing 25%
damage
3%
Source: Brütsch et al. "Insulation Failure Mechanisms

of Power Generators", DEIS July/August 2008
> Electrical root causes are significant when it comes to a failure of the equipment

Introduction Groundwall Insulation Tests
> Why conducting electrical tests on Motors and Generators? > Insulation Resistance (DC) Groundwall Insulation
50% > Voltage withstand test (AC)

45%
40%
35% > Power/Dissipation Factor (AC)
30% Epoxy resin
25%
20% > Partial Discharge Test
15%
10%
5%
Laminated core
Conductor strand
0%
(High Voltage)
(Ground)
Mica tape
Motors from Petro Motors in General
Source: Thorsen, Dalva. "A Survey of Faults on Induction Motors in Offshore Oil Industry, Petrochemical Industry,
Gas Terminals, and Oil Refineries", IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 31, NO. 5, 1995
> Electrical root causes are significant when it comes to a failure of the equipment
Outer corona protection
(OCP) on Ground
Power/Dissipation Factor (PF/DF) Measurement

Agenda
> The electric representation of the
> Introduction groundwall insulation consists of:
> An ideal capacity for the vacuum field Groundwall Insulation
> Capacitance – Power/Dissipation Factor Measurement > Polarization
– Measurement principle > An electric conduction through the
dielectric and surface
> Voltage Withstand Test Epoxy resin
– Measurement principle
> What can be identified?
Laminated core
Conductor strand
(High Voltage)
> Use case 105 MW Hydro Generator
(Ground)
Mica tape
Outer corona
protection (OCP) on
Ground

Power/Dissipation Factor Measurement How is the tangent(δ) measured?
> The insulation of a rotating > The dissipation/loss factor is measured by comparing a test object to a
electrical Machine can be known reference
modelled as:
> The reference should be an ideal capacitor
> A capacitor
> A resistor > Typical a compressed gas capacitor is used
> Nearly no polarization losses in gas (tan(δ)≤1*10-5)
> Simply speaking: The higher the
> Good stray current suppression using guard electrodes
resistive current the bigger the
dielectric losses
High voltage source

CTest
CRef
> tan =
RTest
> cos(φ) = test

reference object
IRef ITest
Reactive power compensation Dissipation Factor (DF) Measurement
> Simplified diagram of compensation with reactor > One value for the entire insulation of the winding (similar to insulation
resistance)
U2 U1 > Often used as FAT for coils and bars, were guarding techniques can be
V1
applied
V2
> With limits according to IEC 60034-27-3:2015
Lcomp W2 W1
> On entire windings strongly influenced by the surface current on the end
potential grading
> No specific limits for entire windings
> Dissipation factor tip-up can be an indirect indication of PD activity

Factory acceptance test Tangent delta measurement
> Checking the curing state of the single bars or coils > Tangent delta for rotating machines is a fingerprint measurement
> Guarding techniques are used
end potential grading (EPG)
> Guard Ring electrodes with slot coating gap
> Driven Guard i
> Guard electrodes on endwinding
Source: IEC 60034-27-3
Source: H.L. Lee, “The Handbook of Dielectric

Analysis and Cure Monitoring”, 2014
Dissipation Factor (DF) Measurement Dissipation Factor (DF) Measurement
> On entire windings the DF is strongly influenced by the surface currents
> Modified Coil

> 6kV
> ~ 0.5 m slot portion length
> Guarding of surface current
is possible at one side
Source: IEC 60034-27-3:2015)

Dissipation Factor (DF) Measurement Use I case fingerprint measurement
> No specific limits for dissipation factor measurement > 1st measurement in 2007, 2nd measurement 2017
> Examples of three different measurements on three different machines
> Phase comparison and/or trending are the most powerful tools to check for
0,90
0,85
Tan Delta [%] potential faults.
0,80
0,75
0,70
% 0,65
0,60
0,55
0,50
0,45 U-VWE
0,40
0
1000
2000
3000
4000
5000
6000
7000
V 8000
Voltage Withstand Test

Agenda
> One of the key issues in the diagnosis is: "Can
> Introduction the insulation withstand the voltage during
operation and under non-stationary
> Capacitance – Power/Dissipation Factor Measurement
conditions?"
> Voltage Withstand Test > This question can be answered by means of a
– Measurement principle voltage withstand test with " Yes " or "No".
> What can be identified? > On new machines, this test is carried out
> Use case 105 MW Hydro Generator within the framework of an acceptance test.
> In older machines, the test ensures that the
aged insulation is still reliable. The test voltage
for older machines is lower than for new
machines.

Voltage Withstand Test Voltage withstand test
> The voltage is increased via a ramp function in certain steps to the desired
test voltage. Then, the test voltage is maintained for one minute. > If the voltage is increased in steps,
the steps should be lower than 5%
> According to IEC 60034-1, the
of the test voltage.
test shall be commenced at a
voltage not exceeding half of > Test voltage levels for new and
the full test voltage. aged machines are defined in
V
IEC 60034-1.
> The duration of the ramp V
Center Conductor
should not be less than > The test is passed if there was
t > 1 min ttotal > 10 s
10 seconds. no flashover or breakdown.
Test voltage
Test voltage
< 5 % of the
test voltage
Initial voltage
< 50 % test voltage Initial voltage
t < 50 % test voltage
t
Ramp Stable voltage
Ramp Stable voltage
Voltage withstand test

Agenda
> The test levels can be challenging for old machines
> Introduction
> Standard claims that test with DC is possible by adding a certain safety
factor Æ Use Case > Capacitance – Power/Dissipation Factor Measurement
> Voltage distribution is not the same than for AC – Measurement principle
> Common practice in the world of the rotating machines are test voltages in > Voltage Withstand Test
the range of: – Measurement principle
> Phase to ground rated voltage Un/√3 > What can be identified?
Center Conductor
> A factor (e.g. 1,2) times phase to ground rated voltage 1,2 x Un/√3
> Nominal voltage Un > Use case 105 MW Hydro Generator
> Combined dissipation factor and voltage withstand test

> If the machine has no breakdown during the PF/DF measurement and
PF/DF measurement results are ok, test is passed

What can be measured What can be measured with HV-AC-Tests
Conductor strand
Insulation integrity: (High Voltage)
> Voltage withstand test (AC) Groundwall Insulation
> Power/Dissipation Factor (AC) > Path existing, but not yet conductive
enough
Epoxy resin > Detectable with Insulation resistance
and PF/DF measurement
Laminated core
> Use Case I (Ground)
Laminated core
Conductor strand
(High Voltage)
(Ground)
Creepage path in the insulation: Conductor strand
(High Voltage)
> Breakdown not fully formed
Mica tape
> Only detectable with HV-AC tests
such PF/DF or voltage withstand
> Use Case II
Laminated core
(Ground)
Outer corona protection
© OMICRON Page 24
(OCP) on Ground © OMICRON Page 25
Test overview
Agenda
Machine part tested What should be checked Recommended electrical tests
Partial discharge (PD) ■ ■ ■
Contamination ■ ■ ■ ■ ■ > Introduction
Insulation degradation ■ ■ ■ ■ ■
Stator winding Voltage withstand ■ > Capacitance – Power/Dissipation Factor Measurement
Insulation integrity ■ ■ ■
Connection problems ■
Inter turn faults ■ > Voltage Withstand Test
Inter turn faults ■ ■
Rotor winding
High contact resistance ■
> What can be identified?
n
tio
rp
> Use case 105 MW Hydro Generator

so
or
ab
ct
fa
ric
er
ct
nt
t
is
en
w
le
ys
en
ts )/ po
ts ie
m
em
al
is
en d d
re
em
g
an
ys
ur
su
g
in
m nδ
ur
al
as
ri n
st
re a
se
ea
an
as
te
e
su ex
ito
re ta
in
on
m
)m
e
en
m
su r (
st
nd
se
ea nd
on
sp
e
te
ea t o
nc
on
ta
m
m ni
re
(P
m fac
op
nc
s
ta
sp
tio tio
PD
th
cy
e
dr
sis
ta
re
i
n
ra riza
rg
en
is
io
ne
le
re
ha
c
ge
es
at
qu
la
Po
tri
l i
ng
sip
sc
n-
tr
lta
po
ec
fre
O
di
di
ac
Vo
is
el
e,
in
p
,d
nt
al
Di
ee
nc
w
rt i
Co
ce
Sw
ta
DC
Pa
an
sis
cit
re
pa
n
tio
Ca
la
su
In

Use Case II: Hydro Generator (105 MVA) Use Case II: Hydro Generator (105 MVA)
> Decommissioned Hydro Generator > Before the lifetime test, following tests where performed:
> Very high PD values > Insulation resistance
> 45 years in operation > DC Ramp test
> 160.000 operating hours > PF/DF measurement
> Peak load generator (11.000 starts and stops) > PD probe measurement
> 12,5 kV
> Playground for several investigations > No indication of breakdown or week spot after AC-voltage endurance test
with 1,2 Un and 20,5 hours, repeating the mentioned tests.
> Beside the Power/Dissipation Factor measurement 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Following common insulation tests where performed:

600 350 250 150 100 150 200 250 300 200 150 180 200 250
900 800 700 650 450 450 300 200 250 200 100 200 300 300
400 350 300 200 150 100 80 60 100 140 180 180 200 150
200 180 150 120 110 100 80 130 140 150 150 120 140 150
200 180 120 120 100 90 100 120 150 40 50 120 130 150
> PD measurement
150 120 100 100 100 100 180 100 100 80 80 80 80 80
70 70 90 100 100 100 100 100 90 50 50 50 50 50
80 80 80 80 80 80 80 50 50 50 50 50 50 50
150 150 120 80 70 60 60 80 100 110 110 130 150 130
> Insulation resistance DC (5 kV)

80 80 80 80 80 80 70 80 70 50 50 50 50 50
150 140 130 100 90 90 80 50 60 60 70 70 70 70
80 80 80 80 80 80 80 50 50 30 40 50 50 50
80 80 80 80 70 80 80 80 80 80 50 30 30 30
100 80 80 60 50 50 50 50 30 30 30 30 50 50
> DC-Ramp test DC (up to 32 kV)

150 100 100 80 80 70 60 50 50 40 50 50 50 60
100 90 80 50 50 50 50 50 50 50 40 40 50 50
60 60 60 50 50 50 70 70 80 70 50 50 50 50
50 50 50 50 50 50 50 50 50 70 60 60 60 60
50 50 40 40 50 40 40 40 30 40 40 50 50 50
> PD probe measurement

50 50 50 50 50 50 50 40 40 40 40 40 40 40
60 50 50 60 50 50 50 50 50 50 50 50 50 50
60 80 90 90 80 90 80 60 60 60 60 50 50 50
70 60 60 60 50 50 50 50 50 50 50 50 50 50
60 60 50 60 50 50 50 50 50 50 50 50 50 50
70 60 60 60 60 60 50 60 50 50 50 50 50 60
60 70 70 60 70 60 60 60 60 60 60 60 50 50
> AC-Voltage endurance test until breakdown

110 100 90 90 100 90 100 80 80 80 80 60 80 70
120 150 120 110 100 90 80 60 70 60 60 50 50 50
60 50 50 50 50 50 40 40 40 40 40 40 40 40
50 40 40 40 30 30 30 30 30 30 30 30 30 30
40 50 40 30 30 30 30 30 30 30 30 30 40 40
40 30 30 30 30 30 30 30 30 30 60 70 80 90
50 40 30 30 30 40 30 40 40 40 30 30 40 40
60 30 30 30 30 40 40 40 40 40 40 40 40 40
40 40 50 40 30 30 30 30 30 30 30 30 30 30
50 40 30 30 30 30 30 30 30 30 30 30 40 40
30 30 30 30 30 30 30 30 30 30 30 30 30 30
30 30 30 40 30 30 30 30 30 30 30 30 30 30
80 80 80 80 80 80 70 70 80 80 80 80 80 80
100 90 80 70 70 60 60 70 60 60 70 70 60 60
80 80 70 60 60 70 60 50 60 60 70 100 110 100
80 80 70 80 70 70 70 70 70 70 70 70 70 70
180 140 140 100 100 90 80 80 70 70 100 120 120 130
180 150 120 110 90 80 70 70 70 70 100 120 120 110
90 90 80 80 70 70 70 70 70 70 70 80 100 120
100 70 80 80 70 70 90 90 80 70 70 90 80 100
90 80 80 80 80 80 70 60 60 60 60 70 100 100
110 90 90 80 80 70 70 60 60 50 50 60 100 100
100 80 80 80 80 80 80 70 50 50 100 100 100 100
90 80 80 80 70 70 70 70 50 50 50 100 100 100
90 80 70 70 60 70 60 60 50 50 50 50 100 100
110 90 80 80 60 60 50 50 50 50 90 110 110 120
90 80 80 70 70 60 50 60 50 60 90 110 110 100
70 70 60 60 50 50 50 60 50 60 90 100 90 90
60 60 50 50 60 60 70 70 60 70 70 70 70 70
90 80 70 70 60 50 50 50 60 70 80 80 100 100
70 70 70 70 80 90 90 90 80 90 80 80 80 70
120 100 100 100 100 90 90 90 80 70 60 70 70 60
100 100 90 100 100 100 90 90 80 80 80 70 70 60
Use Case II: Hydro Generator (105 MVA) Breakdown during: Testing vs. Grid
> Power/Dissipation Factor Measurement and combined > Advantages of breakdown during testing vs. in operation:
Voltage Withstand Test > Much less energy in the breakdown point Æ small damage
> Increase of the test voltage > Offline testing often during bigger maintenance
> Maintenance team and spear parts are available Æ no additional standstill
> Sudden Breakdown of a week spot at 1,5 x Un (18,75)
> Prior stress with 1,2 x Un for several hours > Pictures from Breakdown during operation
> Then all the prior tests were repeated Æ without results
Source: Maughan, “Theroset Stator Bar Insulation Systems”,

DEIS, Amrch/April – Vol. 34, No.2
Source: Combined cycle journal, Maughan, August 2013

High Voltage AC Tests Thank you
> High voltage AC Tests are beneficial to ensure a reliable operation until
the next service interval
> In combination with PD measurement they are enabling a reliable
condition assessment of the insulation
> PD testing and High voltage AC tests are complementing each other
> There are defects not detectable with PD measurements
> There are defects not detectable with PF/DF or voltage withstand test (next
slides)
> High Voltage AC tests can uncover week spots which are not detectable
with DC tests
> Different voltage distributions
> Single Value for the entire winding or winding section

> No information if the matter is concentrated fabian.oettl@omicronenergy.com
> No information about type of defect
Agenda
> Pole Impedance measurement
> DC Resistance measurements
> High Voltage AC Tests

> Power/Dissipation Factor measurement
> Voltage Withstand test
> High Voltage source for PD measurement
> Use Case

Benefits of a portable high voltage source

© OMICRON Page 2

Pole Impedance Measurement Pole Impedance Measurement
> Detection of inter turn faults on Rotor Poles > Connection of the CPC 100
> Comparison between poles
> A defective pole shows significant lower impedance > 2 Options for source
> 130 V AC
> Difference between measurements with pulled rotor and > 500 V AC
rotor inside the
e machine
Pole Impedance Measurement Agenda
> Pole Impedance measurement
> DC Resistance measurements
Old Machine
> Voltage Withstand test
> Use Case

> No international standard
> No specific limit Source: Zikulnig; „Erdschluss- und Windungsschlusserkennung an Läuferpolwicklungen
mit der Frequenzganganalyse (FRA)“ Omicron Diagnosewoche 2015
> ± 2,5% seems to be a common
practice

DC – Resistance Measurement DC – Resistance Measurement
> Performed on: > Different outputs for different resistance ranges:
> Stator winding to detect possible soldering > 1 μΩ to 10 mΩ Æ 100A DC
problems > 10 mΩ to 10 Ω Æ 6A DC
> Rotor winding to detect bad contacts
> As FAT to check the I²R component and after heat
run
> Due to the small resistance a 4 wire measurement

has to be establishes =
> Evaluation
> Comparison between the different phases (if
starpoint can be disconnected)
> Comparison between different connectors
> Comparison between other measurement data
(trending)
RContact RContact
RContact
V
V V
DC – Resistance Measurement DC – Resistance Measurement
> Connection example Winding resistance measurement with 50 A

> 5,6 MVA Hydro Generator
Significant higher resistance value in phase W
0,05Ω
0,045Ω
0,04Ω
0,035Ω
0,03Ω
U
0,025Ω V
W
0,02Ω
0,015Ω
0,01Ω
0,005Ω
0,0Ω

DC – Resistance Measurement Damage due to bad soldering contact
> Injection of 250 A for several hours

> Thermographic detection of the fault
> Root cause: Bad soldering contact
Source: Binder & Zikulnig, “Übersicht über die häufigsten Fehler an Hydrogeneratoren der VHP”, Omicron Diagnostic Week, 2015
Damage due to bad soldering contact Agenda

Root Cause: > Pole Impedance measurement
> Poor realization of the soldering contact
> Mechanical stress and high temperature leads to fatigue cracks > DC Resistance measurements
> Continuous deterioration of the soldering point until breakdown

High temperature Ö > Voltage Withstand test
Contact-Burn / Meltdown Ö
Arcing Ö > Use Case
potentially machine burn
> State of the art in “older” machines

> Brazing of coils and bars in new machines

Combined C and PF measurement Facts and Figures
> Light-weight
> Less than 29 kg/64 lbs
> Easy to transport
> Easy storage
> Multifunctional system

> Time-saving
> Easy handling
> Automated reporting and test procedures
> Tests can be prepared in the office
Combined C and PF/DF measurement Combined C and PF measurement
> CP TD1 can drive 300mA at 12kV

> For Capacitance & power/dissipation > Capacity of rotating machines is usually much higher than 80nF
factor (PF/DF) measurement
> Compensation has to be performed
> output voltage 12kV in conjunction with
CPC 100 (booster) 1
=
> Gas-isolated Reference capacitor 2∗ ∗ ∗
> weight CP TD1 = 25 kg / 55.2 lbs
> mounted with CPC 100 on trolley > Compensation calculation is performed by the device
CP CR500
CP CR500
CPC 100 CP TD1
(tan delta)
CP CR500

CP CR 500 - Facts Winding insulation capacitance
> Up to 12 kV (maximum voltage of the TD1) > Capacitance of stator to ground

> Used Frequency 15 Hz – 400 Hz
> Current compensation up to 1 A
> WxHxD = 455 x 275 x 220 mm
> Weight: 36 kg
> Available as
> 2 x 40 H
> 40 / 80 H
> 2 x 80 H
> Even though the values are discrete, EVERY capacitance within the specs
can be tested at rated frequency
Reactive power compensation Compensation 50 Hz

> For testing at rated frequency
> Simplified diagram of compensation with reactor
U2 U1 U2 U1 CPC + TD1 + 40 II 40 H
V2 V1 V2 V1
W2 W1 Lcomp W2 W1 CPC + TD1 + 80 II 40 H
CPC + TD1 + 40 H
CPC + TD1 + 80 H
CPC + TD1
0 100 200 300 400 500 600

Capacitance of the DUT in nF

How to connect the Devices? How to connect the Devices?
U V W U V W
HV Input 1 HV Input 2 HV HV Input 1 HV Input 2 IN-B HV

IN-A Booster IN-A Booster
CP CR500 CP CR500
Serial Serial
Safety B Safety A CP TD1 Safety B Safety A CP TD1
LV Output LV Output
CU
To CPC 100 To CPC 100
Terminate last CR500 in the safety loop Terminate last CR500 in the safety loop
with CPC 100 dongle with CPC 100 dongle
CU-V
High-voltage area Safe area High-voltage area Safe area
How to connect the Devices? Test Object
7 8
U V W
1 1
HV Input 1 HV Input 2 HV Input 1 HV Input 2 HV
IN-A Booster
CP CR500 CP CR500
Serial
Safety B Safety A Safety B Safety A CP TD1
LV Output 5 LV Output
1
4
3 2
To CPC 100
Terminate last CR500 in the safety loop
with CPC 100 dongle Serial Booster
4
Safety
CPC 100
1
High-voltage area Safe area

Excursion to PTM Examples of successful practical measurements
Type of Rated Max. Test Rated Capacitance PF/DF of the
machine Voltage Voltage Power of test test object
object
in % at
in kV in kV in MVA in nF
Utest,max
Motor 10 12 14 118,7 1,84
Hydro 10,8 12 43 176,0 2,53
Turbo 18 12 438 503,6 2,91
Hydro 9 9,9 35 1289,7 2,69
Hydro 10,5 12 56 607,1 4,06
Turbo 6,3 7,2 29,63 290,3 1,24
Turbo 10,5 12 90,75 419,7 1,01
Motor 6,3 72 13,05 113,2 1,36
Hydro 16,8 11,8 130 920,7 2,24
At the above listed machines the CPC + TD1 + CR500 was measuring the
power/dissipation factor, as voltage withstand test (except one), as well as it
was used as high voltage source for PD testing
Agenda Use Case
> Pole Impedance measurement > Hydro Generator

> 6,4 kV
> 1981 Class F
> DC Resistance measurements > Rotor was pulled for a bigger maintenance
> Electrical diagnosis was requested:
> High Voltage AC Tests > Insulation resistance
> Power/Dissipation Factor measurement > Dissipation Factor Measurement
> Voltage Withstand test > Voltage withstand test
> Partial discharge Measurement
> Use Case
© OMICRON Page 30

Use Case Use Case
> Insulation Resistance > Cleaning of the endwinding area

measurement
> PDC measurement
> @ 600s:
> Phase U: 5,8646 μA
> Phase V: 4,8075 nA
> Phase W: 4,8549 nA
© OMICRON Page 32
Use Case Use Case 7,00

Tan Delta [%]
6,00
> Insulation Resistance measurement
> PDC measurement > Dissipation Factor Measurement 5,00
> Similar results for V and W 4,00

%
> @ 600s: > Phase U significant higher 3,00
> Phase U Before: 5,8646 μA 2,00
> Phase U After: 513,067 nA 1,00 U-VWE
0,00
> Still a hundred times higher than the
1000
2000
3000
4000
5000
6000
other two phases! V
7,00
U-VWE Tan Delta [%]
3,50
6,00 V-UWE Tan Delta [%]
3,00
W-UVE
Before 5,00
2,50
4,00
% 2,00
3,00 %
After 2,00
1,50
1,00
1,00
0,50
V-UWE
0,00 0,00
0
1000
2000
3000
4000
5000
6000
1000
2000
3000
4000
5000
6000
V V
© OMICRON Page 35

Use Case Use Case
> Voltage withstand test > Fault localization

> Up to 1,2 UN
> Breakdown at UN
Detection of smoke
Breakdown at
6400V
Lower section 0.01A
Upper section 2.04A
© OMICRON Page 37
Use Case Use Case

> Reparation > Verification of the successful reparation
> Separation of the defective winding segment
U-VWE before repair
Separation of defective winding bar Re-soldering with healthy lower bar
all phases after repair

Isolation of soldered parts Completed repair

Thank you

Unconventional Measurements on Rotating Machines

© OMICRON Page 40
Unconventional Tests on Rotating Machines Agenda
> What are unconventional Tests on Rotating Machines? > Dielectric response analysis
> DIRANA Measurement Measurement Principle
> Dielectric response measurements Influence of different parameter
> Diagnosis of ground wall insulation
> SFRA Measurements > Sweep Frequency Response Analysis

> Various possibilities for fault detection such as: Measurement Principle
> Inter turn short in the stator winding
> Inter turn short in the rotor winding How to connect?
> Broken bars in the damper cage Assessment methods
> ...
> Research topics

> No dedicated standards regarding these topics

Dielectric Spectroscopy Dielectric response analysis
I (f)
Dielectric Spectroscopy: Dielectric spectroscopy U, I U(t) I(t)
IR(f) IC (f)
measurement of tan(G) at > Dielectric spectroscopy =
U f R C different frequencies measurement of tan(G) at different
frequencies
t
> Frequency Domain Spectroscopy
10
(FDS) in frequency domain
Im IR (f)
IR(f) 90°- G
> Polarization/Depolarization Current
tan(G,f)
IC (f) IC(f) 1 (PDC) measurement method in time
= 1
I (f) domain
Dissipation factor
tan(G)
0,1
G(f)
0,01
U
0,0001 0,01 1 100 0,001
Re Frequency in Hz 0,001 1000
Frequency in Hz
Frequency Domain Spectroscopy (FDS) Frequency Domain Spectroscopy (FDS)
Frequency Domain Spectroscopy (FDS) Advantage of FDS:

> Dielectric spectroscopy = measurement > Fast and accurate at high frequencies
of tan(G) at different frequencies U ~ Insulation
under test > Resistant to disturbances
Frequency Duration of 1 sine wave
> Frequency Domain Spectroscopy (FDS)
in frequency domain 5000 Hz 0,2 ms
A Disadvantage of FDS: 1000 Hz 1 ms
I
> Very slow at low frequencies 50 Hz 20 ms
Principle
1 Hz 1s
> Apply sinusoidal voltage of different U, I U(t) I(t)
frequencies f1, f2, ... 0.1 Hz 10 s
10 mHz 100 s
> Determine tan delta (tanG) at the
frequencies f1, f2, 1 mHz 16,7 min
t
0.1 mHz 2,7 h
10 μHz 27 h
90°- G

Polarization/Depolarization Current (PDC) Factors affecting the measurement (1)
Polarization/Depolarization Current (PDC) > Temperature

> In Metals: T↑ Æ R↑
> Dielectric spectroscopy = measurement of > In insulation: T↑ Æ R↓
tan(G) at different frequencies U Insulation
under test
> Polarization/Depolarization Current (PDC)
measurement method in time domain
A
I > Moisture Source: IEEE 43-2013
> Only effecting old asphaltic or shellac insulation system

Principle Polarization Depolarization
> Moisture can affect the surface condition
> DC step voltage applied do winding terminal Æ Perform test at temperatures above the dew point
U, I U(t)
and ground
> Measure the charge current at times t1, t2, ... I(t)
t1 t2 t3 t4 t5 t6
t i
> Calculate the tan(G) at the corresponding > Surface condition
frequencies f1 = 1/t1 f2 = 1/t2 PDC > Contamination in the endwinding area
rea
using Fourier transformation measurement > Insulation design i
Factors affecting the measurement (3) Influence of Temperature
> Pre-polarisation (existing charge)
R
Phase
10 min.
U 40,2 GΩ
V 38,3 GΩ
W 46,0 GΩ
> Spare bars, resin rich technology, rated voltage of 10.5kV
Sumereder et. al: OBSERVATION OF DIELECTRIC PARAMETERS

AT GENERATOR STATOR WINDINGS, ISH 2009

Temperature and Relative Humidity of the Air Influence of insulation systems
> Spare bars, resin rich technology, rated voltage of 10.5kV
Sumereder et. al: OBSERVATION OF DIELECTRIC PARAMETERS

AT GENERATOR STATOR WINDINGS, ISH 2009
Dielectric response (dissipation factor) on a Fault Polarization Index PI
> As result from PDC measurement

> No effect of voltage magnitude on healthy windings
> Changes in resistivity over voltage magnitude indicates a possible weakness
Æ Measurement at different voltages
Voltage Magnitude: 80V Voltage Magnitude: 5000V

Measurement Time
Resistance PI Resistance PI
1 min 2690 MΩ 2720 MΩ

U–E 2,85 2,89
10 min 7670 MΩ 7860 MΩ
1 min 2680 MΩ 2750 MΩ

V–E 2,83 2,86
10 min 7580 MΩ 7870 MΩ
1 min 2790 MΩ 2930 MΩ

W–E 3,13 2,92
10 min 8750 MΩ 8550 MΩ

Agenda How does SFRA work?
> Dielectric response analysis How SFRA unit analyzes frequency response
Measurement Principle > Injection of sinusoidal excitation voltage with continuously increasing
Influence of different parameter frequency into one end of the machine winding
> This can be the rotor or multi-turn stator windings
> Sweep Frequency Response Analysis
> Measurement of signal returning from the other end
Measurement Principle
How to connect?
Assessment methods Sine generator, Machine Results
variable frequency (complex network)
How does SFRA work? Measurement setup stator
How SFRA unit analyzes frequency response

> Comparison of signals generates unique
frequency response which can be
compared to reference data A B C
CMC RMC
> Deviations indicate electrical changes within

the rotor or stator
Coaxial cables Rotating machine stator (complex RLC network)
> No additional data processing
required due to direct measurement
in the frequency domain Amplitude
Output Reference Measurement

Channel Channel
50Ω
Rref=50Ω Rme=50Ω
U1 U2 Injection signal
U Reference signal
Measurement signal
Results Phase SFRA instrument

Measurement setup rotor Measurement setup single pole
CMC RMC CMC RMC
Rotating machine
rotor (complex
Coaxial cables RLC network) Coaxial cables
Rotating machine pole
(complex RLC
network)
Output Reference Measurement Output Reference Measurement

Channel Channel Channel Channel
50Ω 50Ω
Rref=50Ω Rme=50Ω Rref=50Ω Rme=50Ω
U1 U2 Injection signal U1 U2 Injection signal
U Reference signal U Reference signal
Measurement signal Measurement signal
SFRA instrument SFRA instrument
Measurement between Stator and Rotor Methods of SFRA measurements
General
> Multiple connection possibilities:
> Stator: All phases separated
> Stator: Winding connected in star
> Stator: Winding connected in delta
> Rotor: Entire winding measurement
> Rotor: Single pole measurement
> Measurement between Rotor and Stator
> Depending on the connection, the measurement frequency response is

influenced and has different sensitivity to specific failure modes
Phases separated
Star connection
Delta connection

Methods of SFRA measurements Methods of SFRA measurements
General 1. Stator winding

> Measurement configuration > Single phase measurement: frequency response is obtained for each
> Always use the same measurement connection individual winding
> Always use the same measurement direction U1 V1 W1
> Documentation in the software
> Use the same Rotor position
> Keep none involved phases floating
> Ground connection

> Low inductive ground connection
> Same arrangement for highest reproducibility
U2 V2 W2
Phase U individually
Phase V individually
Phase W individually
1. Stator winding 1. Stator winding

> Star point closed > Delta connection
U1 V1 W1 U1 V1 W1
U2 V2 W2 U2 V2 W2
Phase U star Phase U delta

Phase V star Phase V delta
Phase W star Phase W delta

1. Stator winding 2. Rotor winding (entire winding)

> Overview about possible winding interconnections > If e.g. rotor is still installed in the machine
> ATTENTION: Results may differ between installed and pulled rotor.
Terminal for
Name of the Terminal for
Interconnection measurement
Source and
Response
Reference
III U U1 U2
Separated phases III V V1 V2
III W W1 W2
Y UN U1 U2-V2-W2
Star connection Y VN V1 U2-V2-W2
Y WN W1 U2-V2-W2
DU U1 U2
Delta connection DV V1 V2
DW W1 W2
2. Rotor winding (single pole measurement) 3. Measurement between Stator and Rotor
> If rotor is pulled and the poles are accessible > If the rotor winding is accessible
> ATTENTION: The position of the pole can influence the results > ATTENTION: The position of the rotor is strongly influencing the
measurement results
Comparison of six healthy poles

Methods of SFRA measurements Assessment methods
3. Measurement between Stator and Rotor 1. Time-based: earlier measurement of the same machine
> Only time based assessment possible
> Broken damper bars may become loose during operation and can damage
the entire machine
2. Type-based: measurement of a type-equal machine
3. Phase comparison: measurement of other phases or poles
Phase U vs. Phase V

Phase V vs. Phase W
Phase W vs. Phase U
Assessment methods Time-based assessment
1. Time-based assessment > Hydro Generator with multi-turn loop winding.
Reference
measurement
(fingerprint)
New
measurement
Comparison
Healthy winding
Winding with artificial fault

Time-based assessment Time-based assessment
> Hydro Generator with multi-turn loop winding. > Entire rotor winding with 10 salient poles
Turn to turn short in

phase U
> Page 36
Healthy winding
Winding with artificial fault
Time-based assessment Time-based assessment

Measurement between Stator and Rotor Measurement between Stator and Rotor
> Broken Damper Bar Detection with SFRA > Comparison of measurements performed
at the same rotor position

Assessment methods Type-based assessment
2. Type-based assessment Test Object 1:

> Example: Flat copper 6 pole rotor
DUT A
DUT B
Comparison
Healthy rotor Turn to turn short in one

pole
Type based assessment SFRA Measurement

Test Object 2: Test Object 2:
> Test rotor with 10 poles > Comparison between the entire healthy winding (green trace) and a short
circuit between 4 turns in one pole (blue trace)
> Winding was accessible to simulate different faults

SFRA Measurement SFRA Measurement
> Inter turn fault with 4 turns bridged (red) and healthy pole (green) > Inter turn fault with 4 turns bridged (red), healthy pole (green) and resistance
between the connections (0,5 Ω black; 1 Ω blue; 50 Ω purple)

> Measurement of entire winding with (red) and without fault (blue) > Test rotor with 8 poles and 240 turns per pole
> Winding was accessible to simulate different faults

> Comparison between healthy poles > Inter layer fault with 40 turns bridged
SFRA Measurement Measurements

> One turn of the 240 is bridged (black trace) and compared to the healthy > The influence of non-
pole (blue trace) ideal short circuit is
simulated by connecting
> Small difference between 1 and 10 kHz
some resistances
> Turn-to-turn short circuits
gets closer to the healthy
trace the higher the
resistance
> Same sensitivity to
conventional methodes
One turn
0,5 and 1 Ohm in shorted
between the contacts
Healthy
pole

Assessment methods Phase-based assessment
3. Phase-based assessment > Small motor with known defect in phase U (orange)
Phase U vs. phase V
> Comparison between the

three phases of the same
machine
U1 V1 W1
U2 V2 W2
Thank you

Introduction of Partial Discharge Measurements
OMICRON Electronics Asia Ltd. July 24, 2018

© OMICRON Page 54

Electric fields What is partial discharge ?
> Partial discharge (PD) is a localized dielectric breakdown of a small portion

of a solid or liquid electrical insulation system under high voltage stress.
kV kV
E 0,5 E 20
m m
> Definition from IEC 60270 Specification:

Localized electrical discharge that only partially bridges the insulation
between conductors and which can or cannot occur adjacent to a conductor.
photo: NWS
Evidence of partial discharges Evidence of partial discharges
photo: VATech Hydro Austria

photo: IPH Berlin photo: IPH Berlin
tracking
structures

Evidence of partial discharges Video ...
video: IPH Berlin
electron
Video ... Important units -
+
Charge Q [1Coulomb] = [1As]:
-
2nd orbit
Charge in movement = CURRENT!!! nucleus
1st orbit
The more electrons per TIME interval, the higher is the current
Current = Charge per Time

[A] = [Coulomb] per [s]
[A] = [Coulomb] / [s]
Charge = Current times Time

Charge = Current x Time
q= =
video: IPH Berlin

Charge – How to measure? Time Domain vs. Frequency Domain
> Time Domain Integration q= =
t1 t2
Ideal PD impulse (Dirac) consists of all frequencies!
Time Domain Ù Frequency Domain .... via Fast Fouier Transformation
= = ( )
SI prefixes Charge – How it is measured
> Frequency Domain Integration

Most used prefixes:
> More details in chapter FILTER = = ( )
kilo Volts [kV]
milli Ampere [mA]

micro Farad [μF]
nano Coulomb [nC]
pico Coulomb [pC]

Analog PD measurement systems ”Digital” PD measurement systems
Filter is analogue!
”Digital PD measurement system PD measurement with the OMICRON MPD software

with digital filtering
disturbed
voltage
signal
voltage
reading
test voltage
over time
PD activity
over time

What are partial discharges? Types of PD
Partial discharge: > Internal PD

> Void discharges, „electrical treeing“ Solid insulation
> Local electrical stress in the insulation or on the surface of the insulation
> External PD
> Always generates electromagnetic signals > Corona
> Surface discharges
> Often accompanied by an emission of sound, light, heat and chemical

reactions
Outer semicon
Solid
insulation
Inner
semicon
Image Source: Andreas Küchler – Hochspannungstechnik

External PD Internal PD
Internal discharge Cavity / void discharge

in laminated material
Surface discharge Corona discharge
Treeing

PD classification Thank you
‫ܭ‬n = 2
Lines of
electrical field
Semicon layer protrusion Void

(stress concentration at the tip) (field strength doubling)

What is important to PD on RotM?
PD sources / Noise separation is important
A: small voids on edge

B: Delamination: conductor – main insulation
C: Delamination of tape layers
D: Treeing in layers
E: Broken strands
G: Micro voids
F: Slot discharge, semicon paint abrasion
H: Discharges in Cooling Duct
I: Delamination of Insulation in Elbow
Seminar for Diagnostic Measurements and J: Endwinding surface discharge - contamination
On-line Monitoring on Rotating Machines K: Insufficient Spacing, Tracking and Sparking

L: Connection area between slot corona protection and endwinding
Advanced source separation method
corona protection

© OMICRON Page 2
Overview about PD faults Delamination: Conductor and main insulation (S1)
Outer corona protection Inner corona protection > PD intensity decreases with increasing temperature or load
> Predominant in the positive half cycle of the applied voltage
S1
S2
S1
S3
S3
S4
> Cause: thermal cycling or impregnation fault
Core winding: mica, paint, epoxy Laminated core / stator slot wall
> Insulation unsolders from the conductor

Delamination: Conductor and main insulation (S1) Delamination: Conductor and main insulation (S1)
Outer corona protection Inner corona protection > PD intensity predominant in the
positive half wave
> Degree of hazard is considered as
S1 dangerous according to IEC
S3
Core winding: Mica, paint, epoxy Laminated core / stator slot wall
Phase relation of PD with respect of void position Delamination of tape layers (S2)
> Voids next to the conductor Outer corona protection Inner corona protection
appear in the positive cycle.
> PD occurs in both half cycles if

voids are located in the middle.
S2
> Voids next to the laminated core
appear in the negative cycle.
S3
Core winding: mica, paint, epoxy Laminated core / stator slot wall

Delamination of tape layers (S2) Delamination of tape layers (S2)
Source: Brütsch et al. "Insulation Failure Mechanisms of

Power Generators", DEIS July/August 2008
Delamination of tape layers (S2) Slot discharge: Outer corona protection abrasion (S3)
Outer corona protection Inner corona protection

> No polarity predominance
> Cause:
> Inadequate resin impregnation during manufacturing
> Thermal influence during operation
> Delamination of tape layers
S3
S3
S2
Core winding: Mica, paint, epoxy Laminated core / stator slot wall

Slot discharge: Outer corona protection abrasion (S3) Slot discharge: Outer corona protection abrasion (S3)
> Test bar modification
> Removed outer corona
protection with slot simulation
Abrasion
Conductive cover
Slot discharge: Outer corona protection abrasion (S3) Micro voids (S4)
> PD intensity increases with increasing load Outer corona protection Inner corona protection
> Predominant in negative half cycle of the applied voltage
S3
S3
> Cause: S4
> Mechanical force on the coils
> Electrical deterioration of conductive layer (OCP)
> Damage of outer corona protection (OCP) or end-winding protection (EPG) Core winding: mica, paint, epoxy Laminated core / stator slot wall
> Loose coils in the slot


Micro voids (S4) Overview about PD faults - Winding overhang
End winding protection (EPG)
> Micro Voids are present in every mica-epoxy insulation system E3 Spacer
Outer
> Often in combination with other PD sources
corona
> Degree of hazard low according to IEC protection
E2
E2 E1
S3
E1
Conductor
Spring finger
© OMICRON Page 17 © OMICRON
Laminated core Page 18
End winding surface discharge: (E1) End winding discharge: (E1)

End winding protection (EPG) > Bar to bar activity
Spacer
Outer > In the end winding due to
corona insufficient spacing
protection
E1
E1
Conductor
Spring finger
© OMICRON
Laminated core Page 19 © OMICRON Page 20

End winding surface discharge: (E1) End winding surface discharge: (E1)
> Cause: dust or other contamination > Discharge present in both half
or bad performance of the end potential grading waves; no real predominance
> Predominance in the positive half cycle > Both shapes below are a possible
symptom of the defect
> Additional similar phenomenon in the negative half cycle possible (see next slide)
> Example on the right with inner

delamination and surface
discharges
Insufficient spacing: Corona (E2) Streamer Discharge (E2)

End winding protection (EPG)
Spacer > Erosion of Outer corona protetction (OCP) and/or end potential grading
Outer (EPG) at the slot end
corona
> Discharges to the press fingers
protection
> High discharges amplitude > 50 nC
E2
E2
Conductor
Spring finger
© OMICRON

Surface discharge: Bad connection (E3) Surface discharge: Bad connection (E3)
Insulating tape
End winding protection > Bad connection area between
E3 outer corona protection (AGS, in
Outer Spacer
the slot) and end winding corona
winding protection (EPG)
protection
> Modification of a test bar in
combination with EPG
Conductor
Mounting parts
© OMICRON
Surface discharge: Lost connection between Summary of PD sources

gradings(E3) Abbreviation Source Explanation
Delamination at the boundary between conductor and main insulation.
Delamination of insulation tapes
S1 Often result of overheating or extreme mechanical forces leading to
from winding conductor larger separation of the layers.
Delamination within the main insulation can be caused by incomplete
S2 Delamination of insulation tape layers curing of the insulation system during manufacture or by mechanical or
thermal overload during operation.
Slot PD’s are caused due to abrasion of the outer corona protection
Abrasion of slot corona
S3 between stator bar and slot of the stator core mainly due to operational
protection tape / paint vibration or movement.
Micro voids / cavities are internal PDs generated within air or gas filled
S4 Micro voids / cavities pockets within the main insulation. They result from the manufacturing
process and do not indicate aging factors.
End-winding / surface PD’s generated on the surface of the insulation
End-winding surface discharge / material and are located naturally in the end-winding section of a
E1 machine. They result often from conductive contamination (carbon, oily
tracking
dust, abrasion etc.) or from damaged field grading materials.
This type of activity will occur between two bars in the winding
End-winding discharges in
E2 overhang, or between a bar and the press finger or sharp edges of the
gas / sparking stator core but also within the gap between the bar/ coil and slot wall.
Arise at the juncture between OCP and EPG, if the field control system
Discharge between Outer Corona
E3 is not adequate or effective, because of poorly designed contact
Protection and End potential grading points, contamination, porosity, thermal effects
Noise. From external source conducted or irradiated noise signals
N1 Noise (asynchronous noise) which is asynchronous with AC voltage. Disturbances are not caused
by stator winding insulation system.
Disturbances. Typical 6 pulses caused by thyristor switches of the
N2 Excitation / converter noise exciter / converter by synchronous generators. They can be also 12
and 24 pulses.
© OMICRON Page 27

Summary of PD sources What is important to PD on RotM?
high S1 high S2 high S2 high S3
PD sources / Noise separation is important
Delamination of insulation tapes Delamination of insulation tape Delamination of insulation Abrasion of slot corona
from winding conductor layers; single sources tape layers protection tape / paint
low S4 low S4 medium E1 medium E1
End-winding surface End-winding surface

Micro voids / cavities Micro void – different scaling discharge / tracking discharge / tracking
medium E1 medium E1 medium E2 medium E3
End-winding
medium surface End-winding surface End-winding discharges in Bad conection between OCP and
discharge / tracking discharge / tracking gas / sparking EPG
medium E3 medium E3 Disturbances Disturbances
Discharge between Discharge between Noise (asynchronous noise) Excitation / converter noise
OCP and EPG OCP and EPG
Attention: all PRPD shapes are displayed in unipolar and linear scaling
P
Challenge: Overlapping of PD sources Coupling of PD into other phases

PD Software Display 3PARD Analysis Principle
Capacitive
Capacitive &&coupling
Capacitive
Inductive Inductive
Inductivecoupling
coupling couplingtotothe
thewindings
wingdings
Capacitive & Inductive coupling
Noise ofofthe
same
other
phase
two phases
200 pC
C
Noise, asynchronous with AC

E
E
medium
900 pC
C
H
Inner PD + Endwinding surface

discharge - contamination
3 2 1 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 36 35 34 33 32
high
PD
CORE
500 pC
Delamination of tape layers- L1

bigger inner micro void L2
L3
© OMICRON Page 31 neutral© OMICRON Page 32

Coupling of PD into other phases Coupling of PD into other phases
3PARD Analysis Principle 3PARD Analysis Principle
Inner PD Source Outer Noise
3PARD 3PARD
L1>L2>L3 L1 §L2 §L3
L2 L1 L2 L1
Outer Noise
Inner PD in L1
L2 L2
L3 L3
L3 L1 L3 L1
timeframe timeframe
3PARD = Three Phase Amplitude Relation Diagram 1 μs
3PARD = Three Phase Amplitude Relation Diagram 1 μs
Because of the complex propagation and coupling characteristics of PD impulses, every source of PD Electrical interferences and outer noise usually have the pulse-like nature of PD but
impulses has unique impulse to couplings ratio and consequently a characteristic 3PARD signature occur with almost same amplitude on all three phases of the machine.
What is important to PD on RotM? Example – Using 3PARD in a synchronous 4-Channel

PD Measurement
Signal Separation Technique – 3 PARD (3 Phase Amplitude Relation Diagram)

Example – Fault Separation with 3PARD Example – Back transformation of cluster A
Cluster B
Cluster C
Cluster A
Example 1 – Back transformation of cluster B Challenge: Overlapping of PD sources

B
Fault separation: 3CFRD/3FREQ
Ck
CD
MPD 600
Source 1
Source 3
MPD 600
Source 2

Methods 3CFRD (3FREQ) – How it works
Fault separation: 3CFRD/3FREQ Phase-resolved pattern 3FREQ
fc2
FFT pulse 3
FFT pulse 1
FFT pulse 2
Surface discharge Corona discharge Internal void

PD source 1
PD source 2
PD source 3
fc3 fc1
Thank you

Partial Discharge Measurements on Rotating Machines

© OMICRON Page 43

PD Measurements on Rotating Machines
Agenda
Why should PD-measurements be conducted on rotating
> PD Measurement on Rotating Machines electrical machines?
– Why should PD-measurements be conducted on rotating electrical
machines? > Electrical field is exceeding the local
– I have already done an insulation resistance measurement and a electrical field strength in the
dissipation factor measurement, why should I perform an additional PD insulation Epoxy
resin
measurement? > Due to imperfections during the
– Which is the most suitable way to acquire typical PD of stator impregnation
windings? > Surface discharges
Conductor strand
– What else is influencing the PD measurement?
> By-Products, especially Ozone: Mica
– Should I measure PD on my new machines too? flakes
> Very corrosive to metals
> Polymers are attacked Æ
embrittlement
Theoretical
> Can be symptom for mechanical discharge channel
defects
Outer corona
> OFF-line and ON-line measurements protection (OCP)
are possible
Benefits of PD-measurement Benefits of PD-measurement
> At least two PD sources are present when measuring an epoxy mica In stator windings of HV rotating machines usually we are distinguishing between:
insulation of a HV rotating machine :
> “Normal” Partial discharges:
> Some kind of „noise“ caused by the measurement environment
> Micro voids, generated during the manufacturing process Can already inject in the range of 0,2 ... 0,3 UN/√3
Are not critical regarding the lifetime of the asset
Can influence the diagnosis of the “relevant” PD sources (mapping)
> Critical Partial discharges:
Known and verified by international publications (see also IEC/TS
Source: Vogelsang et al. “Performance testing of high
voltage generator- and motor insulation", Dec 2005 60034)
Due to unsymmetrical behavior, dimensions of the faults, properties of
> As epoxy mica is a compound insulation system, the mica is acting as a the electrodes and other influences, the faults and damages can be
barrier against PD differentiated from “normal” PD
Interpretation of the measurement

Source: R.Vogelsang; Detection of electrical tree propagation by
partial discharge measurements; ETEP; 2005

Benefits of PD-measurement Benefits of PD measurement
All PRPD shapes
S1 S2 S2 high S3
high high high are displayed
unipolar and
linear
Delamination of insulation tapes Delamination of insulation tape Delamination of insulation Abrasion of slot corona
from winding conductor layers; single sources tape layers protection tape / paint Risk assessment
low S4 low S4 medium E1 medium E1 according to IEC MICAMAX plus, Monitoring System for
condition Based Maintenance, PD Tech
Power Engineering AG, September 2002

Micro voids / cavities Micro void – different scaling discharge / tracking discharge / tracking
medium E1 mediu E1 medium E2 medium E3

m
R. Brütsch, T. Hillmer, Von Roll, High

End-winding
mediusurface End-winding surface End-winding discharges in Bad conection between OCP
Voltage Main Stator Insulation, EEIM
discharge / tracking discharge / tracking gas / sparking and EPG Seminar, Berlin, Oktober 2005
medium medium
E3 E3 Disturbances Disturbances
Discharge between Discharge between Noise (asynchronous noise) Excitation / converter noise
Benefits of PD measurement Use Case: High Speed Motor

I have already done an insulation resistance measurement and > High speed Synchronous Motor, key asset for
a dissipation factor measurement, why should I perform an production
additional PD measurement?
> Previous tests:
> DC insulation resistance @600sec.:
> L1: 117 GΩ
> PD measurement is a powerful tool to check the insulation condition without
> L2: 97 GΩ
dismantling the machine > L3: 90 GΩ
> Polarization Index:
> The measurement indicates defects long before the above mentioned methods > L1: 5,9
> L2: 7,2
> PD measurement does not provide a single integrative value which is > L3: 3,9
representing the whole winding > Dissipation factor measurement:
> L1: 1,69% @ UN/√3
> Single fault types can be identified > L2: 1,67% @ UN/√3
> L3: 1,63% @ UN/√3
> High partial discharge activity in the end-winding

> In the range of 100 nC
> Indication for completely eroded stress grading

High Speed Motor Influence of the PD-measurement approach
The Motor was in a critical condition: Which is the most suitable way to acquire typical PD of stator
> High PD is deteriorating the insulation windings?
> PD is causing By-products such as Ozone

> Ozone is attacking metals, silicone and other > PD-measurement equipment with LF filter
materials frequencies (10 kHz ... 1 MHz) with quasi-
integrating approach; Measurement of
„apparent“ charge
> In this particular case a chemical reaction > PD-measurement equipment with VHF filter
caused nitrites (small salt grains) which were frequencies (30 MHz ... 300 MHz) with peak
acting like a sand blast in the cooling circuit of detecting approach; Measurement in mV
the machine
> PD-measurement equipment with HF range
> A new Motor was ordered (3 MHz ... 30 MHz) and UHF range (300
> Meanwhile the star point of the Motor was MHz ... 3 GHz) are not treated in this
changed to enable additional lifetime, until the presentation
replacement is ready.
Influence of the PD-measurement approach Influence of the PD-measurement approach
> Injection of a known PD-pulses along the winding of a decommissioned 10nC at a), Measurement at XB 10nC at b), Measurement at XB
generator (105 MVA) 0,2 0,2
155 mV
Voltage at XB in V
60 mV
Voltage at XB in V
0,15 0,15
0,1 0,1
0,05
9,78 nC 0,05
9,21 nC
a) 0 0
-0,05 -0,05
-0,5 0,0 0,5 1,0 1,5 -0,5 0,0 0,5 1,0 1,5
DSO Time in μs Time in μs

b)
> Very strong attenuation of the high frequency parts already
after the first bar (original peak was 2,6 V)! a)
XB > With such strong signal attenuation it can be assumed, that DSO
most of the PD-sources “deeper” in the winding are not

detected.
b)
> Quasi integration of the signal seems to be more robust. XB

Influence of the PD-measurement approach Other Influences
> The VHF part of PD impulses is dominated by the winding section, which is
located very close to the terminal. What else is influencing the PD measurement?
> This could be the right diagnosis but not necessary!
> Environmental influence:
Erosion of the grading in the
slot exit
> Humidity (especially for external discharges); the lower Æ the better
> Temperature
High mechanical force at the
spacer 2001 @ 15% RH Æ app. 25 nC 2003 @ 30% RH Æ app. 10 nC
Eroded outer corona protection due

to a loose bar or vibration sparking
> The LF-filter bandwidth is recommended in the IEC standard (IEC 60270
Source: E.Binder, OMICRON Diagnostic Week, Bregenz, March 2017
and IEC TS 60034-27)

Other Influences Comparison of 4 sister machines
> Voltage shape

> real-time sinusoidal shape Should I measure PD on my new machines too?
෡
௏
> Ratio between and
ଶ
ܸோெௌ is a good indicator
> YES! In order to get a fingerprint value. Trending over time is the most powerful
method to detect changes in the insulation system
> YES! A factory acceptance test ensures the quality of the produced and installed
machines
> Use case from industry:

> Large industrial facility in Europe
> Over 500 HV-Motors
> Outage costs of one production line up to 150.000 €/d
> Operator has PD measurement equipment and is

> Changes in the single PD sources during the measurement testing on a regular base ON-line and OFF-line
> Gas pressure in the voids
> Changing surface condition when slot discharges occur > New machines are tested offline before they are
> ... installed in the production line

Comparison of 4 sister machines Conclusion
> Partial discharges are often early warnings and symptoms for other defects
> 4 NEW Sister Machines
(6kV rated Voltage) > Defects and damages can be identified due to their typical pattern and
classified regarding their risks according to international standards
> Manufacturer was faced
with the results > Not only the discharge itself, but also the By-products can be dangerous to
the machine
> The PD measurement method is strongly influencing the results; a quasi-
integrating method is highly recommended
Very good conditions for
this particular acquisition > Always record the parameters, if you are comparing two measurement
results
> Fingerprint measurements with periodical measurements is the most
powerful tool to check for any changes in the insulation system
No problems with further
purchases of machines
Thank you

Propagation of Partial Discharges in
Rotating Machines Windings
© OMICRON Page 20

Motivation Test Object
> Partial discharge measurement as a useful Instrument of rotating electrical > Small Hydro Generator taken out of service:
machine diagnosis > Form wound coil winding with 156 slots; two coils per slot; two turns per coil
> Single foreign matters can be detected > Inner diameter 2,2 m; Core length 57 cm
> Specific pattern as typical symptom depending on the type of defect > 6,4 kV; 5,4 MVA
> Maintenance Tool for non-invasive testing
> Measurement requires an interpretation

> Awareness of the effect of the environment, the measurement technique
and parameters
Coupling Mechanism Cross coupling in the end winding

Different propagation mechanism of high frequency impulses in rotating Measurement setup
machines windings are known:
> Direct injection into the winding
> Cross coupling over the end winding
> End winding loops are representing a single turn transformer, coupled via air > Measurement on two channels
> Depending, if rotor is installed or not (offline) with digital oscilloscope
> Cross coupling via slot > Build up of an “artificial end

> PD is “jumping” to other phases also via the slot winding” (AEW)
> This can be seen if the propagation speed is measured
> AEW not connected to the
> Propagation in the winding winding, only coupling via air
> The PD signal is travelling along the winding
> Highest signals when measurements are taking
> Focus on this propagation way in this presentation
XB
DSO

Cross coupling in the end winding Measurement Setup
Installation of an artificial end-winding Two main investigations were performed:
> Fast coupling mechanism after 12 ns the > Injection of artificial impulses
signal was visible at the artificial end > Access the winding directly in the end
winding (blue) winding area
> Partial discharge measurement with a
> Strong damping frequency selective device
> High impedance time domain
> Propagation speed in the winding appr. measurement with DSO (2,5 Gs/s)
100 m/μs > Coupling Impedance was simulating a
single channel PD-measurement
> Provocation of “real” PD events

> With Outer Corona Protection (OCP)
Island in the slot
> Partial discharge measurement with a
frequency selective device
> High impedance time domain Measurement on
measurement with DSO (2,5 Gs/s) coupling capacitor
> Coupling Impedance was simulating a
single channel PD-measurement Injection points of 10nC
for attenuation matrix
„Real“-PD events
Measurement Setup Measurement Setup

Injection of artificial impulses Provocation of “real” PD
> Calibrator impulse with < 10ns rise time and 34 V peak > Electrode arrangement in the same slot part than the artificial impulses were
injected
> Known apparent charge of 10 nC
> Inception voltage between 1100 V and 1300 V, same size of the copper foil
> 15 injection points along the winding
> Calibration arrangement included > 12 electrodes along the stator winding
HFCT on Trigger
Source: E.Binder, “Aspekte zur Analyse von Teilentladungsaktivitäten von

grounding braid
Staänderwicklungen,“ in OMICRON Diagnosewoche, Bregenz, 2017

Thread
Discharge
OCP
Island
Stator
with
core
.
.
.
.
copper
Stator
strip
winding
Ground
wall
insulation

Results Results
Partial Discharge Measurement of artificial impulses Partial Discharge Measurement of “real” PD events
> Outcome: Attenuation Matrix Æ Normalized attenuation of the different > Outcome: Attenuation Matrix Æ Normalized apparent charge of the “real” PD
injection points measured with different filter frequencies at the terminals events, measured with different filter frequencies at the terminals
Position of the PD event along the winding

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
1,2 1,2
1 1
q(x) / q0
0,8
0,8
q(x) / q0
0,6
0,6
0,4
0,4
0,2
0,2
0
0
100 kHz - 200 kHz 800 kHz - 1450 kHz 50 kHz - 90 kHz 50 kHz - 90 kHz 130 kHz - 170 kHz 400 kHz - 440 kHz
Results Results
Time Domain Measurement Time Domain Measurement, Data Processing with different sliding averages
Original Sampling Window: 100
Trigger Impulse values
U2 U1 DSO
100 % 0% CC
> High frequency components are overlapping the measurement Sampling Window: 500 Sampling Window: 1500
values values
> Data processing to analyze the low frequency components

Results Results
Time Domain Measurement of artificial impulses Time Domain Measurement of “real” PD events
Impulse Distribution
Impulse Distribution
0,6
4
0,4
3
Voltage in V
0,2
Voltage
2
in V
12 0
1 10 12
8 1
- 80
0 6
4 0,2 6
-5 2 - 4
0 5 10 2
15 20 25 0 0,4-5 0
30 5 10 15 20 0
25 30
Time in μs Time in μs
Results Interpretation
Voltage peak measurement BEFORE data processing
Partial discharge measurement
> Highest peak value for normalization
> Artificial impulses: 34 V > Good winding coverage with low filter frequencies
> “Real” PD events: 17,1 V > With 50 kHz – 90 kHz the injected impulse of 10 nC was measured as > 5nC over
50% of the winding
Position of the event > Benefit of a frequency selective measurement system
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% > By using higher frequencies, only the PD pulses close to the measurement point
1 are recorded with reasonable amplitude.
0,8 > Typical resonance spots are observed

Vpeak(x)/ Vpeak0
> Due to superimposition of the forward impulse component and the reflected
0,6 > Only applicable for this specific stator
0,4
0,2
0
"Real" PD Artificial Impulses

Interpretation Interpretation
Similar results of apparent charge between artificial impulses and “real” PD
Time domain measurement
> Signals are stretched due to damping, cross coupling, dispersion and
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% travelling wave reflections
1,2 > Peak value decreases after a short winding portion to less than 20% of the
1 original peak
> Coupling over the endwinding
0,8 > Increase at the starpoint due to the position of the connection bars
q(x) / q0
0,6
> Low frequency components visible over the entire winding
0,4 > Run time measurement not really successful
> Cross coupling was not considered
0,2
0
100 kHz - 200 kHz artificial impulse 50 - 90 kHz "real" PD
50 kHz - 90 kHz artificial impulse 130 - 170 kHz "real" PD
Conclusion Thank you
> A method to investigate the behavior of PD in a machine is compared to PD

events provoked by high voltage
> Good correlation regarding the partial discharge measurement and the time
domain measurement
> The particularities of both – the PD measurement with quasi-integration and

the time domain measurement – are presented
> Slower frequency components are visible over a large winding section
> Signals are "stretched and shrinked" after a short winding portion
> Typical shape of impulses according to their origin
> The advantages of a frequency selective PD measurement are shown
> Same parameters should be used to compare measurement results
> Strictly speaking the results can only be obtained for the investigated machine fabian.oettl@omicronenergy.com
> The results are matching the outcome of similar investigations

Why Partial Discharge on Rotating Machine?
Why partial discharge is vital

Thermal
> Stator insulation faults are the second most common damage
17%
cause
Insulation
damage Mechanical
> On-line PD measurement/monitoring can access the 56% damage
condition of a stator winding insulation system 24%
under real service conditions

> Allows long-term planning of maintenance and outages Bearing
damage
> Enables condition-based vs. time-based maintenance Other
3%
stator
High faults
10% Rotor faults
Condition-based maintenance 10%
Insulation strength
Bearing faults
Other
Seminar for Diagnostic Measurements and Condition trigger
41%
12%
On-line Monitoring on Rotating Machines Time-based maintenance Stator

insulation faults
PD monitoring system on Rotating machines Failure

27%
Operating stress
t = Period between routine
Source: CIGRE, 2009 Survey of
Low t time-based maintenance checks
Hydro Generator Failures
© OMICRON Page 2
Why Partial Discharge on Rotating Machine? What is important to PD on RotM?
Partial Discharge is an EARLY indicator of stator insulation failure PD sources / Noise separation is important
> PD occurs in the stator insulation system of motors and generators, where
local electric field stress exceeds the local electrical strength.
A: small voids on edge
> Void, Cavity, Treeing
> Delaminated tape layer B: Delamination: conductor – main insulation
> Abrasion C: Delamination of tape layers
> Contamination or bad connection on end winding/overhang D: Treeing in layers
> An increase of PD activity can indicate insulation degradation caused by E: Broken strands
aging processes such as overheating, load cycling and coil vibration G: Micro voids
F: Slot discharge, semicon paint abrasion
H: Discharges in Cooling Duct
Slot 99 Slot 234
I: Delamination of Insulation in Elbow
J: Endwinding surface discharge - contamination
K: Insufficient Spacing, Tracking and Sparking
Visual inspection confirms surface L: Connection area between slot corona protection and endwinding
PD activity on stator slots corona protection
Source: PJB (PT Pembangkitan Jawa-Bali)

Summary of PD sources Summary of PD sources
Abbreviation Source Explanation high S1 high S2 high S2 high S3
Delamination at the boundary between conductor and main insulation.
Delamination of insulation tapes
S1 Often result of overheating or extreme mechanical forces leading to
from winding conductor larger separation of the layers.
Delamination within the main insulation can be caused by incomplete Delamination of insulation tapes Delamination of insulation tape Delamination of insulation Abrasion of slot corona
S2 Delamination of insulation tape layers curing of the insulation system during manufacture or by mechanical or from winding conductor layers; single sources tape layers protection tape / paint
thermal overload during operation.
Slot PD’s are caused due to abrasion of the outer corona protection
low S4 low S4 medium E1 medium E1
Abrasion of slot corona
S3 between stator bar and slot of the stator core mainly due to operational
protection tape / paint vibration or movement.
Micro voids / cavities are internal PDs generated within air or gas filled
S4 Micro voids / cavities pockets within the main insulation. They result from the manufacturing
process and do not indicate aging factors. Micro void – different scaling
Micro voids / cavities discharge / tracking discharge / tracking
End-winding / surface PD’s generated on the surface of the insulation
End-winding surface discharge / material and are located naturally in the end-winding section of a
medium E1 medium E1 medium E2 medium E3
E1 machine. They result often from conductive contamination (carbon, oily
tracking
dust, abrasion etc.) or from damaged field grading materials.
This type of activity will occur between two bars in the winding
End-winding discharges in
E2 overhang, or between a bar and the press finger or sharp edges of the
gas / sparking stator core but also within the gap between the bar/ coil and slot wall. End-winding
medium surface End-winding surface End-winding discharges in Bad conection between OCP and
discharge / tracking discharge / tracking gas / sparking EPG
Arise at the juncture between OCP and EPG, if the field control system
Discharge between Outer Corona
E3
Protection and End potential grading
is not adequate or effective, because of poorly designed contact medium E3 medium E3 Disturbances Disturbances
points, contamination, porosity, thermal effects
Noise. From external source conducted or irradiated noise signals
N1 Noise (asynchronous noise) which is asynchronous with AC voltage. Disturbances are not caused
by stator winding insulation system.
Disturbances. Typical 6 pulses caused by thyristor switches of the Discharge between Discharge between Noise (asynchronous noise) Excitation / converter noise
N2 Excitation / converter noise exciter / converter by synchronous generators. They can be also 12
and 24 pulses.
Attention: all PRPD shapes are displayed in unipolar and linear scaling
© OMICRON Page 6
What is important to PD on RotM? Challenge: Overlapping of PD sources

PD Software Display
PD sources / Noise separation is important Noise
200 pC
C
Noise, asynchronous with AC
medium
900 pC
C
Inner PD + Endwinding surface

discharge - contamination
high
500 pC
Delamination of tape layers-

bigger inner micro void
© OMICRON Page 7
P © OMICRON Page 8

What is important to PD on RotM? What is important to PD on RotM?
Signal Separation Technique – 3 PARD (3 Phase Amplitude Relation Diagram) To see wide range of stator winding
What is important to PD on RotM? How do we measure/monitor PD?
PD activity happening in deep inside of stator winding Fixed Monitoring On-line measurement
MONGEMO vs. OMS 605
iPD
wrong filter settings

A good filter settings
(too high) f
= 9 Continuous monitoring purpose 9 Usable for on-line measurement /
2 ( )
Low capacitance 9 No human involvement & periodic monitoring purpose
e.g. 80pF(40MHz~) Automatic classification 9 Portable solutions for multiple machines
High capacitance 9 Accessible / flexible to customer’s
e.g. 2nF(1.5MHz~) network
© OMICRON Page 11

OMS 605 system configurations OMS 605 system configurations
Generators and motors High-voltage cables and accessories

(Single or multiple machines)
OMS 605
Coupling capacitors
(3 x per machine) Terminal box (optional)
Notebook PC
with software
Terminal box (PC not included)
HFCT sensor (3 x per
(optional)
accessory group)
OMS 605
Notebook PC
with software
(PC not included)
Note: A terminal box can be used if the sensors are permanently installed.
© OMICRON Seite 13 © OMICRON Seite 14
OMS 605 system configurations Working Procedure
Bushings and power transformers

(Conventional PD and UHF measurements) Bushing tap Installation (Machine is off)
adapter
(1 x per channel) Installation of
Coupling Capacitors Calibration(Divider factor) System Check
/Terminal box
Terminal box
(optional) Commissioning (Machine is on)
HFCT
UHF drain valve (optional - for
sensor (optional) PD detection Tune measuring Find
Save setting file
or gating) Frequency Best SNR
UHF extension
kit (optional)
Periodic Measurement (Machine is on)
Analysis with
Load Saved setting file Record Stream
3PARD/3FREQ
Notebook PC OMS 605
with software
(PC not included)
© OMICRON Seite 15 © OMICRON Page 16

MONGEMO : System configuration MONGEMO : System Integration
Extendable system for one or several machines at one location Easy integration with third-party devices
and SCADA systems
Single First Second Third Fourth
machine machine machine machine machine > Integration of several types of monitoring sensors
> i.e. temperature, voltage, power, H2 pressure, etc.
> Send data from MONGEMO central computer

to other systems (i.e. SCADA)
Coupling
capacitors > Supports multiple industrial communications
Add
machines to standards:
Protected
be monitored > Modbus (RTU, TCP/IP) Coupling
enclosure with
acquisition unit
> DNP3 capacitors
> IEC 61850
Protected
Central computer Central enclosure with
with monitoring monitoring acquisition unit
software server and 3rd Party monitoring sensors
software (i.e. temperature, voltage,
power, H2 pressure, etc.) Central computer
Multiple machine monitoring system with monitoring
SCADA system
software
MONGEMO : Monitoring Software MONGEMO : Advanced Technique
Data visualization Automated Cluster Separation

> Configurable real-time and trend charts show PD data and alarm status > Automated data evaluation and
for each monitored phase/channel cluster separation are performed
regularly, periodically or when
> Time stamp and phase-related PD data available for each data point triggered by an alarm (violation of
> All data records and corresponding charts can be saved and exported pre-set threshold).
> For each cluster, the MONGEMO
monitoring software identifies the
phase of signal origin.
Real-time PD data (seconds, minutes) Historical PD trend data (hours, days, years)

MONGEMO : Advanced Technique Automated pattern classification
Pattern Classification How does it work?

> PatClass uses as input data PD Step 1: Automated
data sets recorded periodically PD separation Undefined
by the Monitoring Software. NO
Sufficient
YES
data?
> Very short time needed to PD is
generate a detailed report Step 2: Discrimination Noise
Noise is detected Noise detected
Identification
regarding the origin and type of between PD and noise type
Deterministic
the PD signals detected. Deterministic decision decision is
is achieved not received
Knowledge-
> Users can scan observe/track PD based Analysis
the suspected clusters by Step 3: Knowledge Source
themselves. based analysis result

Pattern
> Simplifies the analysis of data for Recognition
non-specialist users. Step 4: Pattern recognition Probabilistic decision is achieved
approach result
PD
Source
Thank you

Benefits from Continuous Monitoring of Rotating Machines

© OMICRON Page 23

Agenda Partial Discharges Monitoring System
> Introduction The crucial characteristics to guarantee the best insulation

assessment :
> PD Monitoring System Architecture
A synchronous multi-channel data
> Case study: Turbo generator 180 MVA, 16.6 kV acquisition stage:
• ensure high flexibility for different installation
> Case Study: Turbo generator 100 MVA, 13.2 kV conditions;
• provide advanced tools for in-depth analysis.
> Case Study: Hydro generator 126 MVA, 16.5 kV
A server based monitoring software to:
> Automated Pattern Classification System • trend and store significant data for analysis
and comparison of different machines over
> Conclusions their operating life;
• be easily accessible remotely for periodical
evaluations and/or system tuning.
Page 2 Page 3
Turbo generator: 180 MVA, 16.5 kV Turbo generator: 180 MVA, 16.5 kV
Preventive maintenance based on PD monitoring
Before maintenance Maintenance After maintenance
Page 4 Page 7

Turbo generators: 100 MVA, 13.2 kV – air cooled Turbo generators: 100 MVA, 13.2 kV – air cooled
Combined PD data trend from Gen. 1 and Gen. 2
One – year PD data trend PD data trend Sept. – Oct. 2013 PRPD diagrams for central freq. of measurements: 2 MHz and 7 MHz
¾ Both units commissioned in 1981

¾ PD monitoring system installed in 2012
Page: 15 Page: 6
Hydro generator: 126 MVA, 16.5 kV Hydro generator: 126 MVA, 16.5 kV
UNIT 7 and UNIT 8 were installed in 1997 in one of the largest

Unit 7 showed a steep increase in the trend of PD activity in phase S,
hydroelectric power plants in Southeast Asia (1008 MW) result confirmed by the comparison with the results from Unit 8
By means of 3PARD
• identify and evaluate the reason of the
separation technology increased PD activity;
The installation of PD Monitoring completed in the 2nd half of 2012 and PD Pattern analysis,
root cause investigation • winding.
assess the overall conditions of the stator
has been carried out to:
Page 8 Page 9

Hydro generator: 126 MVA, 16.5 kV Hydro generator: 126 MVA, 16.5 kV
Root cause investigation Overall assessment of the stator

01/03/13 12/04/13 28/05/14
By rejecting only the investigated PD activity, the overall conditions of the stator
winding can be assessed
3PARD cluster
01/03/13 12/04/13 28/05/14
under
investigation
PD Source
under
investigation
• root cause of the growing trend: efficiently
isolated and analyzed in time;
Results • conditions of the stator winding: removing
the investigated PD, general assessment can
an anomalous activity has been isolated in Phase S and it’s persistence
be performed.
confirmed by the historical data
Page 10 Page 11
Hydro generator: 126 MVA, 16.5 kV Partial Discharge Monitoring System
Diagnosis
Hardware Architecture
Investigated PD Overall conditions of the
activity ground-wall insulation
• Internal PD activity uniformly 1st generator 2nd generator 3rd generator 4th generator
• Surface Tracking evolving distributed on the three phases
in time;
Visual Inspection Dissipation Factor test Capacitive sensors
Fiber optic Fiber optic

cable cable
Central computer & monitoring software
Page 12 Page: 13

Partial Discharge Monitoring System Partial Discharge Monitoring System
Web GUI Monitoring software architecture
Page 14 Page 15
Partial Discharge Monitoring System Partial Discharge Monitoring System
Analysis Services Trend Analysis
Trend PD Source Pattern Alarm / Threshold PD Source

Analysis Separation Classification Warning Separation
Q: Has something Q: How many Q: What are the Q: How dangerous

happened on problems are problems – defects are the problems?
HV asset? there? or noise?
Page 16 Page 17

Partial Discharge Monitoring System PD Measurements – 3PARD analysis
PD Source Separation
(automated OPTICS clustering)
> © OMICRON
E
E
PD Pattern
Classification
3 2 1 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
36 35 34 33 32
TE
CORE
L1
L2
L3
neutral
Page 18 Page 19
PD Measurements – 3PARD analysis PD Measurements – 3PARD analysis
phase of PD origin
PD cross-talk
PD cross-talk
∆t 1/∆t ~ selectivity
PD cross-talk between phases PD cross-talk between phases
Page 20 Page 21

PD Measurements – 3PARD analysis PD Measurements – 3PARD analysis
3PARD Inner PD-source 3PARD Noise

L1>L2>L3 L1 ≈ L2 ≈ L3
L2 L1 L2 L1
noise
Inner PD source in L1
L2 L2
L3 L3
L3 L1 L3 L1
timeframe timeframe
1 μs 1 μs
Because of the complex propagation and coupling characteristics of PD Electrical interferences and outer noise usually have the pulse-like nature of PD
impulses, every source of PD impulses has unique characteristic 3PARD but occur with almost the same amplitude on all three phases of the machine.
signature
Page 22 Page 23
© OMICRON 5/30/2018 5/30/2018
PD Measurements – 3PARD analysis Automated Cluster Separation Using 3PARD

L2
Cluster 3
Cluster 2
Cluster 1
L1 L3
Cluster 1 Cluster 2 Cluster 3
Synchronous multi-channel and multi-spectral PD source separation

Page: 24 Page: 25

PD Pattern Classification PD Pattern Classification
Reference Database Reference Database

(examples of PRPD patterns - stator winding PD defects) (examples of disturbances)
Name PRPD illustration Name PRPD illustration
Delamination of the
rmS1 insulation on the Noise
copper side N1
(asynchronous noise)
Delamination of the
rmS2
insulation layers
Cause of PD in Stator Winding Excitation /
N2
Delamination of the converter noise
rmS3 insulation on the
core side
Discharges in micro Disturbances

rmS4 N3
cavities (synchronous noise)
End-winding surface
rmE1
discharge (tracking)
End-winding
rmE2 discharges /
sparking
Discharge between
corona protection
rmE3
and stress grading
layers
Page 26 Page 27
Patterns Features (descriptors) Algorithmic Concept
PRPD
Hq - charge vs. Hn – impulse amount vs. Hqn – avarage charge vs.

impulse number phase angle phase angle
Page 28 Page 29

Algorithmic Concept: attributes for knowledge-based analaysis Algorithmic Concept. distance plots
robust pattern
recognition dececision
uncertain pattern
recognition dececision
Page 30 Page 31
Report example – PD PD Data Evaluation
Page 32 Page 33

PD Data Evaluation PD Data Evaluation
01/2017 02/2017 03/2017
3PARD
Case study 2
PRPD
End-winding surface discharge End-winding surface discharge End-winding surface discharge
(tracking) – E1 (tracking) – E1 (tracking) – E1
Result
Decision Path: Decision Path: Decision Path:
Sufficient data → Not noise → Sufficient data → Not noise → Sufficient data → Not noise →
Knowledge-based decision - unclear case Knowledge-based decision - unclear case Knowledge-based decision - unclear case
(between E1 and E2) → Probabilistic decision (between E1 and E2) → Probabilistic decision (between E1 and E2) → Probabilistic decision
by Pattern Recognition (E1 50%) by Pattern Recognition (E1 100%) by Pattern Recognition (E1 50%)
Page 34 Page 35
PD Pattern Classification Conclusions

¾ Continuous PD monitoring of stator windings in large generators has
proven to be an essential diagnostic tool for condition-based
Decision Reliability
maintenance in order to increase their reliability.
• The decision is based on separated initial data corresponded to only one ¾ Based on the PD monitoring results obtained before and after the
PD source (automated clustering) maintenance of the machine, the effectiveness of the maintenance
program can be evaluated and future maintenance can be properly
• Insufficient data is rejected at the early stage of analysis scheduled.
• Only few approved PD features (attributes) are considered for the decision
¾ A versatile PD acquisition unit can be used for PD measurements on
• Reliable deterministic decision is provided for clear cases by knowledge- different types of rotating machines.
based analysis
¾ A modular, distributed monitoring software system allows reliable long-
• Probabilistic decision is provided for unclear cases with different levels of term storage of monitoring data and provides access via a web
reliability defined by the distance plot
interface for detailed data evaluation.
¾ The main key to performing appropriate diagnosis of the insulation

state is to accurately separate and identify different PD sources from
noise. To achieve this, synchronous multi-channel 3PARD evaluation
techniques can be applied.
Page 36 Page 37

Thank you
fabian.oettl@omicronenergy.com
© OMICRON Page 38

Makalah Seminar Omicron 24-25 Juli 2018

Uploaded by

Copyright:

Available Formats

Makalah Seminar Omicron 24-25 Juli 2018

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Makalah Seminar Omicron 24-25 Juli 2018

Uploaded by

Copyright:

Available Formats

Seminar for Diagnostic Measurements

and On-line Monitoring on Rotating Machines

July 24 - 25, 2018

Fabian Öttl received his MSc degree in Electrical

Seokhoon Hong is a Regional Application Specialist on

Introduction in electrical testing of rotating machines

High voltage AC-Tests

10:30 - 11:00 Coffee break

Benefits of a portable high voltage source

11:45 - 12:30 Customer Case Study

12:30 - 14:00 Lunch

Unconventional measurements on rotating machines

15:00 - 15:30 Coffee break

Partial discharge (PD) measurement

16:15 - 17:00 Customer Case Study

Advanced source separation methods

Partial discharge measurements on rotating machines

10:30 - 11:00 Coffee break

11:00 - 11:45 Customer Case Study

11:45 - 12:30 Customer Case Study

12:30 - 14:00 Lunch

Propagation of partial discharges in rotating machines

15:00 - 15:30 Coffee break

PD Monitoring system for rotating machines

Benefits from continuous PD Monitoring

Terms and Conditions

Electromagnetic induction principle (linear movement)

Seminar for Diagnostic Measurements and

OMICRON Electronics Asia Ltd. July , 2018

Generator principle Generator principle

Electromagnetic induction principle (rotational movement) Winding arrangement

© OMICRON Page 3 © OMICRON Page 4

© OMICRON electronics Asia Limited 2018 Page 1

© OMICRON Page 5 © OMICRON Page 6

Roebel bar Roebel bar

Decreased primary Iron

> Self-induced currents increase at outside of conductor

© OMICRON Page 7 © OMICRON Page 8

© OMICRON electronics Asia Limited 2018 Page 2

> Stator winding strand Facts

Outer Corona Protection,

Reduction of dielectric stress Illustration off electric field distribution

© OMICRON Page 9 © OMICRON Page 10

Inner potential grading (IPG) Main insulation

Copper 1,7 x 10²

Outer Corona Protection (OCP) Iron 0,1

Mica 1015 – 1018

© OMICRON Page 11 © OMICRON Page 12

© OMICRON electronics Asia Limited 2018 Page 3

> Epoxy resin has medium electrical

Theoretical discharge > Simulated treeing behavior under

Source: R. Vogelsang, Electrical breakdown in HV Winding Insulations of

© OMICRON Page 13 © OMICRON Page 14

Resin rich systems Vacuum Pressure Impregnation (VPI)

Resin rich technique Facts of VPI

> Through high pressure unused with conventional methods

© OMICRON Page 15 © OMICRON Page 16

© OMICRON electronics Asia Limited 2018 Page 4

Arrangement > Without OCP

End potential grading (EPG)

OMICRON Electronics Asia Ltd. July , 2018

OMICRON Electronics Asia Ltd. July , 2018