Shaft Current Protection: Tin Rabuzin
Shaft Current Protection: Tin Rabuzin
Shaft Current Protection: Tin Rabuzin
TIN RABUZIN
Degree project in
ICS
Master thesis
Stockholm, Sweden 2015
XR-EE-ICS 2015:009
Abstract
Shaft current protection in hydro and turbo generators is an important gen-
erator protection issue. Currents flowing in the generator shaft might damage
generator bearings which, in turn, could reduce operating time and cause large
financial losses. Therefore, it is important to prevent operation of the generator
under conditions of high shaft currents.
In this project, task was to develop measurement and protection system that
is able to operate under certain conditions. Measurement device has to be able
to accurately measure currents lower than 1 A in a generator shaft that can
vary in diameter from 16 cm up to 3 m. Also, those currents might appear in
frequencies equal to multiples of line frequency. Device is to be located in a
limited space and in a proximity of the generator. Thus, stray flux is expected
which might influence measurements. Furthermore, since currents that have
to be measured are low, output of a measurement device is usually a low level
signal. Such signal had to be catered for and adapted in a way that it can be
used with numerical relay.
After literature review and overview of possible solutions, Rogowski coil was
chosen as the measurement device which will be further analysed. Two other
current transformers were considered which served as a good comparison with
Rogowski coil. Several different tests and measurements were made on men-
tioned measurement devices. Also, it was investigated how IEC61850-9-2 and
Merging Unit (MU) could be used in this application. Upon this investigation,
complete protection systems were assembled in the laboratory and they were
tested.
To asses the behaviour of different systems in the real environment, test in-
stallation was built in the hydro power plant, Hallstahammar. This installation
included traditional systems, with measurement signals connected to the relayå,
and the one which utilized concepts of Process Bus and Merging Unit. Measure-
ments and tests that were made there served as a final proof of successfulness
of protection systems.
Results showed that Rogowski coil was a suitable choice for a measurement
device due to its beneficial mechanical and electrical properties. Also, tests made
with actual shaft current showed advantages of using Rogowski coil in pair with
Merging Unit and process bus over traditional protection systems. Nevertheless,
it was confirmed that both types of systems satisfy project requirements.
During my stay at ABB SA Products, I have also had a lot of help and
guidance. Sincere thank you to my supervisor, Zoran Gajić, who entrusted me
with an opportunity to work on this project and for his valuable comments and
advice. Also, special thanks to Odd Werner-Erichsen and Mohammad Khorami
who had enough patience to answer my endless questions and helped me to
carry out this work. I am also thankful to Kjell Westberg for arranging my
training and for all of his efforts regarding my work.
Last, but not the least, my deepest gratitude goes to my parents Zdravko and
Silvija Rabuzin, brothers Filip and Lovro Rabuzin, girlfriend Andrea Cvek and
grandparents Josip and Marija Hajduk. They have been my biggest emotional
support and encouragement throughout my studies. This journey would not be
possible without them.
Contents
Abstract i
Referat ii
Acknowledgements iii
Acronyms ix
1 Introduction 1
B Laboratory Set-up 65
List of Figures
CT Current Transformer
EMF Electromotive Force
EMI Electromagnetic Interference
HMI Human-Machine Interface
are then used by protection relay which, in case of excessive currents, trips the
unit. Certain issues are associated with shaft current transformers. Installation
of such large and heavy transformer in confined space, where it is usually placed,
is difficult and time consuming. Furthermore, it has to accurately measure low
level primary currents (less than 1 A) in a large conductor which is shaft. This
causes its secondary current to be at a very low level. Also, because it is
installed in the proximity of the generator, there is a possible high influence of
the stray flux from the generator on measurements. Figures 1.1 and 1.2 show
previous ABB’s solution with the shaft current transformer, ILDD, and the
shaft overcurrent relay, RARIC.
In this project, the new solution for shaft current protection had to be found.
The requirement of the solution for the new current sensor was that it has to
show better mechanical and electrical properties than the shaft current trans-
former. It shall be easy to assemble and mount. Also, since the use of numerical
relay for protection purposes is assumed, its secondary quantities must be suit-
able for its inputs. In this project, ABB’s relay REG670 2.0 was used. Appendix
A provides a reader with an overview and possible configuration of this relay.
Also, results of the investigation of relay’s measurement performance in low
current or voltage level range is presented.
In Chapter 2, literature review of possible current sensors that could be used
in this case is given. Selected sensors are then analysed and results are shown
in Chapter 3.
Chapter 4 presents complete shaft current protection systems consisting of
measurement devices, protection relay and/or amplifier. Those systems are
tested and different measurements are performed. Also, this chapter gives an
Introduction 3
2.2.1 Frosting
Frosting is the most common type of bearing damage due to shaft currents.
Damage is not noticeable to human eye. Microscopic view of the damage re-
veals craters whose bottoms are round and shiny. This indicates melting of the
material. This kind of damage where material is removed occurs during voltage
discharge. Sometimes, damage due to chemical attacks might be mistaken for a
shaft current damage. Microscopic view of frosting damage is shown in Figure
2.2.
2.2.2 Pitting
Pitting damage shown in Figure 2.2 on the left, appears similar to the frosting
damage except it is much larger in size since its source is powerful. Pitting
occurs more randomly and does not cover entire area as frosting. It is sometimes
possible to count the number of discharges.
2.2.4 Welding
Welding is caused by very high currents (hundreds of amperes). It is easily
noticed by human eye. Welded parts usually have to be split by sledgehammer
or other mechanical means.
Figure 2.2: Bearing damage - pitting, frosting and spark tracks [23] [9]
Shaft Currents and Voltages in Generators 7
Grounding Brushes
Grounding brushes can be used for practical and economical shaft grounding.
They provide low impedance path from the shaft to the ground. However, there
are some well known issues that come with using brushes such as oxidation
buildup, mechanical wear because of contact with shaft, reduced effectiveness
due to contaminants built up on metallic bristles and required maintenance [17].
Some of those issues are reflected on and improvements are suggested in [16].
Shaft Currents and Voltages in Generators 8
Grounding Rings
Grounding ring is a recent method of grounding shafts. Its advantages are pre-
sented in [14]. Grounding ring uses microfibres to establish electrical contact
with the shaft. Due to very small diameter of fibres, surface roughness does not
affect resistance between two conducting materials, unlike in the case of ground-
ing brushes. Because of this reason, there is no need to apply pressure to keep
electrical contact which results in low friction and negligible wear. Authors also
state that microfibres are resistant towards contamination. One other benefit
of using grounding ring is that even if fibres loose physical contact with the
shaft surface, ”breakdown due to local field emission will occur and reestablish
electrical contact”.
Conductive Grease
Third method of grounding the shaft is through grease in the bearings. Such
electrically conductive grease provides path for the current through the bearing
without causing arcing. It should be said that this, in comparison with previous
examples, is not shaft grounding per se. In this case shaft is grounded only if
pedestal of bearing is grounded. And then, as it was mentioned, arcing in bear-
ings is prevented. According to [17], conductive particles might cause lubricant
to be ineffective and, therefore, this method has been abandoned.
Turbine Grounding
Previously mentioned grounding methods were applicable to turbo generators.
In case of hydro-generators, water in the turbine provides good grounding for
the shaft. Therefore, no additional grounding is applied to the shaft.
Current Transformers
First obvious choice is to use current transformer to measure shaft current.
Generator shaft, as a conductor of shaft current, acts as a primary winding with
one turn. Magnetic core with secondary and test winding is mounted around the
shaft. Principle of operation is the same as of any current transformer and it is
well known. Therefore, it will not be discussed further. Depending on the shaft
size, core is built out of two or four parts. Because of its weight and difficulties
in connecting the core parts on site, current transformer with magnetic core is
not mechanically attractive.
As it was mentioned in Chapter 1, this solution was previously used by ABB.
Diagram of shaft current transformer ILDD is shown in Figure 2.3.
Main issue with ILDD was its low sensitivity at desired trip and alarm
levels. As it is described in [10], new solutions to this problem were investigated.
Among other solutions, current transformer design was improved such that it
Shaft Currents and Voltages in Generators 9
Rogowski Coil
One of the proposed solutions in [10], is Rogowski coil. Rogowski coil is named
after German physicist Walter Rogowski. It was used since the beginning of
Fig. 1. Turbine generator shaft grounding current and voltage sensing.
the 18th century. However, it has only gained popularity in the recent years
due to solutions of certain problems associated with Rogowski coil. For pro-
tection purposes it could not have been used earlier due to low power output
which could not be used as an input to electromechanical relays. However, with
numerical relays, power requirement of input devices decreased and they are
gaining popularity [26]. They also exhibit behaviour that gives them advantage
over conventional CTs [11].
Some of the benefits and drawbacks of using Rogowski coil as a measurement
device as found in literature are presented in Table 2.1.
Principle of operation has been well known and covered by the number of
papers and review is given in [24].
Following Ampere’s law it is written
I
1 −−→ → −
i(t) = B(t) · ds (2.3)
µ0
What Equation 2.3 says is that if magnetic flux density is integrated around
closed contour, this results in a current which is flowing inside closed contour.
Advantages Drawbacks
No iron core - linear, no saturation
Necessity of integration circuit
and magnetizing current
Low production cost Low voltage output
Mechanical fexibility, small size Low frequency noise
and weight from integrator circuit
Electrically safe when open Sensitivity to conductor position
Wide bandwidth Limited rejection of external fields
Non-disturbing for primary circuit Inability to measure DC currents
Can endure large overload without
damage
Ampere’s law is the reason why Rogowski coil can be flexible and mechanically
attractive, unlike, e.g. CTs.
Change in magnetic flux which is passing through turns in Rogowski coil
will, according to Faraday’s law, induce voltage. This is written in following
equation.
dΦ
u(t) = N (2.4)
dt
If laws stated in Equations 2.3 and 2.4 are combined, they result in next equation
Z
u(t) = N Ḃ ~ dA ~ = N A · µ0 · i(t)
~˙
(2.5)
s
where N is the number of turns in Rogowski coil, A is the area surrounded by
a single turn and s is the coil length. Equation 2.5 states relationship between
measured current and voltage induced in Rogowski coil.
To get the waveform of the measured current, output voltage has to be
integrated. There are different methods possible to perform integration each
of them with their own benefits and drawbacks [21]. Problems associated with
integration circuits were the reason why, in the beginning, Rogowski was not
used much.
In [7], Rogowski coil is analysed under non-ideal measurement conditions
such as non-uniform turn distribution, terminal opening, tilted conductor and
influence of external fields. All of those are applicable to the possible solution
of shaft current problem.
Nevertheless, it is shown in [18] that Rogowski coil yields accurate measure-
ments of shaft currents compared to measurement described in IEEE Standard
112-2004. Moreover, some measurements of shaft currents with Rogowski coils
were already successful [13].
Both of the above mentioned measurements utilized Rogowski coil relatively
long time ago. However, it has not been used for shaft current protection pur-
poses until now due to its low power output which was not sufficient to drive
electromechanical and solid state relays. Now, with numerical relays, Rogowski
coil can also be considered for those purposes.
CHAPTER 3
Analysis of Shaft Current Measurement Devices
Three measurement solutions that could be used to measure shaft currents were
presented in Chapter 2. Based on the possible performance and ease of instal-
lation, Rogowski coil was chosen to proceed with. In addition to Rogowski coil,
two other CTs were tested also. Purpose of testing CTs is to be able to compare
Rogowski coil with older solution and to investigate possible retrofit solutions
of ILDD with REG670 or Zelisko GWR3 with RARIC.
First, devices will be described and relevant technical data will be presented.
Then, different measurements will be shown and analysed for each of these de-
vices such as linearity measurements, frequency response, magnetization curve,
waveforms and step response. Furthermore, sensitivity test was performed for
current transformers to check their compatibility with old ABB relay, RARIC.
It is important for devices to be able to reject external fields because they will
be placed in areas affected by high stray fields from the generator. Therefore,
this ability is also checked. In the end, devices are compared and conclusions
are drawn.
x10 x1
Type LR 015/1.5
Sensitivity (mV/A) 200.0 20.0
Peak Current (A) 30.0 300.0
Noise typ. (mVRMS ) 1.0 0.5
Phase Error at 50 Hz (deg.) 7.0
Peak di/dt (kA/µs) 0.01 0.1
Low Frequency Bandwidth - fL (Hz) 4.0
High Frequency Bandwidth - fH (kHz) ≈ 40
Typical Linearity (Hz) ±0.05% (full scale)
Typical Accuracy (Hz) ±0.3%
Minimum Output Loading (kΩ) 100
Analysis of Shaft Current Measurement Devices 14
As it can be seen on the left side of Figure 3.2, this current transformer
is fairly large and bulky. It is made out of two parts which are electrically
connected by wires. Magnetic connection is shown on the right side of Figure
3.2. Two core part ends are shaped in form of a letter L and they are pressed
together with two metal plates with four screws. This way air gap is reduced and
mechanical robustness is increased. Tightening screws are made out of plastic
to avoid creating additional short circuited winding around the core.
High number of secondary winding turns results in low secondary currents.
Therefore, on it’s secondary winding, 82 Ω resistor is connected and this sec-
ondary voltage is measured. Resistance is chosen to be the same as an input
resistance of RARIC. Connecting resistance on the secondary side of the CT
increases secondary voltage. This voltage is then easier to measure than low
level currents.
Later, in Chapter 4, possibility of using secondary current as a measurement
signal will be investigated. However, in this chapter, only secondary voltage
measurement will be shown.
It also consists of two parts connected electrically with wires. Core joints in
case of ILDD 096 are not manufactured in the same way as in the case of Zelisko
GWR3. It is obvious if one looks at the joint displayed on the right side of Figure
3.3. Distance between lamination ends is much greater thus increasing influence
of air gap on magnetic performance. Also, mechanically, it is much harder to
assemble and connect two parts together. Plastic plates connected with metallic
screw are used to tighten laminations of different core parts together.
As in the case of Zelisko GWR3, due to the fact that secondary winding has
500 turns, currents are very small. Therefore, 82 Ω resistor is connected to the
secondary. Then, voltage is measured as a secondary signal.
Magnetization Curves
70
ILDD 096
Zelisko GWR3
60
50
Secondary Voltage (V)
40
30
20
10
0
0 50 100 150 200 250 300 350
Secondary Current (mA)
250
Secondary Voltage (mV)
200
150
100
50
0
0 0.5 1 1.5
Primary Current (A)
which is due to the lack of iron core. Therefore, curve given in Figure 3.5 is
applicable to all frequencies.
Judging by measurements with Zelisko GWR3, it can also be concluded
that they are linear. However, in this case, they are dependant on frequency.
Of course, as frequency of the primary current is increased, lines are steeper
which indicates better performance of the transformer.
Measurements with ILDD 096, on the other hand, are not linear at levels up
to 1 A. Around trip and alarm levels, non-linearity is very high. Transformer
exhibits similar behaviour when it comes to frequency dependency as Zelisko
GWR3.
Changes in measurements with changes in frequency can be explained by in-
creased magnetization impedance. Very simplified model of current transformer
is shown in Figure 3.7. If it is assumed that primary current of any frequency
140
120
120
100
100
80
80
60
60
40
40
20 20
0 0
0 0.5 1 1.5 0 0.5 1 1.5
Primary Current (A) Primary Current (A)
Ip0 Lm 82Ω
Im Is
is the same, Kirchhoff’s law states that this current has to be equal to the
sum of magnetizing current and secondary current. If frequency of the current
is increased, both magnetization impedance and, as a consequence, secondary
currents are increased.
This model is simplified to a great extend and cannot be used to accurately
predict behaviour of CT. Nevertheless, it serves as a good explanation of phe-
nomenon seen in Figure 3.6.
200
25
150 20
15
100
10
50
5
0 0
0 200 400 600 800 1000 0 200 400 600 800 1000
Frequency (Hz) Frequency (Hz)
data for all current levels except 1.5 A. It is also shown that phase difference is
reduced with increase of primary current level and frequency. This is even more
clear in the Figure 3.8.
In case of Zelisko GWR3 there is a change in amplitude of output voltage
with a change in frequency for the same primary current levels. Secondary
voltage is stable after approximately 200 Hz. However, higher frequencies will
not be measured and, therefore, we can expect non-linear behaviour at lower
frequencies. This is reflected in the Figure 3.6, where frequency dependency was
already noticed. When it comes to phase difference between primary current
and secondary voltage, it decreases with frequency until it is settled at around
10◦ . For instance, at 50 Hz and 1 A, secondary voltage is shifted for 28◦ .
Even worse results were obtained with ILDD 096. Secondary voltage is
settled at even higher frequencies and, at lower levels, it is very non-linear.
50
Phase Difference (deg.)
140
120
40
100
30
80
60 20
40
10
20
0 0
0 200 400 600 800 1000 0 200 400 600 800 1000
Frequency (Hz) Frequency (Hz)
120 50
100 45
80 40
60 35
40 30
20 25
0 20
0 200 400 600 800 1000 0 200 400 600 800 1000
Frequency (Hz) Frequency (Hz)
Figure 3.11: RARIC sensitivity curve with displayed results for Zelisko GWR3
limit is 130 mA and for 80 mV at resistor it is 250 mA. We can see that all
currents form Table 3.5 for Zelisko GWR3 are well bellow current limits. ILDD
096 current at 50 Hz and 40 mV at resistor is above current level limit. Other
ILDD 096 currents satisfy requirement, but they are very close to the limit.
Based on test results, it is concluded that Zelisko GWR3 passes the sensitiv-
ity test. However, ILDD 096 does not. As it was already mentioned, Rogowski
coil requires minimum output loading of 100 kΩ and, therefore, this test was
not done with it.
Figure 3.12 illustrates how this test was done for Rogowski coil. Numbers
from 1 to 15 indicate position at the circumference where external conductor
was placed. External conductor is marked with red ellipse. Conductor was not
placed in the area surrounding T-connection. According to the manufacturer
data,this area is highly sensitive.
Figure 3.13 shows how the same test was done for CTs. There is less mea-
surement points at which external conductor was placed. This is due to the fact
that no more points are needed to capture the behaviour of CTs with respect
to position of external currents, as it will be seen later.
Measurements under influence of external fields for Rogowski coil are shown
in Figure 3.14. Left part shows measurements with external current of 50 A at
different distances from circumference of Rogowski coil. Right part shows mea-
surements at 5 cm from circumference of Rogowski coil with different currents.
What can be noticed is that, as external conductor is moved further away
from T-connection starting from the node 1, influence is decreased. When com-
50 50
0 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Node Number Node Number
ing back to T-connection from the other side, external influence is increasing,
until node 15 (closest to the T-connection). At node 15, measurements are
sharply reduced. However, conductor was then moved even closer to the T-
connection (not shown in measurements) and influence was increased again.
Anomaly at node 15 (reduction of influence) cannot be explained without know-
ing details of construction of Rogowski coil, however, this is most likely due to
geometrical position of the external field and end turns at the T-connection. It
should be noted that distance between each node and between T-connection is
approximately 20 cm.
Different attempts were made to protect T-connection of Rogowski coil
against external fields. Two solutions of Faraday cage were tried out. One
of them was rigid cage made out of steel and the other one was aluminium tape.
No significant improvement was noticed.
External conductor was also placed in the plane of Rogowski coil around its
circumference. Again, as this conductor is moved closer to the T-connection
its influence on the measurement is increased. Small improvement of external
rejection was noticed with Faraday cage made of steel. Faraday cage with
aluminium foil did not show any improvement. This is due to the fact that
aluminium has lower magnetic permeability than steel. Magnetic permeability
is the parameter that influences how well Faraday cage can protect against low
frequency magnetic fields such as those studied here. However, improvement of
Faraday cage made of steel is very small. To keep the mechanical flexibility of
the solution, it is decided not to proceed with using Faraday cage to protect the
T-connection.
Another observation was made when measurements were being done. Ro-
gowski coil was much better at rejecting external fields from conductor that was
in its plane than those from conductor perpendicular to its plane. In fact, if
this conductor was not close to T-connection, no change is measurement was
noticed and, therefore, they are not shown here.
Analysis of Shaft Current Measurement Devices 24
300 300
200 200
100 100
0 0
1 2 3 4 1 2 3 4
Node Number Node Number
450
400 400
350
300
300
200 250
200
100
150
0 100
1 2 3 4 1 2 3 4
Node Number Node Number
Step Response
160
140
Voltage (mV)
120
100
80
60
40
20
0
−0.1 0 0.1 0.2 0.3 0.4 0.5 0.6
Time (s)
around steady state value in case of Rogowski coil. This variation should not
be mistaken for consequence of poor damping in the system. It is actually noise
from operational amplifier. Manufacturer states that this noise is distributed
around low frequency bandwidth, fL = 4 Hz and peak-to-peak magnitude of
this noise is equal to 6.1 mV. Attempt was made to isolate and analyse this
noise. Results shown in Figure 3.18 correspond to expectations with minor
deviation in noise amplitude and low frequency bandwidth. Peak-to-peak noise
was equal to 5 mV and frequency of harmonic with maximum amplitude is 5.64
Hz. Afterwards, frequency analysis was done for output voltage after reaching
steady state from Figure 3.17. Same low frequency noise was identified thus
confirming variations of output voltage in steady state.
15
3
10
2.5
5
Voltage (mV)
2
|V(f)|
0
1.5
−5
1
−10
0.5
−15
0
0 5 10 15 20 −20
0 2 4 6 8 10
Frequency (Hz)
Time (s)
200 1 100 1
Secondary Voltage (mV)
0 0 0 0
−200 −1 −100 −1
VRogowski VZelisko
Iinj Iinj
−300 −1.5 −150 −1.5
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (ms) Time (ms)
ILDD 096
1.5 1.5
VILDD
Iinj
1 1
Secondary Voltage (mV)
0.5 0.5
Primary Current (A)
0 0
−0.5 −0.5
−1 −1
−1.5 −1.5
0 10 20 30 40 50 60
Time (ms)
1000
Rogowski coil
900 Zelisko GWR3
ILDD 096
800
Secondary Voltage (mV)
700
600
500
400
300
200
100
0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Primary Current (mA)
200 70
Zelisko GWR3
Rogowski coil ILDD 096
180
Zelisko GWR3 60 Rogowski coil
ILDD 096
160
50
Secondary Voltage (mV)
120 40
100 30
80
20
60
10
40
20 0
0 200 400 600 800 1000 0 200 400 600 800 1000
Frequency (Hz) Frequency (Hz)
levels they are very non-linear. While amplitude response of Zelisko GWR3
settles after a certain frequency, this cannot be seen in the range of frequencies
that were applied here in case of ILDD 096. Similarly, phase response is the
best in case of Rogowski coil and the worst in case of ILDD 096. Despite non-
linearities noticed in frequency response, those will not cause significant errors
in complete protection system. As it will be shown later, relay will filter out
desired frequency and therefore, non-linarity of frequency response will not make
significant difference in performance of CT and Rogowski coil.
It should be noted that primary conductor was moved around the area cir-
culated by Rogowski coil and two other current transformers. No significant
deviations from expected measurements were detected.
To conclude, advantage of Rogowski coil is obvious in comparison to other
two current transformers. Both in terms of mechanical and electrical properties.
It is expected that it will perform well as a part of the systems discussed in
Chapter 4. Same can be expected from Zelisko GWR3, even though at the lower
level. On the other hand, ILDD 096, most likely, will not show satisfactory level
of performance. Biggest concern would be magnetization and non-linearity of
ILDD 096 at current levels and frequencies which are to be measured.
CHAPTER 4
Shaft Current Protection Systems
in all cases is largest possible voltage range (such that the maximum possible
amplification is achieved), except in the last case. There, largest possible current
range is selected.
Reason for choosing two voltage input ranges for further testing is in the
way how they are set. When using Setting 1, one has to, in addition to selecting
position of DIP switches, adjust potentiometer ”Span” and reduce the maximum
input voltage from 1 V to 250 mV. On the other hand, in case of Setting
2, there was no need to use potentiometer to adjust input range. Mentioned
potentiometer is very sensitive and it required some time to precisely set the
input voltage. Therefore, by using both Setting 1 and 2 in further measurements,
stability of amplifier’s gain was tested.
8 8
7 7
Output Voltage (V)
6 6
5 5
4 4
3 3
2 2
1 1
0 0
50 100 150 200 250 1 2 3 4 5
Input Voltage (mV) Input Current (mA)
next sections.
It can be seen that measurements are linear. Amplifier gain corresponds to
specified settings. There are some small deviations at lower input voltages (not
visible in Figure 4.1) for settings 1 and 2. However, those voltages will not be
measured in the final set-up. Also, they are not significant, compared to the
absolute measured value.
Larger inaccuracies can be seen in measurements with amplifier setting 3 in
lower range. However, that is due to imperfections of injected currents at low
values and not due to performance of the amplifier. This was concluded based
on observations of low amplitude current waveforms.
Setting 3
Output Voltage (V)
6 −6
5 −8
4 −10
3 −12
2 −14
1 1 2 3
−16 1 2 3
10 10 10 10 10 10
Frequency (Hz) Frequency (Hz)
4.2.1 System A
Connection diagram for System A is shown in Figure 4.3. It is based on Ro-
gowski coil.
Analogue current output of CMC 256-6 is used to inject primary current
that simulates real shaft current. Rogowski coil is mounted around conduc-
tor through which primary current is flowing. Double shielded co-axial cable
connects Rogowski coil with the Integrator box. Integrator box performs inte-
gration and amplification of the signal from Rogowski coil. Signal is amplified
20 times and it can be measured at the output BNC connector. Using this con-
nector, signal is further transmitted to amplifier discussed in previous section
via second single shielded co-axial cable and BNC-to-screw terminal adapter.
Output of the amplifier is, depending on the settings, connected to voltage or
current channel of the TRM module as shown. DC/DC converter is used to
transform the DC voltage from 48 V power supply to 12 V and 24 V and to
supply it to Integrator Box and amplifier.
4.2.2 System B
Second system is based on current transformers ILDD 096 and Zelisko GWR3.
Connection diagram in both cases is the same and, therefore, only one of them
is shown in Figure 4.4.
It is also mostly the same as the one in Figure 4.3. Difference is in the
way that secondary signal is connected to the amplifier. In this case, secondary
terminals S1 and S2 are connected directly to the amplifier. Also, there is no
need to supply DC voltage to the current transformer as it is the case with
Integrator
Box KNICK TRM of REG670
∫ 2
4
5
6
CH1
7 8
CH11
DC/DC
Converter
48 V
82 Ω 2 5
S2 CH1
4 6
7 8
CH11
ILDD/GWR3
DC/DC
Converter
48 V
7
200
6
5 150
4
100
3
2
50
1
0 0
0 0.5 1 1.5 0 0.5 1 1.5
Primary Current (A) Primary Current (A)
Linearity with Settings 1, 2 and 3 Linearity with Setting 3 and without amp.
7 200 8
Setting 1 Without Amplifier
Setting 2 180 Setting 4 7
6 Setting 3
160
6
Secondary Voltage (mV)
140
5
120
4
100 4
3
80
3
2 60
2
40
1 1
20
0 0 0
0 0.5 1 1.5 0 0.5 1 1.5
Primary Current (A) Primary Current (A)
age was not large enough to be registered. In this case, however, such behaviour
is not limited only to the low ranges of primary currents. Also, there was no
large difference in linearity of measurements when secondary current was used
instead of voltage. Measurements without an amplifier and secondary current
were attempted, but they were very low and unstable and thus not shown here.
Measurements for System B with ILDD 096 are shown in Figure 4.7. They
are done in a same way as those with Zelisko GWR3. Again, non-linearity
discussed in Chapter 3 is noticeable in the low ranges of primary current. Reso-
lution issues are, also, present without an amplifier. It should be mentioned that
non-linear behaviour in range of primary currents lower than 0.5 A is captured
well with voltage inputs. In case of Setting 4, where current is used as a sec-
ondary signal, resolution of those measurements is low. However, it is improved
at values larger than 0.5 A.
In the end, quantization of linearity was attempted for different systems and
Shaft Current Protection Systems 35
Linearity with Settings 1,2 and 3 Linearity with Setting 3 and without amp.
6 8
Setting 1 140 Without Amplifier
Setting 2 Setting 4
7
5 Setting 3
120
6
4 100
5
80
3 4
60
3
2
40 2
1
20 1
0 0 0
0 0.5 1 1.5 0 0.5 1 1.5
Primary Current (A) Primary Current (A)
Coefficient R2 tells us how much variance in data (in percentage) linear ap-
proximation predicts in measured data. Therefore, it can be used to compare
linearity of different systems and amplifier settings. Results are shown in Table
4.3. Column ”Set 0” corresponds to measurements without an amplifier.
What can be concluded with certainty is that System B with ILDD 096 is
less linear than both System A and System B with Zelisko GWR3. Also, we can
see that current measurements are less linear than voltage measurements with
ILDD. There are very small deviations of coefficient of determination in System
A and System B with Zelisko GWR3 with all amplifier settings. Furthermore,
calculated value of R2 is very close to 1, which suggest that measurements are
very linear.
1 1
Primary trip current (A)
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 1 2 3 4 0 1 2 3 4
Setting number Setting number
4
Primary Current (A)
−2
−4
0 5 10 15 20 25 30 35 40 45 50
Time (ms)
made but the ones shown here represent findings from all of them. Cells marked
yellow are those with measurements that have error greater than 50 mA and
cells marked red are those with error greater than 100 mA.
Behaviour of systems, when it comes to filtering tests, is clear. At higher
magnitudes of primary currents, all systems gave measurements with errors
within reasonable limits except System B with ILDD and amplifier setting 4.
When primary current is decreased to 0.5 A, results are still within limits for
System A and System B with Zelisko GWR3. However, measurements at 0.5
A of System B with ILDD 096 have high errors. Errors go up to 45 % of
actual primary current. Poor performance of ILDD 096 can be associated with
non-linearities shown in 3.6 and distortions of waveforms shown in 3.20.
15 Voltage Waveform
Filtered RMS Voltage
10
5
Voltage (V)
−5
−10
−15
0 10 20 30 40 50 60 70 80 90 100
Time (ms)
higher than the maximum input value, amplifier limits the output voltage to
the maximum possible value. This results in, of course, limited filtered RMS
voltage level. However, this voltage is high enough for protection function to
recognize the fault. It should also be mentioned that such high currents, as they
are assumed here, are not expected in reality.
4.7.1 Overview
The topic of IEC61850 standard is substation automation based on modern nu-
merical relays exploiting benefits of communication networks within substation.
Communication network of a modern substation, according to IEC 61850, is split
into three levels. Those are process, bay and station levels shown in Figure 4.12.
Part 9-2 of above mentioned standard defines lower part of the communication
system hierarchy within modern substation, process level. What is also defined
in Part 9-2 is process bus. Process bus is used for ”communication between
high voltage process and components interacting with it” [4]. Since process bus
will be analysed as a part of the solution to the problem this project deals with,
only Part 9-2 will be considered in the following discussion.
To begin with, essential device for existence of process bus will be discussed,
merging unit. It is used as an interface between conventional or non-conventional
CTs or VTs and IEDs, just as it is shown in Figure 4.12. Signals from mentioned
sensors are digitalized and synchronized and then sent across process bus to bay
level IEDs. In that sense, part of merging unit has similar role as TRM and
ADM of REG670 discussed in Appendix A. Difference now is that, instead of
data ”travelling” through backplane module of IED, it travels through Ethernet
link which is part of the process bus. Data sampling and protection functions
are now located in different devices. Sampled Values according to IEC61850 are
primary quantities and they are sent using multicast from the merging unit to
all IEDs [4].
Since merging units do not have to perform sampling at the same time as
all other merging units in substation, they have to be synchronized. Synchro-
nization will not be discussed here since it is not important for this particular
Introduction to process bus
application. There is only one voltage signal from measurement device that is
used for protection purposes. Therefore, phase shift of this signal, with respect
What is process bus?
to some other signal, is not important.
Process level MU MU MU MU MU MU
NCIT NCIT
4.7.2 Implementation
In this project, Rogowski coil can be considered to be non-conventional trans-
former. Therefore, merging unit could be used as an interface between Rogowski
coil and REG670 2.0. First, capability of REG670 2.0 will be investigated in
order to check if the same application can be used with Sampled Values as the
one described in Appendix A.
For this reason, system depicted in Figure 4.13 was assembled and config-
ured. It consists of OMICRON CMC256-6 test set, industrial Ethernet switch
and REG670 2.0. Both test set and relay are connected to the switch via optical
fibres.
Test set has option to behave as a MU. Sampled Values are generated by the
test set according to [1]. Since [1] refers to the subset of IEC61850-9-2, this kind
of Sampled Values is called 9-2 Light Edition (LE). Multicast SV are mapped
to the secondary voltage outputs of the test set and current output group A.
Therefore, all physically injected currents and voltages at the test set are also
scaled to the primary values according to the user settings and multicast across
process bus.
Figure 4.14 shows parameters for SV configuration of test set. First pa-
rameter, Sampled Value ID, specifies ID of the merging unit which sends the
sampled value. Since, as described earlier, SV are sent as multicast, multicast
MAC Address is specified in the second parameter. According to [1], Applica-
tion Identifier (APPID) always has to be set to 16384 (0x4000). IEC 61850 can
utilize virtual LAN (VLAN) according to IEEE 802.1q. This allows traffic seg-
regation and prioritization [28]. Therefore identifier of virtual LAN and priority
of data was entered as shown. According to IEC 61850-7-3, quality of data can
be specified. Different values of quality tag and their meaning can be found
in [12]. If the value is set to 0, validity of the data is assumed to be ”good”. In
the end, simulation flag can be used to indicate that SVs come from test device.
Upon applying configuration in 4.14 to test device, SV frames were captured
by Wireshark. One of the frames is shown in Figure 4.15.
It can be seen that frame structure corresponds to the one described by
[1]. Therefore, it is expected that such SV stream will be compatible with
REG670 2.0. 64 bytes under section seqData is actually data corresponding to
4 voltage and 4 current channels. At the time of capture, it can be seen that the
instantaneous value of current is equal to 1145 mA (0x0479). This information
is found in first four bytes of seqData.
Next, REG670 2.0 was reconfigured. Merging unit was added in Hardware
Configuration of the IED as an additional hardware card. Following this hard-
ware configuration, only Sampled Value ID was changed to be the same as the
one set at the test set. Now, merging unit channel can be used just like a TRM
channels.
This is shown in Figure 4.16. Rest of the application is configured as shown
in Appendix A. Only difference is that now, instead of using signals from the
TRM, one can use signals from the Merging Unit. Therefore, no additional
change is shown here.
Finally, Sampled Values streamed from the test set were display on the HMI
SMAI2_SHAFT_3P
SMAI2 SMAIHPAC
SHAFT_CURRENT BLOCK G2AI3P G3P AI3P SMAIHPAC2_SHAFT_I3P
REVROT G2AI1 SMAI2_SHAFT_IL1 BLOCK AI1 SMAIHPAC2_SHAFT_IL1
GRP2L1 G2AI2 AI2
GRP2L2 G2AI3 AI3
TRM_40.CH1(I) GRP2L3 G2AI4 AI4
GRP2N G2N O:3|T:8|I:2
Shaft Current Protection
O:1|T:8|I:14 Systems 43
C
E
4.8 Summary
Two systems with three different measurement devices were described in this
chapter.
1
Also, for all of those2
and for various amplifier 3
settings, linearity
4
mea- 5
surements, ramping and filtering Project
test
Thesis Project
were performed. What can be concluded
Responsible department
ABB Ltd.
Technical reference
is that both Systems ReplacingA and System B with Zelisko GWR3 showed satisfactory Created by
Rev. Modification
performance.
Rel. date
However,
Created by
System
Based on
Thesis BProject.Substation.Voltage
with ILDD 096 is not recommended to use
Level.Bay Approved by
detail.
First, Rogowski coil (model LFR/015/1.5/1690) was put on top of existing ILDD
and strapped with cable strips which is shown in Figure 5.2. Also, additional
conductor shown on the right side of Figure 5.2 was placed between shaft and
ILDD and Rogowski coil. This conductor was used to inject primary currents
for ramping and linearity tests shown later.
In addition to cables connecting RARIC and secondary of ILDD, two other
cables were connected in parallel so that the performance of ILDD can be ob-
served. This can also be seen on Figure 5.2.
Intentional
A Shaft Gorunding
Bearing
Rotor
Coaxial cable
Rogowski coil
Injection wire
RARIC
Turbine
Since injection current cannot accurately represent shaft current, shaft was
grounded intentionally while generator was in service. In between the shaft
and the grounding point there was an amperemeter installed. More detailed
description will be given in the following section.
The rest of measurement and protection equipment was placed on the outside
of the compartment where measurement devices were located. Overview of
this equipment is shown in Figure 5.3. Secondary of ILDD 052 was connected
to the amplifier, in parallel with RARIC. Output of the amplifier was then
connected to the voltage channel on the TRM. Similarly, output from integrator
box of Rogowski coil was connected to second amplifier. Output from the second
amplifier was also connected to the voltage channel on TMR of REG670. In
parallel, the same voltage from Rogowski coil was connected to the Merging
Unit which digitalized signal and sent it via optic fibers to Optical Ethernet
Module on REG670 (see Figure 5.3). It should be mentioned that for some
measurements, amplifiers shown here were bypassed and signal was connected
directly to the TRM.
What is not shown on Figure 5.3 is OMICRON test set. Its current outputs
were connected to injection wire shown in Figure 5.1. Also, its analog voltage
inputs were used to record different analog signals in this system.
Furthermore, changes were also made on the RARIC relay in the control
room. First of all, trip outputs from this relay were blocked so as to be able
to test performance of both existing and new installation without tripping the
unit. Moreover, to be able to compare performance of RARIC and REG670,
trip outputs were redirected to additional numerical relay. This relay was used
only for recording of secondary voltage from ILDD on RARIC inputs and trip
outputs. This way, one can use ILDD voltage as a reference signal when com-
paring recordings done by REG670 and additional numerical relay in the control
Secondary of ILDD
from Figure 7.1
Merging
Unit
room.
Lastly, for the reader be able to understand figures shown in following sec-
tions, Table 5.2 is shown. It summarizes trip and alarm values that were set for
REG670 with different devices.
Primary Current Waveform − Rogowski coil Amplitude Spectrum of Primary Current − Rogowski coil
4 1.6
3 1.4
2
1.2
Primary Current (A)
1
1
|I(f)| (A)
0
0.8
−1
0.6
−2
0.4
−3
−4 0.2
−5 0
0 5 10 15 20 25 30 0 200 400 600 800
Time (ms) Frequency (Hz)
Secondary voltage Waveform − ILDD Amplitude Spectrum of Secondary Voltage − ILDD 096
250
800
600
200
400
Secondary Voltage (mV)
200 150
|V(f)| (mV)
−200 100
−400
50
−600
−800
0
0 5 10 15 20 25 30 0 200 400 600 800
Time (ms) Frequency (Hz)
measured with amperemeter which served as a confirmation that the shaft was
actually grounded. Information on number of stator segments was not available
which, unfortunately, results in inability to check the accuracy of the theory
presented in Chapter 2 and [3].
Figures also show that ratios between different frequency components of
shaft current in case of Rogowski coil and ILDD 096 are not the same. This
is, as it was mentioned before, due to the fact that secondary voltage of CTs
depends on frequency of primary current.
It was also observed how the relay reacts to the actual short-circuit cur-
rent. Unfortunately, behaviour of the part of the installation that incorporates
Merging Unit could not be observed. SMAIHPAC filter had temporary issues
which caused its inability to filter the SV data coming from the MU. For this
reason, this part of the installation could not have been used for testing with
shaft-current (150 Hz). However, its behaviour without the filter was observed
Hydro Power Plant Tests 49
500
V (mV)
−500 Rogowski
ILDD 052
0 2 4 6 8 10
Time (s)
Filtered Secondary Voltage
400
300
V (mV)
200
100 Rogowski
ILDD 052
0
0 2 4 6 8 10
Time (s)
ILDD 052 Trip Signal
1
0.5
0
0 2 4 6 8 10
Rogowski Trip Signal
1
0.5
0
0 2 4 6 8 10
RARIC Trip Signal
1
0.5
0
0 2 4 6 8 10
150 6
100 4
50 2
0 0
0.2 0.4 0.6 0.8 1 1.2 0.2 0.4 0.6 0.8 1 1.2
Primary Current (A) Primary Current (mA)
1 1
Primary trip current (A)
0.6 0.6
0.4 0.4
0.2 0.2
0 0
1 2 3 4 1 2 3 4
Attempt number Attempt number
from SMAI preprocessing block whose filter length is much shorter than the one
from SMAIHPAC filter. For this reason, protection function is much faster. It
is, however, expected that primary trip currents would have been higher in case
where SMAIHPAC would have been used (with ∆t = 200 ms). Nevertheless,
advantage of using merging unit over other set-ups will be shown in the following
section.
0 5
V (mV)
−200 0
0 20 40 60 80 100
Time (ms) −5
Voltage waveform form Rogowski without an amplifier
−10
200 0 20 40 60 80 100
V (mV)
Time (ms)
0
Voltage waveform form ILDD 096 with an amplifier
−200 10
0 20 40 60 80 100
Time (ms) 5
Voltage waveform form ILDD 096 without an amplifier
V (mV)
200 0
V (mV)
0
−5
−200
0 20 40 60 80 100 −10
0 20 40 60 80 100
Time (ms)
Time (ms)
14
20
Secondary Voltage (mV)
12
10
|V(f)| (mV)
10
0
8
−10
6
−20
4
−30 2
−40 0
0 50 100 150 200 250 300 0 50 100 150 200
Time (ms) Frequency (Hz)
In this chapter conclusions based on the work presented in previous chapters will
be given. Also, this solution cannot be considered as a final product. Therefore,
following conclusions, future work that needs to be done will be stated.
6.1 Conclusions
To start with, assumption that Rogowski coil was suitable choice for measure-
ment device was investigated throughout this project.
In terms of its mechanical properties, Rogowski coil outweighed CTs used
in this project but, also, some other possible solution presented in Chapter 2.
Similarly, in terms of electromagnetic properties it was proven that it is more
suitable for this application. Its secondary voltage output after integrator is
higher than those of CTs. Also, lack of magnetic core results in flat frequency
responses and linearity on the whole measurement range considered. Low fre-
quency noise was detected with Rogowski coil. However, since it is very low and
it is not at the frequencies that are to be measured, it does not affect overall
performance. Some concerns were raised regarding its rejection ability when
external fields exist around T-connection. Those concerns were disputed by
measurements made at Hallstahammar. It was shown that stray flux in this
particular power plant was not high enough to cause errors in measurements.
Three different CTs were also investigated. They are mechanically much
less attractive than Rogowski coil. The weight of CT makes it hard to mount
it around the shaft in confined and dark space, as it was encountered in Hall-
stahammar. Also, they are made out of 2 or more parts. Those core parts of
the CTs have to be accurately connected and pressed together to reduce the
air gap as much as possible which. This might be, depending on the design,
more or less difficult (see Figures 3.2 and 3.3). Also, voltage outputs of those
transformers are lower than those of Rogowski coil. Since they do have magnetic
core, measurements are dependant on frequency of primary current.
Nevertheless, Zelisko GWR3 and older ILDD 052 were proven to be suitable
to use as a measurement device. ILDD 052 can be considered as a part of retrofit
installation with REG670. Also, Zelisko GWR3 could be used with old ABB’s
relay RARIC and REG670. This was not the case for ILDD 096 because of its
low voltage output and reduced magnetic performance.
Conclusions and Future Work 54
For the purpose of generator shaft protection, numerical relay REG670 Version
2.0 made by ABB was used [19]. Following text will be dedicated to its hardware
configuration, application configuration and ability to measure low level signals.
both equipped with LDCM modules. GPS time synchronization module (GTM)
could also be found on one of the ADM slots.
On the front of this IED, Local Human-Machine Interface (LHMI) could
be found. LHMI consists of display, various buttons and LED indicators and
Ethernet communication port.
SMAI1_SHAFT_3P
SMAI1
SMAIHPAC
BLOCK SPFCOUT
SHAFT_VT DFTSPFC G1AI3P G3P AI3P SMAIHPAC_SHAFT_3P
REVROT G1AI1 SMAI1_SHAFT_UL1 BLOCK AI1 SMAIHPAC_SHAFT_UL1
GRP1L1 G1AI2 AI2
GRP1L2 G1AI3 AI3
TRM_40.CH11(U) GRP1L3 G1AI4 AI4
GRP1N G1N O:2|T:8|I:1
O:1|T:8|I:13
CURRENT INPUT
SMAI2_SHAFT_3P
SMAI2 SMAIHPAC
SHAFT_CURRENT BLOCK G2AI3P G3P AI3P SMAIHPAC2_SHAFT_I3P
REVROT G2AI1 SMAI2_SHAFT_IL1 BLOCK AI1 SMAIHPAC2_SHAFT_IL1
GRP2L1 G2AI2 AI2
GRP2L2 G2AI3 AI3
TRM_40.CH1(I) GRP2L3 G2AI4 AI4
GRP2N G2N O:3|T:8|I:2
O:1|T:8|I:14
way of connecting inputs of the TRM other than Ph-N. Parameter SetFrequency
can be used to specify which frequency component one wants to extract from
the input signal.
FreqBandWidth can be used to set the bandwidth around the desired fre-
quency in which frequency components will be looked for. It is important to
mention that filter has its own natural bandwidth which depends on Filter-
Length.
FilterLength sets the time it will take for the filter to store all samples
needed for calculation. As the filter length is increased, natural bandwidth
is decreased. In the case of 1.0 s of filter length and 0 Hz of set frequency
bandwidth as it is here, total frequency bandwidth of the filter will consist only
of natural frequency bandwidth and that is ±3 Hz. If there are more than one
frequency components within the bandwidth, filter will choose the one with the
highest magnitude. However, if there is no clear highest magnitude, filter will
return zero for the magnitude and phase angle. Signal frequency will have value
equal to minus one in that case.
Parameter OverLap defines how often the calculations will be performed.
What will happen in the case of 20 % overlap is as follows. Frames of data
corresponding to filter length of 1.0 s will be filled with data. Last 20 % of the
previous data will be taken into the new frame. Therefore, two frames will be
overlapped. When the new frame is full, calculations will be done again and
this way number of times that the filter performs calculation per filter length is
increased.
GRP1_LE
SRMEMORY TIMERSET
ROG_TOV2 HM1L01R
SR_NOUT SET OUT INPUT ON TIMER ROG_TOV1 HM1L01Y
TIMER RESET NOUT OFF HM1L01G
O:914|T:3|I:1 O:911|T:3|I:1 O:5600|T:3|I:
SR_NOUT
GRP1_LE
ILDD_OV2 HM1L02R
ABB REG670 - Generator Protection Relay 60 ILDD_OV1 HM1L02Y
HM1L02G
B O:5600|T:3|I:
Binary Outputs
BOM_4.BO1
SMBO
SHAFT_TOV1 BO1 BO1
SHAFT_TOC1 BO2 BO2
BO3 BO3
BO4 BO4 BOM_4.BO2
BO5 BO5
BO6 BO6 GRP1_LED1
BO7 BO7
BO8 BO8 DISTREC_FULL HM1L14R
BO9 BO9 HM1L14Y
BO10 BO10 HM1L14G
C O:6300|T:3|I:1 O:5600|T:3|I:1
for any trip from protection functions, corresponding contact on the BOM will
close. D
Main Application: LO
variables. Variables that are shown are filtered and unfiltered voltage output
from measurement devices, frequency obtained from the filtered signal, primary
current calculated from filtered voltage, filtered secondary current from mea-
surement devices, and phase angles of current and voltage signals. Display is
shown in Figure A.7.
2 3 4 5 6 7 8
D
Figure A.7: Display on LHMI
35 400
SMAI Output SMAI Output
SMAIHPAC Output SMAIHPAC Output
30 350
300
25
Noise Voltage (mV)
250
20
200
15
150
10
100
5 50
0 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)
Set-up for the linearity and accuracy measurements is shown in Figure A.9.
All signals were generated by high accuracy unit - OMICRON CMC 256-6.
Voltage and current outputs of CMC 256-6 were directly connected to current
and voltage inputs on the TRM of REG670. They were Channel 1 and Channel
11, respectively. Measurements were recorded from the graphical display on the
relay. They are shown in the Figure A.10.
On the left side of the figure, it can be seen that the measurements are linear.
Also, frequency of the signal does not affect measurements shown. Right side of
the figures shows deviations of the measurements from the actual values injected.
No measurement deviates more than 10 mV from the actual injected value at all
four frequencies tested. Attempt was made to measure linearity and accuracy
for the low current levels as well. However, measurements were not stable.
After examination of the waveform of the injected current, it was noticed that
the waveforms had glitches and irregularities. Therefore, measurements with
low level currents were not done.
What can be concluded is that the REG670 provides accurate and linear
measurements at the low level voltages. It is expected that it will perform well
with different measurement devices with secondary voltage outputs. However, it
was noticed that at voltages below lowest nominal voltage, measurements were
less accurate and stable. Even though it was not possible to test performance at
low level currents, attempt of using secondary currents of current transformers
as output signals is shown in Chapter 4.
OMICRON CMC 256‐6
Current Output
Voltage Output
4.5
10
4
3.5
Measured Voltage (V)
5
Voltage Error (mV)
2.5 0
2
−5
1.5
1 50 Hz 50 Hz
−10
60 Hz 60 Hz
0.5 150 Hz 150 Hz
180 Hz 180 Hz
0 −15
0 1 2 3 4 5 0 1 2 3 4 5
Input Voltage (V) Input Voltage (V)
Physical arrangement which was built in the lab to test measurement devices
and systems A and B discussed in Chapters 3 and 4 is shown in Figures B.1
and B.2.
[13] P.J. Link. Minimizing electric bearing currents in adjustable speed drive
systems. In Pulp and Paper Industry Technical Conference, 1998. Confer-
ence Record of 1998 Annual, pages 181–195, June 1998.
[14] A. Muetze and H.W. Oh. Design aspects of conductive microfiber rings for
shaft-grounding purposes. Industry Applications, IEEE Transactions on,
44(6):1749–1757, Nov 2008.
[15] P.I. Nippes. Understanding shaft voltage and grounding currents of turbine
generators. Magnetic Products and Services, Inc.
[25] R.F. Schiferl and M.J. Melfi. Bearing current remediation options. Industry
Applications Magazine, IEEE, 10(4):40–50, July 2004.
[26] V. Skendzic and B. Hughes. Using rogowski coils inside protective relays. In
Protective Relay Engineers, 2013 66th Annual Conference for, pages 1–10,
April 2013.
[27] SKF. Electrically insulated bearings from skf. ”http://www.skf.com/
binary/98-111303/Motor-Insocoat-hybrid-6160_EN.pdf”. Accessed:
2015-04-20.
Bibliography 69