1
Detection of generator bearing inner race creep by
means of vibration and temperature analysis
Georgios A. Skrimpas, Ivaylo G. Dragiev, Reynir Hilmisson, Christian W. Sweeney,
Bogi B. Jensen, Nenad Mijatovic, Joachim Holbøll
Abstract—Vibration and temperature analysis are the two
dominating condition monitoring techniques applied to fault
detection of bearing failures in wind turbine generators. Relative
movement between the bearing inner ring and generator axle
is one of the most severe failure modes in terms of secondary
damages and development. Detection of bearing creep can be
achieved reliably based on continuous trending of the amplitude
of vibration running speed harmonic and temperature absolute
values. In order to decrease the number of condition indicators which need to be assessed, it is proposed to exploit a
weighted average descriptor calculated based on the 3rd up
to 6th harmonic orders. Two cases of different bearing creep
severity are presented, showing the consistency of the combined
vibration and temperature data utilization. In general, vibration
monitoring reveals early signs of abnormality several months
prior to any permanent temperature increase, depending on the
fault development.
Index Terms—Condition monitoring, vibration analysis, angular resampling, rotational looseness, bearing creep
I. I NTRODUCTION
Wind energy has seen continuous development over the past
two decades reaching 320GW of installed cumulative global
capacity in 2013 [1]. The vast majority of the newly designed
wind turbines operate under varying speed conditions. A
typical variable speed wind turbine drive train consists of one
or two main bearings, a multistage gearbox and a fast rotating
generator. In order to overcome the faults related to gearboxes,
the concept of direct drive turbines has been adopted by
numerous manufacturers [2]. Direct drive wind turbines are
usually equipped with permanent magnet (PM) or synchronous
generators, whereas single or double fed induction and PM
generators, are installed in geared systems.
The availability and reliability of wind power systems are
essential parameters for their competitiveness compared to
conventional energy sources. Reports presenting reliability
data from various wind turbine types provide an insight in
the component failure occurrence rate and severity [3], [4].
Based on this statistical data, generator faults correspond to
approximately 5% of the total number of failures. However,
G. A. Skrimpas is with the Remote Monitoring Group, Brüel and Kjær
Vibro A/S, 2850 Nærum, Denmark and the Centre of Electric Power and Energy, Department of Electrical Engineering, Technical University of Denmark,
2800 Lyngby, Denmark
I. G. Dragiev, R. Hilmisson and C. W. Sweeney are with the Remote
Monitoring Group, Brüel and Kjær Vibro A/S, 2850 Nærum, Denmark
B. B. Jensen is with the Department of Science and Technology, University
of the Faroe Islands, 100 Tórshavn, Faroe Islands
N. Mijatovic and J. Holbøll are with the Centre of Electric Power and Energy, Department of Electrical Engineering, Technical University of Denmark,
2800 Lyngby, Denmark
there is not any differentiation between bearing defects or
electric failures, although it is expected that bearing related
issues are more frequent especially in fast rotating machines.
Bearing faults can be generally divided into two categories:
a) rotor dynamic failures and b) bearing subcomponent defects. Rotational and structural looseness, tilted inner race on
the shaft or outer race on the bearing housing are typical
failure modes of the first category [5], whereas all bearing
subcomponents, namely inner race, outer race, ball and cage
are subjected to faults of varying severity [6]. Monitoring
of bearing vibrations reveal additional failure mechanisms,
such as misalignment between gearbox and generator, rotor
imbalance and electric erosion, as well as excessive vibrations
generated by electric causes, such as open rotor phases in
double fed induction generators [7].
Bearing inner race creep is assessed to be one of the most
severe failure modes from an escalation and consequential
damages standpoint. It is defined as the relative movement of
the inner race relative to the shaft provoked due to insufficient
fit or deformation of bearing and shaft [8]. A characteristic
secondary damage is excessive shaft wear above tolerance,
resulting in generator replacement or machining of the shaft
and placing a sleeve. Both solutions generate substantial
downtime when a new generator is not in stock or the proper
equipment and personnel are not available respectively.
There is very limited literature in regards to detection of
bearing creep and its severity assessment. In [9], tapered
roller bearing cone creep is analysed employing temperature
readings and vibration measurements. The development rate is
divided into three stages based on the relative speed between
the axle and inner race. The speed is less than 0.25rpm in
stage I and reaches gradually 8rpm in stage II, where the
axle outer diameter is decreased by approximately 0.4mm. A
slight temperature increase is observed in stage II, whereas
both stages show stable vibration levels. The axle wear and
creep speed are 1.5mm and 23rpm at the end of stage III.
Both temperature and vibration levels present erratic behaviour
reaching 150◦ C above ambient and 25Gs respectively. However, the spectral bandwidth is not specified and the exact
vibration pattern is not presented.
The present paper deals with the detection of generator
bearing inner race creep by means of vibration and temperature
analysis. A condition indicator based on the weighted average
of the 3rd to 6th running speed harmonics is proposed towards efficient monitoring of rotational looseness. Exponential
moving average is employed for temperature level assessment.
Two study cases are presented illustrating different severity
2
levels along with the corresponding frequency spectra and
temperature trends.
The structure of the paper is as follows. Section II introduces
a general concept in condition monitoring of generator bearings. The mechanism of bearing creep and its development
are described in section III. The tools utilized towards the
detection of inner race bearing creep are presented in section
IV. Fault detection of two different bearing creep severity
cases is shown in section V. Finally, section VI presents the
conclusions of this work.
II. G ENERATOR B EARING C ONDITION M ONITORING
Vibration analysis has been the most wide-spread condition
monitoring technique applied on wind turbines generators
employing accelerometers installed radially at the load zone,
as shown in Fig. 1 [10]. Tracking of speed related spectral
components describing the shaft dynamics, such as the first
and higher orders running speeds, is usually implemented
in condition monitoring systems along with broadband measurements in various frequency ranges for overall vibration
evaluation and early stage bearing defects [7]. Alternatively,
an envelope can be set over the considered healthy vibration
signature, where an alarm is triggered when a frequency
component exceeds the above mentioned predefined limit.
deformation of the two components under radial load [8], [16].
Furthermore, if the bearing is manufactured with normal level
of hardening, permanent growth can occur in the rings if they
are subjected to high temperature levels for certain period
of time. The second mechanism is encountered in cases of
interference fit. Creep is generated when the shearing stress
on the fitting surfaces exceeds the tangential frictional force
due to contact pressure [8]. In the two above mechanisms,
the inner ring lags with respect to the shaft. In [16], a third
bearing creep mechanism is proposed due to travelling waves
generated by the forces acting on the ring surface when rolling
elements cross over it. It is noted that in this mechanism the
creep direction coincides with the ring rotational direction.
Fig. 2 depicts a ball bearing where c is the clearance
between shaft and inner ring. In addition, the shaft, inner ring
and cage speeds are denoted as nshaf t , nir and nc respectively.
Fig. 2
Fig. 1: Positioning of accelerometers on generator bearings.
The accelerometers are installed at the load zone for more
efficient and consistent fault detection.
In addition to vibration based monitoring, temperature sensors are placed on the bearings’ housings in order to assess the
condition of the bearings from a thermal standpoint. Due to
the location of the temperature sensors, the measured values
reflect approximately the outer ring temperature, when in fact
a temperature difference in the range of 10◦ C up to 40◦ C is
expected between the outer and inner rings. Other techniques
applicable on generator condition monitoring, such as current
signature analysis [11] - [14] or utilization of thermal imaging
cameras [15], have been proposed, but with limited field
applications.
III. B EARING C REEP
A. Bearing Creep Mechanisms
There are three mechanisms mentioned in literature causing
bearing creep. The first mechanism is induced due to increased
clearance between the axle and the inner ring as result of
The theoretical calculated inner ring linear displacement
per one full shaft revolution is equal to πc [16]. Assuming
inner ring inner diameter equal to Di = 170mm, clearance
c = 1.5mm and speed nshaf t = 1620rpm, the angle of
displacement of the inner ring relative to the shaft per shaft
revolution is calculated as:
φ
Di
= πc ⇒ φ ≈ 0.055 rad
2
(1)
The creep speed, which is equal to the difference between
the shaft and inner ring speeds, is then computed as follows.
ncreep =
φ
Tshaf t
≈ 14.3 rpm
(2)
where φ is the creep angle over one full shaft revolution
and Tshaf t is the period corresponding to shaft speed equal
to nshaf t = 1620rpm.
The above calculated speed value is based on theoretical
considerations of the creep mechanism; nonetheless it is
assessed as conservative from a practical position. In [9], creep
speed reached 23rpm in case of tapered roller bearing under
1.4mm clearance for a shaft spinning at 1852rpm and wheel
diameter of 860mm. Although a direct comparison between
ball and cylindrical bearings should be avoided, it is expected
that the actual creep speed is relatively higher compared to
3
the theoretical one, at least at the stage where the shaft outer
diameter wear encounters for the clearance.
B. Bearing Creep Development and Fault Diagnosis Techniques
Bearings in wind turbine generators are installed applying
interference fit between the shaft and the inner ring, implying
low clearance, whereas the fit between the outer ring and
the housing is referred as loose. Furthermore, internal radial
bearing clearances are also present as shown in Fig. 3. The
aforementioned clearances are considered to be unequal, i.e.
ai 6= bj 6= ci for i = 1, 2 and j = 1, . . . , 4. It should be
emphasized that any of the depicted clearances might also
vary depending on the angle of measurement. For example,
clearances c1 and c2 could have different values if measured
horizontally and vertically.
temperature spikes. The motion is expected to be intermittent
depending on the operating condition and it can last for a
few weeks up to few months. The detection of this stage is
of essential importance in terms of minimizing any secondary
damages. If the bearing remains in operation, the condition is
irreversible and both shaft and inner ring suffer from excessive
wear, reaching up to few millimetres. The effects of this
stage include increased heat generation, potentially extreme
vibrations, lubrication deficiency, misalignment and finally
generation of smoke due to rubbing surfaces.
IV. A NALYSIS BASED ON AVERAGED V IBRATION AND
T EMPERATURE VALUES
Bearing creep is manifested in vibration spectra as increasing high order running speed harmonics [17]. Tracking
these spectral components along with continuous temperature
monitoring offers a sound representation of the bearing status.
The presence of harmonics is unavoidable in the vast majority
of applications due to inherent mass imbalance, misalignment,
poor installation and hardware imperfections. Furthermore,
both temperature and vibration trends may show erratic behaviour depending on the operational condition and speed
fluctuations of a wind turbine.
A. Averaging of Vibration Data
Fig. 3: Fitting and internal bearing clearances
The shaft, bearings and bearing housings are usually manufactured from different materials, such as steel and cast
iron, thus their thermal expansion rates are not equal. Due to
varying operational and thermal conditions, speed fluctuations,
installation issues, lubrication efficiency and bearing type, the
clearances between the above components vary over time.
Focusing on the geometrical differences of the inner ring and
shaft, it can be concluded that the inner ring expands relatively
faster. If the inner ring is subjected to numerous temperature
peaks and depending on the manufacturing process, the expansion can become permanent.
Although the dynamics and development of bearing creep
under normal operation have not been studied in depth, it has
been proposed that the mechanism follows various severity
stages which are detectable via vibration analysis and temperature readings [9]. During the initial stage of the phenomenon,
microscopic movement of the ring relative to the shaft takes
place resulting in fretting of the shaft and inner ring fitting
surface. From a vibration point of view, this condition is
usually characterized by slowly increasing or erratic vibration
levels. On the contrary, temperature peaks may be present at
this stage, however there is not any permanent rise. The degree
of clearance due to shaft wear and inner ring expansion determines the stage where the relative displacement transforms
form microscopic to macroscopic. At this stage the ring spins
slowly on the shaft generating higher vibrations and frequent
A typical vibration spectrum of a bearing subjected to
rotational looseness is characterized by increased 3rd order
and higher running speed harmonics. In order to minimize the
number of condition indicators established towards monitoring
of looseness patterns, it is proposed to compute a weighted
averaged value of the 3rd up to 6th order speed harmonics.
The weighted arithmetic mean of a data set D =
x1 , x2 , . . . , xN with corresponding non-negative weights
w1 , w2 , . . . , wN is
PN
x̄ = Pi=1
N
w i xi
i=1
By normalizing the weights so as
simplified to
x̄ =
N
X
(3)
wi
PN
i=1
wi0 xi
wi0 = 1, Eq. 3 is
(4)
i=1
PN
where wi0 = wi / i=1 wi for i = 1, . . . , n.
The efficiency of the above method is highly dependant on
the proper weight selection. In this work, the assignment of
weights relies on the argument that high weights should be
assigned to lower orders, as the harmonic order is assessed to
be inversely proportional to the condition severity.
Based on the above, the proposed weights for the 3rd , 4th ,
th
5 and 6th order speed harmonics are:
wi =
1
, i = 3, . . . , 6
i
(5)
Therefore, the averaged condition indicator is calculated as:
4
PN
CIavg =
i=3 xi /i
PN
i=3 1/i
(6)
where N is equal to six and xi stands for the ith running
speed harmonic. At this point, it should be noted that one
of the above running speed harmonics may coincide with a
spectral component of electromechanical origin. In this case,
it is essential to disregard it in order to obtain consistent and
reliable description of the bearing condition.
B. Averaging of Temperature Data
Wind turbine generator bearings’ temperatures data are
employed as safety triggers and fault diagnosis tools. Due
to continuously varying load and running speed conditions,
temperature periodic and transient fluctuations are repeated
leading in thermal stress of the bearing and its lubrication. In
order to accomplish integrated temperature monitoring, it is
crucial to take into account absolute values as well as averaged
trends, both under various conditions.
The proposed method is exponential moving average which
assigns exponential weights on latest and past data, as shown
in Eq. 7. The coefficient α represents the degree of weighting
decrease, where high value discounts older observations faster
and low α corresponds to slow decay.
Si = Ei ,
i=1
Si = α · Ei−1 + (1 − α) · Si−1 , i = 2, ..., N
progressing bearing looseness. The trends do not increase
monotonically, but they rather follow an erratic pattern with
numerous peaks and valleys. The latter could be due to the
expected intermittent microscopic and macroscopic movement
and the operation of the generator. The same behaviour is
encountered in higher power production, which is not depicted
here.
However, the amplitude of the 6th running speed harmonic
is substantially higher due to the fact that it coincides with an
electromechanical component generated in this machine type.
As a result, this spectral component is not taken into account
into the computation of the weighted average condition indicator shown in Fig. 6. Based on evaluation of the trends shown
in Fig. 4, 5 and 6, the averaged condition indicator serves as
reliable representation of its subcomponents and thus the fault
diagnosis can be performed relying solely on it.
(7)
V. C ASE S TUDIES ON D ETECTION OF B EARING I NNER
R ACE C REEP
As discussed in section III-B, the development of rotational
looseness due to relative movement between the axle and the
inner ring follows various stages which can be detected by
the combined employment of vibration analysis and temperature data trending. Although the time interval between the
starting point of microscopic inner ring movement and actual
bearing spinning depends on numerous factors and cannot be
prognosticated consistently, any changes in either vibration or
temperature levels can be utilized towards early fault detection.
The following subsections present two study cases of bearing
inner race creep of different severity, where vibration and
temperature trends are illustrated along with frequency spectra
from accelerometers installed on the generator bearing houses.
A. Late Stage Bearing Creep
In this section, a late stage fault on one of the generator
bearings is presented from a vibration and thermal monitoring
point of view. The shaft was subjected to extensive wear
exceeding 2.0mm leading to the replacement of the generator.
Fig. 4 and 5 illustrate five months vibration trends of 3rd ,
th
4 , 5th and 6th order magnitudes in power bin 1, i.e. power
production below 25% of nominal output. The transfer rate
of these scalar values from the wind turbine to a centralized
server is approximately one hour. The first three harmonics under consideration present increasing trends indicating
Fig. 4: Vibration trends of 3rd and 4th running speed harmonics over five months in power bin 1. The trends present
erratic behaviour with multiple peaks for approximately three
months.
Fig 7 shows the power spectrum of the recorded vibration
signal, where the first six running speed harmonics are pointed
by arrows. The 3rd order harmonic dominates the spectrum
accompanied by increased 5th order in this file, suggesting
severe looseness pattern.
Increased temperature trends usually indicate that the bearing condition has deteriorated to late stage, where in the case
of bearing creep, it corresponds to continuous macroscopic
inner ring movement and shaft wear. Fig. 8 displays the raw
and averaged temperature values in low power production,
where both trends show a clear change of the thermal state
of the bearing. Although the averaged temperature levels are
lower, the overall trend is more stable and can be utilized in
combination with raw values to evaluate the severity of the
fault.
By comparing vibration and temperature data in Fig. 6
and 8 respectively, it can be seen that the weighted averaged
condition indicator shows signs of progression approximately
one month earlier. The early identification of the above men-
5
Fig. 5: Vibration trends of 5th and 6th running speed harmonics over five months in power bin 1. 5th order harmonic
present exponential increase followed by high decline. 6th
harmonic coincides with another spectral component and thus
it is substantially higher.
Fig. 6: Trending behaviour of weighted average condition
indicator excluding the 6th order magnitude in power bin
1. The indicator offers clear representation of the bearing
condition.
tioned development is assessed to be essential in regards to
minimization of potential secondary damages.
B. Early Stage Bearing Creep
Depending on the bearing type, operating conditions, lubrication and installation, bearing creep could follow alternative
motifs in regards to vibration response and development.
Fig. 9 and 10 display the vibration trends of the 3rd , 4th ,
5th and 6th running speed harmonics at low to medium
power production over six months period. Weak signs of
progression can be observed on 3rd , 4th and 6th harmonics,
whereas the 5th order show a clear step change. The bearing
Fig. 7: Vibration spectrum of late stage bearing inner race
creep. The amplitude of higher order running speed harmonics
is comparable to 1st and 2nd orders. The frequency resolution
is 0.1Hz.
Fig. 8: Raw (blue) and averaged (red) temperature trends
in power bin 1. The y axis scales are different in order to
explicitly display any changes.
operated for approximately three months under early rotational
looseness, which resulted in scratches and minor fretting on
the generator shaft. However, due to the challenging task of
predicting the remaining useful lifetime of this bearing and
fault development, proactive replacement is the commonest
and recommended practise.
Fig 11 depicts the power spectrum of bearing vibration
data after the trend increase, highlighting the higher order
speed harmonics. Although this vibration pattern reveals a
potential early stage fault, tracking of the speed related spectral
components over time establishes a reliable fault assessment
technique.
The prevailing issue arising from the above trends is the
confidence level based on whom a wind turbine has to be
6
Fig. 9: Vibration trends of 3rd and 4th running speed harmonics over six months in power bin 2.
Fig. 11: Vibration spectrum of early stage bearing inner race
creep. Only the 5th harmonic amplitude is increased.
Fig. 10: Vibration trends of 5th and 6th running speed harmonics over five months in power bin 2.
Fig. 12: Zoom in the trending behaviour of weighted average
condition indicator in power bin 2.
stopped for inspection and troubleshooting. Making use of a
generic looseness indicator, as shown in Eq. 6, the fault evaluation is rendered less complicated. Fig. 12 shows the trending
behaviour of the averaged weighted condition indicator, where
increase due to early creeping and decrease after the bearing
replacement are noticeable. On the contrary, the information
deduced by both raw and averaged temperature data do not
offer any useful diagnostic information. The latter is most
probably due to the low severity of the fault which did not
reach the state of increased heat generation.
vibration measurements respond earlier compared to temperature data up to several months depending on the progression
of the fault. Consecutive temperature peaks and permanent
temperature increase are clear signs that the bearing has
entered into late stage, resulting in the necessity of bearing
or generator replacement. Due to the fact that there is not
any clear pattern regarding which running speed harmonics
exhibit increasing trends under rotational looseness, it has been
proposed to utilize an weighted average condition indicator
based on the 3rd , 4th , 5th and 6th orders. The employment of
this indicator has reliable revealed the overall change of the
bearing condition, decreasing the number of total descriptors
and hence rendering the diagnosis procedure less complicated.
VI. C ONCLUSIONS
Consistent diagnosis of rotational looseness in wind turbine
generator bearings is crucial in terms of downtime minimization. Two cases of bearings subjected to severe and
early bearing creep respectively have been presented based
on vibration and temperature data. It has been shown that
R EFERENCES
[1] R. L. Arántegui, 2013 JRC wind status report - Technology, market and
economic aspects of wind energy in Europe. European Commission,
Joint Research Centre, Institute for Energy and Transport, 2014.
7
[17] Mobius Institute, Vibration Training Course Book Category II. Mobius
Institute.
VII. B IOGRAPHIES
Georgios Alexandros Skrimpas received the Diploma in electrical and
computer engineering from the Aristotle University of Thessaloniki, Greece, in
2009 and the M. Sc. in wind energy from the Technical University of Denmark
(DTU) in 2012. He is currently pursuing the Industrial Ph.D. degree at the
Centre of Electric Power and Energy at DTU in cooperation with Brüel and
Kjær Vibro. His research interests are diagnosis and prognosis of electrical
and mechanical faults in wind turbines.
Fig. 13: Raw (blue) and averaged (red) temperature trends
in power bin 1. The y axis scales are different in order to
explicitly display any changes.
[2] S. Struggl, V. Berbyuk, and H. Johansson, “Review on wind turbines
with focus on drive train system dynamics,” Wind Energy, 2014.
[3] J. Ribrant and L. Bertling, “Survey of failures in wind power systems
with focus on swedish wind power plants during 1997-2005,” IEEE
Transactions on Energy Conversion, vol. 22, pp. 167–173, 2007.
[4] F. Spinato, P. Tavner, G. van Bussel, and E. Koutoulakos, “Reliability of
wind turbine subassemplies,” IET Renewable Power Generation, vol. 3,
pp. 387–401, 2008.
[5] H. Ma, X. Zhao, Y. Teng, and B. Wen, “Analysis of dynamic characteristics for a rotor system with pedestal looseness,” Shock and vibration,
vol. 18, pp. 13–27, 2011.
[6] K. Marhadi and R. Hilmisson, “Simple and effective technique for early
detection of rolling element bearing fault: A case study in wind turbine
application,” in International Congress of Condition Monitoring and
Diagnostic Engineering Management, pp. 94–97, June 2013.
[7] G. A. Skrimpas, C. W. Sweeney, B. J. Bogi, N. Mijatovic, and J. Holboell, “Analysis of generator bearings vibration data for diagnosing rotor
circuit malfunction in dfigs,” in International Conference on Electrical
Machines, 2014.
[8] J. Murata and T. Onizuka, “Generation mechanism of inner race creep,”
tech. rep., Koyo, 2005.
[9] T. Nagatomo and D. G. Toth, “Investigation of the bearing damage
progression starting from cone creep of a railroad axle journal bearing,”
Quarterly Report of RTRI, vol. 47, pp. 119–124, 2006.
[10] C. Andersson, S. Gutt, and M. Hastings, “Cost–effective monitoring
solution using external surveillance centre,” in The 2nd World Congress
on Engineering Asset Management (EAM) and The 4th International
Conference on Condition Monitoring, 2007.
[11] N. M. A. Freire, J. O. Estima, and A. J. Marques Cardoso, “Opencircuit fault diagnosis in pmsg drives for wind turbine applications,”
IEEE Transactions on Industrial Electronics, vol. 60, pp. 3957–3967,
2013.
[12] B. Corne, C. Debruyne, P. De Baets, and J. Desment, “Stator current
measurements as a condition monitoring technology - the state-of-theart,” in International Conference on Electrical Machines, 2014.
[13] B. Ebrahimi, J. Faiz, M. Javan Roshtkhari, and S. Khatami, “Advanced
eccentricity fault recognition in permanent magnet synchronous motors
using stator current signature analysis,” IEEE Transactions on Industrial
Electronics, vol. 61, pp. 2041–2052, 2014.
[14] S. Djurovic, C. Crabtree, P. Tavner, and A. Smith, “Condition monitoring
of wind turbine induction generators with rotor electrical asymmetry,”
IET Renewable Power Generation, vol. 6, pp. 207–216, 2012.
[15] M. Picazo-Rodenas, J. Antonino-Daviu, V. Climente-Alarcon, R. RoyoPastor, and A. Mota-Villar, “Combination of non-invasive approaches for
general assessment of induction motors,” in International Conference on
Electrical Machines, 2014.
[16] T. Niwa, “A creep mechanism of rolling bearings,” tech. rep., NTN,
2013.
Ivaylo Gregoriev Dragiev received Bachelor and Master degrees in Marine
Engineering from the Technical University of Varna (Bulgaria) from 1997
to 2002. Since 2002 he has been working as marine engineer for various
international companies before switching to the wind industry with Vestas
where he worked for the service department from 2008 until 2014. Currently
he works as Diagnostic Engineer for Brüel and Kjær Vibro in Nærum,
Denmark. His research interests are in condition monitoring systems and
machine fault diagnosis for wind turbines.
Reynir Hilmisson received the B.Sc. and M.Sc. degrees in Electrical Engineering and Acoustics from the Technical University of Denmark in 2007
and 2009. Since 2009 he has been with Brüel and Kjær Vibro as Diagnostic
Engineer. Prior to joining Brüel and Kjær Vibro, he was an Acoustic Engineer
for Tymphany Company and Electrical Engineer Intern for APC. He holds an
ISO Level 3 certification in vibration diagnostics. His research interests are in
development and implementation of condition monitoring solutions specific
to wind power generation.
Christian Walsted Sweeney received the B.Sc. from the University of
Southern Denmark in 2006 and the M.Sc from the Technical University of
Denmark in 2008 both in mechanical engineering. From 2008 to 2010 he was
employed as a diagnostic engineer at Brüel and Kjær Vibro and since 2010
he is the team leader of the diagnostic services group. His research focus is
on the development of condition monitoring systems and handling of large
data quantities.
Bogi Bech Jensen received the Ph.D. degree from Newcastle University,
Newcastle Upon Tyne, U.K., for his work on induction machine design. He
was in various engineering and academic positions in the marine sector from
1994 to 2004. He was at Newcastle University from 2004 to 2010 first as a
Postgraduate, then Research Associate and finally as a Lecturer. From 2010
to 2014 he was Associate Professor and later Head of Research Group at the
Technical University of Denmark (DTU), Lyngby, Denmark. He is currently
Professor of Energy Engineering at the University of the Faroe Islands (UFI),
where he is responsible for education and research in energy.
Nenad Mijatovic received his Ph.D. degree from the Technical University
of Denmark for his work in superconducting machine. After obtaining his
Dipl.Ing. education at University of Belgrade, Serbia, he enrolled as a doctoral
candidate in 2012. Upon completion of the PhD, he has continued to work
in the same field of machine research - superconducting machines, as an
Industrial PostDoc. The 3 year industrial PostDoc grant has been provided by
Hojteknologifonden and supported by Envision Energy Aps., Denmark. Dr.
N. Mijatovic is a member of IEEE from 2008 and his field of interest and
research includes novel electrical machine design, operations and diagnostic.
Joachim Holbøll is associate professor and deputy head of center at DTU,
Department of Electrical Engineering, Center for Electric Power and Energy.
His main field of research is high voltage components, their properties, condition and broad band performance, including insulation systems performance
under AC, DC and transients. Focus is also on wind turbine technology and
future power grid applications of components. J. Holbøll is Senior Member
of IEEE.