2013 IEEE International Conference on Solid Dielectrics, Bologna, Italy, June 30 July 4, 2013

Partial Discharge Investigation and Electric Field

Analysis of Different Oscillation Modes of Water
Droplets on the Surface of Polymeric Insulator under
Tangential AC Electric Field Stress
M. H. Nazemi and V. Hinrichsen
High Voltage Laboratories
Technische Universitt Darmstadt
Darmstadt, Germany
nazemi@hst.tu-darmstadt.de
Abstract For this contribution, different oscillation modes
of a 20 microliter water droplet on the hydrophobic surface of a
silicone rubber insulator under tangential AC electric field stress
were investigated, using a high speed camera. Since the
deformation of water droplet increases the local electric field
stress at the triple zones, an electric field analysis at the triple
zone is performed to find the worst case of water droplet
oscillation modes. Different kinds of water droplet deformation
at four dominant oscillation modes, namely mode No. 1 to mode
No. 4, which were found experimentally, are simulated and
compared. These results show that mode Nos. 1 and 3 provide
higher electric field values at the triple zone in comparison to
mode Nos. 2 and 4. In parallel, partial discharge inception
voltages of a 20 microliter water droplet located on the
hydrophobic surface of a silicone rubber insulator under
tangential AC electric field stress are measured in the frequency
range of 20 Hz to 100 Hz. With the aid of frame analysis of the
captured videos, the mode number of water droplet oscillation at
each frequency is obtained, and the measured partial discharge
inception voltages at each oscillation mode are compared with
electric field simulation results. Partial discharge investigations
confirmed that the water droplets oscillating at mode Nos. 1 and
3 have lower inception voltages compared to PD inception
voltages of water droplets that oscillate at mode Nos. 2 and 4.
These measurements also show that inception voltage makes a
step whenever the oscillation mode of water droplet changes.

the electric field especially nearby the triple zones (the

common line between air, water droplet and insulator surface)
[2-5]. Periodical deformation of a water droplet in an AC
electric field makes it difficult to find an exact value of an
electric field enhancement factor. In [6] the authors reported
four dominant oscillation modes of individual water droplets
on hydrophobic insulator surfaces under high voltage
tangential electric field stress. These oscillation modes, which
are named Mode 1 to Mode 4, were categorized in a wide
range of droplet volume and frequency of the applied electric
field. Fig. 1 shows these four dominant oscillation modes. In
mode No.1 the water droplet elongates towards the electrodes
and develops sharp edges on one side periodically. In mode
No.2 the water droplet oscillates upwards and downwards two
times per cycle of the applied electric field. The water droplet
at mode No.3 moves in a way that two different peaks occur in
its left and right part during one half-cycle each of the applied
electric field. In mode No.4 stronger oscillation occurs at the
top of the water droplet [6].

Keywordswater droplet; partial discharge; oscillation mode;

hydrophobic surface; electric field

I.

Fig. 1. Four dominant oscillation modes of a water droplet on silicone

rubber surface under tangential electric field stress [6]

INTRODUCTION

Application of polymeric insulators in high voltage

engineering has been extended recently because of their low
weight, resistance to vandalism, high contamination resistance
and most important hydrophobic surfaces [1-3]. However,
surface degradation of polymeric insulators due to aging as
well as their surface discharge mechanisms are amongst the
most challenging questions for researchers nowadays. Aging
of the polymeric insulator surface can gradually decrease
hydrophobicity of the surface, which leads to corona and even
surface flashovers. Under wet conditions, water droplets are
formed on the surface of polymeric insulators, which enhance

For this contribution, different oscillation modes of a 20 l

water droplet on the hydrophobic surface of a silicone rubber
insulator under tangential AC electric field stress in the
frequency range of 20 Hz to 100 Hz are investigated, using a
high speed camera. Different kinds of water droplet
deformation at four dominant oscillation modes (mode No.1 to
mode No.4), which were found in the experiments, are
simulated, and the electric field near the triple zone is
analyzed to find the worst case of water droplets oscillation
mode. In parallel, partial discharge inception voltages of a
20 l water droplet located on a silicone rubber surface under

978-1-4673-4461-6/13/$31.00 2013 IEEE

194

2013 IEEE International Conference on Solid Dielectrics, Bologna, Italy, June 30 July 4, 2013
tangential AC electric field stress in the frequency range of
20 Hz to 100 Hz are measured. The measured partial discharge
inception voltages at each oscillation mode are compared with
electric field simulation results.
II.

TEST SETUP

Fig.2 shows the setup for an individual water droplet on an

insulating surface under tangential electric field stress [3, 6].
This setup consists of two high-voltage electrodes, which are
embedded in insulating material plates from opposite sides.
Applying a high voltage of adjustable frequency to the
electrodes generates a virtually homogenous tangential electric
field on the surface. Individual water droplets are placed
between the electrodes, using a high precision pipette. To
make the water droplet more visible, methylene blue is used to
colorize the water droplet (which was proven to have no
impact on the electrical and mechanical properties of the
droplet).
Oscillation of the water droplet under applied electric field
stress is captured by a SHIMADZU Hyper Vision (HPV2)
high speed camera with 1000 frames per second. After each
measurement the surface is cleaned and a new water droplet is
placed on the specimens surface. To measure the PD
inception voltage of the water droplet on the surface, an
accurate PD measurement system based on IEC 60270
requirements is provided (see Fig.3). The MPD 600 Omicron
PD measurement system is used inside a shielded test cabinet
with background noise level of less than 1 pC.

Fig. 2. Configuration of test specimen (Dimensions in mm)

Fig. 3. Partial Discharge measurement circuit

III.

mode are selected (See Table. I). Dimensions of the water

droplet are measured in each frame. The water droplet at
initial condition has 4 mm length, while after voltage
application and due to elongation towards the electrodes its
length reaches 5 mm at mode 1 and 4.4 mm at modes 2, 3 and
4. The E-field distribution is simulated in the following parts
using CST EM-Studio software. The applied voltage in all
the cases is 1 kV (500V). In order to achieve best accuracy of
the results, a thin ring of silicone rubber at the position of
water droplets on the insulating surface is considered, which
allows the software refining the mesh quality at the triple
zones. Fig.4 shows the simulation model. Because the value of
maximum electric field at the triple line reaches to infinity in
the ideal case, the software has a singularity problem at the
triple lines. In other words, by increasing the mesh quality, the
value of maximum E-field strength at the triple lines will be
increased. Special criteria of convergence have to be fulfilled
to be sure that the simulation results are physically correct [3].
For this contribution, the mesh quality step sizes are varied
between three different values in all simulation cases. At
ca. 80% of simulation cases, it is observed that these three
mesh qualities have the same maximum field strength values
at the distance of about 0.1 mm away from the triple line of
the water droplet. It shall be noted that considering the E-field
value at 0.1 mm distance from the edge of the water droplet is
also physically correct because the discharge in triple zones
need a certain minimum volume to develop an avalanche of
electrons. Since the E-field enhancement factor at triple line is
strongly dependent on the software accuracy and the
singularity problem at triple line, considering the converge
point instead of the triple line is the first option to eliminate
the singularity problem, but this converge point is also
dependent on the mesh quality. There are several different
calculated or simulated E-field enhancement factors in the
literature, which may confuse the reader to find the exact Efield enhancement factor. For this contribution four oscillation
modes of a single water droplet on the surface of an insulator
were compared, considering same initial conditions and mesh
qualities in all simulation cases. It shall be noted that the Efield enhancement factors can vary based on different initial
assumptions. The authors are working to develop a method to
find a future better definition of the E-field enhancement
factor based on the integral of an ionization coefficient over a
certain volume near the triple zones. The simulation result of
the test object without water droplet is shown in Fig.5. The Efield between electrodes and at the position of water droplet is
almost merely tangential (parallel to the surface). The E-field
value at the center of the test object can be considered as a
constant value of E0 equals to 0.4 kV/cm. Fig.6 shows a
typical case of simulation with three mesh qualities. In the
following, the average E-field value at 0.1 mm distance from
the triple line from three different mesh qualities is considered
as the maximum E-field value Em.

RESULTS

A. Electric Field Simulation

Since the water droplet oscillates under AC electric field
stress periodically [2, 6], one cycle of this oscillation is
analyzed. With the help of the captured videos four different
frames of water droplet oscillation during one cycle of each

Fig. 4. Simulation model; overal view (left) and zoomed in view (right)

978-1-4673-4461-6/13/$31.00 2013 IEEE

195

2013 IEEE International Conference on Solid Dielectrics, Bologna, Italy, June 30 July 4, 2013
TABLE I.

FOUR DIFFERENT SHAPES OF A 20 L WATER DROPLET UNDER

ONE CYCLE OF TANGENTIAL AC E-FIELD STRESS
Initial condition
(non-deformed)
Frame 1

Frame 2

Frame 3

Frame 4

Mode 1
Mode 2
Mode 3
Mode 4

of non-deformed water droplet on the surface. This

enhancement is 3.10, 3.22 and 3.16 respectively for mode
numbers 2, 3 and 4. It shall be noted that these enhancements
are only due to different shapes of water droplet, when the
lengths of water droplet are constant. Considering that the
length of water droplet can also change during its
deformations, the enhancement E-field factors can be higher.
An E-field enhancement factor of up to 8 is reported in [3]
under different initial conditions. Mode numbers 2, 3 and 4
were found to have only few differences of their k factors at
different frames. This can also be explained by their same
lengths. It is also visible that the field enhancement is higher
for lower contact angles of the water droplets.
2.5

2.5

0.4

mesh 0.1
mesh 0.3
mesh 0.5

2
1.5

0.3

E-Field (kV/cm)

E-Field (kV/cm)

0.2
0.1

1.5

0.5
1

0
2

2.5

3.5

0.5

-20

-10

10

20
0

X (mm)

-10

Fig. 5. Tangential E-field of the test object without water droplet

The maximum E-field at each kind of deformed water

droplet is normalized by the constant E-field at initial
condition (E0). So the field enhancement factor k is defined as
follows:

Em
E0

10

ENHANCEMENT FACTOR OF DIFFERENT FRAMES AT DIFFERENT

MODES

(1)

E-field enhancement factor k

where Em is the maximum tangential E-field at 0.1 mm

distance from the triple line of a deformed water droplet and
E0 corresponds to constant E-field at the center of the test
object without water droplet.

Fr.1

Fr.2

Fr.3

Fr.4

Max.

Mode 1

3.45

3.36

3.25

3.51

3.51

Mode 2

3.10

3.07

3.04

3.02

3.10

Mode 3

3.22

3.19

3.16

3.13

3.22

Mode 4

3.16

3.02

3.04

3.07

3.16

Fig. 7 compares the maximum k factors of four modes. As

can be seen, mode No.1 has a big margin from the other
modes. It is also observable that mode numbers 1 and 3 result
in higher E-field enhancement factor than mode numbers 2
and 4.
3.60
3.50
3.40

k factor

The simulation of a non-deformed water droplet at initial

condition results in a maximum tangential E-field at 0.1 mm
distance from the triple line of 1.15 kV/cm. Thus, presence of
the water droplet on the surface increases the tangential
component of the E-field at the converge point by factor of
about 2.9 (i.e. 1.15/0.4). Since the water droplet shape changes
differently at different modes, four different kinds of water
droplet oscillation at four modes based on Table. I are
simulated and compared with the maximum E-field of a water
droplet at initial condition (non-deformed droplet).
Considering (1), the field enhancement k factors for each case
are calculated. Table. II summarizes the results of k factors
for different degrees of deformation. As can be seen, the k
factor for mode 1 in all of the four cases is higher than for the
other modes. The reason is firstly the length of the water
droplet, which is bigger than that of the other modes, and
secondly the low contact angle of water droplet at mode 1 in
comparison with the other modes. Considering the maximum k
factor, it seems that the deformation of water droplet at
mode 1 increases the local E-field at converge point by factor
of 3.51 which is 22% more than maximum E-field in presence

X (mm)

Fig. 6. Typical case of simulation (mode 1 frame 4) with three values of

mesh quality

TABLE II.

k=

-5

Mode 2

3.20

Mode 3

3.10
3.00
2.90

Fig. 7. Enhancement factor k for different modes

978-1-4673-4461-6/13/$31.00 2013 IEEE

196

Mode 1

3.30

Mode 4

2013 IEEE International Conference on Solid Dielectrics, Bologna, Italy, June 30 July 4, 2013

IV.

CONCLUSION AND OUTLOOK

For this contribution, oscillation modes of a 20 micro liter

water droplet on the surface of silicone rubber insulator under
tangential AC electric field were investigated. Four different
deformations at four dominant oscillation modes, namely
mode No.1 to mode No.4, were simulated and compared with
PD inception voltages of a water droplet at these oscillation
modes. Simulation results show that the mode Nos.1 and 3
provide higher electric field strength at the triple zone in
comparison to mode Nos.2 and 4. Experimentally determined
partial discharge inception voltages in the frequency range of
20 Hz to 100 Hz confirm the simulation results. Water
droplets oscillating at mode Nos.1 and 3 were found to have

lower PD inception voltage in comparison to mode Nos.2 and

4. The measurements also show that PD inception voltage
makes a step whenever the oscillation mode of water droplet
changes. An improved method to determine the E-field
enhancement factor shall be developed in the future work to
find a physically correct value of maximum electric field
strength at the triple zones. These investigations shall be
continued to develop a simulation model of water droplet
deformation and movement on the surface of polymeric
insulators in order to precisely predict the partial discharge
inception field strength.

4.65

PD inceptionE-field kVr.m.s./cm

B. Partial Discharge Measurements

As described in the introduction, an accurate PD
measurement system based on IEC 60270 requirements is
applied (see Fig. 3). Using the MPD 600 Omicron PD
measurement system, the PD inception voltage of a
20 micro liter water droplet located on the surface of silicone
rubber surface under tangential AC E-field is measured.
Before each measurement the surface of specimen is cleaned
to avoid any effect of dust on the experiment. A new water
droplet is placed carefully between the electrodes, using
pipette. The voltage is slowly increased in small steps of about
300 V until the first signal of partial discharges (1 pC) is
detected. This procedure is repeated six times. The average
value, normalized to atmospheric standard conditions
(humidity directly at the triple zone is basically unknown and
was assumed to be the same as in far distance to the droplet),
is noted as PD inception voltage. From this voltage the
inception field strength is determined by calculating of E-field
strength at the center of the test object without water droplet
under this measured PD inception voltage. A large number of
measurements in a wide range of frequencies between 20 Hz
to 100 Hz are performed. Results are depicted in Fig. 8. With
the aid of captured videos, the oscillation modes of the 20 l
water droplet are observed. It has oscillation mode number 1
at 20, 30 and 40 Hz and mode number 2 at 50 Hz. Mode
number 3 is found to occur at 60 and 70 Hz, and at 80 Hz to
100 Hz the water droplet has oscillation mode number 4. As a
first finding, the water droplet oscillates at higher modes with
increasing frequency. Secondly, mode numbers 1 and 3 result
in lower values of PD inception voltage than mode numbers 2
and 4 (see Fig.8). These results are in full agreement with the
simulation results, which show that mode numbers 1 and 3
result in higher E-field enhancement factors than mode
numbers 2 and 4. The reason can be explained by comparison
of the contact angles of the water droplets at different modes.
A water droplet at mode numbers 1 and 3 has lower contact
angles than at mode numbers 2 and 4. The other interesting
result is that the PD inception voltage makes a step when the
oscillation mode of water droplet changes. These steps or
jumping can be seen at frequencies 40, 50 and 60 Hz, at which
the mode numbers change (Fig. 8). It is finally worth
mentioning that inception field strengths may differ for 50 Hz
and 60 Hz power frequency due to possibly different
oscillation modes at these frequencies. More investigations
about these phenomena are reported in [6].

mode 2

4.6

mode 4

4.55
4.5
4.45
mode 3

4.4
4.35
4.3
mode 1

4.25
4.2
4.15
0

10

20

30

40

50

60

70

80

90

100

Frequency (Hz)

Fig. 8. PD inception E-field for 20 l water droplet on the SR surface

under tangential AC electric field stress

ACKNOWLEDGMENT
These investigations are part of a cooperation with other
research groups in the frame of the Collaboration Research
Center Transregio 75 Droplet Dynamics Under Extreme
Ambient Conditions (http://www.sfbtrr75.de/). The authors
would
like
to
thank
the
DFG
(Deutsche
Forschungsgemeinschaft) for the financial support of this
project.
REFERENCES
[1]
[2]

[3]

[4]

[5]

[6]

978-1-4673-4461-6/13/$31.00 2013 IEEE

197

R. Hackam, Outdoor HV Composite Polymer Insulators, IEEE. Trans.

DEI, Vol. 6, pp. 262-268, 1999.
S. Keim, D. Koenig and V. Hinrichsen, Experimental Investigations on
Electrohydrodynamic Phenomena at Single Droplets on Insulating
Surfaces, IEEE. Conf. Electr. Insul. Dielectr. Phenom., pp.133-136,
2003.
S. Feier-Iova, The Behavior of Water Droplet on Insulating Surfaces
Stressed by Electric Field PhD Thesis, TU-Darmstadt, 2009.
http://tuprints.ulb.tu-darmstadt.de/2008/
A. J. Phillips, D. J. Childs and H. M. Schneider, Aging of Non-Ceramic
Insulators due to Corona from Water Drops, IEEE Trans. Power
Delivery, Vol. 14, No. 3, pp. 1081-1089, 1999.
M.H. Nazemi, V. Hinrichsen, Partial Discharge of Water Droplets on
Polymeric Insulating Surfaces, 12th International Electrical Insulation
Conference, INSUCON 2013, Birmingham, UK, May 2013.
M.H. Nazemi, V. Hinrichsen, Experimental Investigations on Water
Droplet Oscillation and Partial Discharge Inception Voltage on
Polymeric Insulating Surfaces under the Influence of AC Electric Field
Stress, IEEE Trans. Dielectric Insul., Vol. 20, No.2, pp. 443-453, 2012.