Journal of Molecular Liquids 284 (2019) 700–708

Contents lists available at ScienceDirect

Journal of Molecular Liquids

journal homepage: www.elsevier.com/locate/molliq

Effect of Al2O3 nanorods on dielectric strength of aged transformer oil/

paper insulation system
Muhammad Rafiq a,b,d,⁎, Li Chengrong b,d, Yuzhen Lv c,d
a
Department of Electrical Engineering University of Engineering and Technology, Taxila, Pakistan
b
Beijing Key Laboratory of High Voltage & EMC, School of Electric and Electronic Engineering, North China Electric Power University, Beijing, 102206, China
c
School of Energy, Power and Mechanical Engineering, North China Electric Power University, Beijing, 102206, China
d
State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources, North China Electric Power University, Beijing, 102206, China

a r t i c l e i n f o a b s t r a c t

Article history: With the commencement of high voltage alternating current (HVAC) and high voltage direct current (HVDC), the
Received 21 October 2018 demands on insulating reliability of power transformer is getting more and more imperative. The mineral oil
Received in revised form 5 April 2019 (MO/TO) in conjunction with paper is the main insulation components in oil-immersed transformers; their insu-
Accepted 6 April 2019
lation properties play a significant role in the safe and stable operation of power transformers. To strengthen the
Available online 11 April 2019
insulation level of ultra-high voltage transformer and to reduce its size and weight, it is imminent to enhance the
Keywords:
insulating performance of transformer oil and oil-impregnated cellulose. Recently, a unique novel effort of sus-
Nanofluids pension of nanoparticles (NPs) into mineral oil (MO) has been carried out and the results have exhibited im-
Topic: proved insulation characteristics of transformer oil. The Al2O3 nanoparticles with favorable features were
Lightning impulse voltage prepared in our lab to develop transformer oil/paper insulation system with better insulation performance.
Streamer The transformer oil-based nanofluids (NFs) were prepared with nanorod shape, oleic acid (OA) surface modifi-
Electron mobility cation and 0.8 g/L concentration of NPs. The pressboards were cut into the desired sizes (85 mm.55 mm). The
Topic: thickness of the pressboards was 2 mm applied in flashover tests. The prepared pressboards were put into an
Trap
oven at 105 °C for almost 48 h to carry out the hot air drying operation. Then they were put in the vacuum drying
Oil-pressboard
at 85 °C for 48 h. The impregnated pressboards were prepared by impregnating them into dried oil and NF under
Creeping flashover
vacuum beneath 1 kPa at 80 °C for almost 48 h to obtain the oil-impregnated pressboard (OIP) and nanofluid-
impregnated pressboard (NIP) respectively. Additionally, the prepared insulation system samples of liquid
(MO and NFs) and solid (OIPs, NIPs) were thermally aged at 130 °C for 30 days. The effect of alumina NPs on ther-
mal aging of transformer oil and pressboard were also investigated. The results showed that Al2O3 NPs can slow
down the aging of oil and pressboard. The average AC and positive LI BDV of aged NFs was 11% higher than aged
MOs. The average LI creeping FOVs of NIPs were 6% higher than OIPs. Additionally, average AC creeping FOVs of
NIPs were 8% higher than that of OIPs. The enhancement in insulating performance of aged NFs and NIPs as com-
pared to MO and OIPs is interpreted in light of electron traps theory.
© 2019 Elsevier B.V. All rights reserved.

1. Introduction these power transformers is cellulose insulation (oil-impregnated

paper/pressboard) and mineral oil (MO) for their reliable operation
TRANSMISSION and distribution system is the most critical, heavily [3]. The transformer collapse data manifested that the average service
loaded and expensive part of the electric network. The power trans- life of the transformers that collapsed due to their dielectric insulation
formers are the most essential element of this electric network system problems is 17.8 years, which is almost half of their expected life [4];
[1]. The potential consequences of this critical element collapse may and the dielectric insulation issues blamed for approximately 75% of
be quite damaging [2]. The most of the transformers in service in the high voltage transformer failures [5]. Both the lifetime and opera-
many industrial states are close to or beyond their designed life, there- tional reliability of the transformers mainly rely on the resistance of
fore, there has been an enhanced focus to improve the functional reli- the aging effect. It is therefore highly essential to enhance the electrical
ability of these existing units. The insulation system normally used in characteristics of insulation material.
The primary insulation medium for power transformers typically in-
cludes a highly refined transformer oil (TO) and cellulosic insulation.
⁎ Corresponding author at: Beijing Key Laboratory of High Voltage & EMC, School of
Electric and Electronic Engineering, North China Electric Power University, Beijing
The transformer oil is required to be stable at high temperature which
102206, China. is essential to suppress arcing, to serve as coolant and electrical insula-
E-mail address: muhammad.rafiq@uettaxila.edu.pk (M. Rafiq). tion [6–9]. The oil applied in transformers also has an effect on the

https://doi.org/10.1016/j.molliq.2019.04.041
0167-7322/© 2019 Elsevier B.V. All rights reserved.
 M. Rafiq et al. / Journal of Molecular Liquids 284 (2019) 700–708 701

dielectric characteristics of paper and pressboard insulations. The oil insulating features of NFs and NFs-impregnated pressboard is essential
used act as electrical insulation between conductive parts and also for their practical applications.
removes the heat produced during the operation of transformer. Addi- Numerous investigations have aimed at the aging mechanisms of in-
tionally, weight, size of the transformer and current density of trans- sulation system of transformers, and they can typically be divided into
former windings depend on the amount of oil and the rate of heat two classifications: first, examines the aging by-products such as
transfer [10]. acids, furfural, characteristics gases and chemical modifications while
The mineral oil (MO) and oil-insulated pressboard (OIP) of oil- aging process, such as the DP; and second, investigations on the influ-
immersed transformers undergoes multiple field stresses aging during ence of aging on electrical insulation traits of oil/paper insulation such
operation (due to electrical stress, thermal stress, chemical stress), breakdown voltage, partial discharge (PD) etc. This chapter produced
thereby accelerating the aging of solid-liquid insulation structure important information on the influence of Al2O3 NPs on the electrical in-
[11–15]. The extensive application of this oil/paper insulation system sulation conduct of oil/paper insulation system. Majority of research
in HV equipment has led to huge research work with focus to enhance studies have focused mainly on the influence of NPs on the MO or in
the dielectric traits of this insulation system. some cases NPs effect on the pressboard but for very short time of accel-
As mentioned transformer play a significant role in power systems in erated aging period. This study provides a comprehensive investigation
enhancing their operational lifetime and reliability which is essential to on the effect of Al2O3 NPs on the electrical features of transformer insu-
prevent downtime and expensive replacements. The lifetime and reli- lation system (oil/pressboard) after accelerated aging at 55 °C and 130
ability of transformers strongly rely on suitable insulation design °C for 30 days. The electrical properties of the aged samples were ob-
[16–19]. The insulation performance quality and physical level of trans- served after certain time intervals.
formers is directly linked to the safety and stability of power grid. The The aging of insulation system will lead to a decline in performance
power transformers are generally reliable during their design life of of transformer insulation system due to deterioration of dielectric
20–35. In reality, their lifetime could be enhanced to 60 years with suit- strength. Additionally, the partial discharge (PD) happens more easily
able maintenance techniques [20]. During operation, the insulation sys- and frequently in aged insulation system. The products which result of
tem (pressboard, oil) always exposed to high working temperature, PD such as humidity, acid and so on in turn accelerate aging process.
therefore the temperature is the most important factor in aging of the Consequently, is becomes highly significant to explore an economic
insulation system of transformer. The research has manifested that in- and essential way to enhance the dielectric performance of this insula-
sulating features of TO would not be affected significantly even if the tion system.
oil has been used for many years [21] but the long term operation at Aging will lead to a decline in working performance of transformer
high temperature will cause apparent effects to paper insulation and oil due to deterioration of dielectric strength and cooling property. Fur-
cause operational complications for normal operation of transformers thermore, partial discharge occurs more easily and frequently in aged
[22]. oil. Products of PD such as moisture, acid and so on in turns accelerate
During operation, the paper insulation is subjected to stresses (elec- aging process of the oil and form a vicious cycle. Consequently, it is ex-
trical, stress, thermal stress due to heat produced by the conductors car- tremely urgent to explore an effective and economical way to improve
rying current or due to hot spot development and chemical stress due to the dielectric capability of aged transformer oil.
breaking of chemical bonds of oil/paper material and producing acidic Power transformers are usually very reliable, with a 20–35 years of
slag). The insulation conduct of oil-paper will decline due to these design life. In practice, the life of a transformer can be as long as
stresses, such as the DP (degree of polymerization) value, when it is ex- 60 years with appropriate maintenance [2]. However, after several
posed to thermal stress and then cause partial discharge [23–24]. After years of service, aging of the cellulose material in the insulating press-
the partial discharge happens, the discharged particles (ions, electrons) board causes the insulation to gradually deteriorate and thus become
influence the surface of insulation material and enhances its aging pro- less effective. Actual data from power distribution networks do indeed
cess. The insulating features of insulation material would impair with indicate that creepage discharge failures become more common as
time inside the electric apparatus. This de-gradation rate is affected by transformers age. Although an oil-impregnated pressboard (OIP) has
the temperature as well as moisture and air [25]. generally desirable electrical properties, it has been discovered that
The surface discharge at the interface of solid liquid insulation is the this material exhibits weak resistance to creepage discharge. Irrevers-
key reason of collapse in transformers [26]. The multiple stresses cause ible damage can be caused by creepage discharge on the pressboard,
physical and chemical degradation of the pressboard materials, causing and this damage may eventually lead to flashover and thus to cata-
surface charge accumulation on its surface and leading to surface dis- strophic transformer failure. Consequently, it is important to under-
charge activity [27–30]. This surface charge accumulation strengthens stand creepage discharge induced transformer failures, especially with
the tangential electric field and creating high magnitude discharges regard to transformer aging.
without producing any BD [31]. The creepage discharge on pressboards In this paper, insulation aging properties of oil samples (MO, NFs)
is one of the failure modes for power transformers. Failure studies gen- and pressboards (OIPs, NIPs) were investigated after aging 6 days at
erally observe the evidences of creepage discharge in the “treelike” pat- the temperature of 130 °C. Based on the trap characteristics in both oil
terns, the carbonized marks left by electrical discharge on insulation samples (MO, NFs) and pressboards (OIPs, NIPs), we propose a modifi-
board [32–33]. Aging will result in reduction in the operational per- cation mechanism of nanoparticles on the insulating properties of trans-
formance of insulation system. Moreover, PD will happen more former oils.
often. The resulting products od PD such as acid, moisture and so
on further accelerate the aging process and initiate a dangerous 2. Materials and methods
cycle. The aging (degradation) process of oil/paper will be expedite
and result in the collapse under the compound effects of the factors 2.1. Formation of transformer oilbased nanofluids
mentioned before. Therefore, it is highly required to find an effective
and economic way to enhance the BD performance of insulation sys- The process of transformer oil-based nanofluids is presented in the
tem of transformers. following sections.
The research has manifested that NPs have the potential to enhance
the insulating features of TO and TO-based pressboard under ambient 2.1.1. Material selection
conditions [33–35]. But as mentioned earlier that the characteristics of The Al2O3 NPs are developed in our laboratory. The mineral oil
insulation system would deteriorate with certain period of time inside (25#Kelamayi) was processed to remove the impurities and to meet
the equipment under operation. Thus, maintaining the superior the requisites of pure oil set by CIGRE board 12.17 [36]. As NPs have
702 M. Rafiq et al. / Journal of Molecular Liquids 284 (2019) 700–708

high surface energy and a tendency to agglomerate, suitable surface mentioned in the earlier sections. The NFs were developed with
modifiers must be chosen to enhance the dispersion characteristics 0.8 g/L concentration of NPs by using the process as presented in the
and stability of NPs [37]. Surfactant techniques have been applied to de- aforementioned sections. The oil samples (pure oil and NFs) were put
velop a glazed shell on the NPs. The rear ends of surfactant molecules re- into vacuum oven at 80 °C, less than 100 Pa for approximately 48 h to
buff each other to ensure a distance between the particles to hinder diminish the influence of gases and moisture appeared during the prep-
them from agglomeration. Surface modification of NPs is an efficient aration process. The pressboard was B3.1 type, pre-compressed sup-
method to avert NPs agglomeration in insulating liquids [38–41]. Never- plied by a company named Weidman [42]. The pressboards were cut
theless, the surface modifiers used for mineral oils cannot be used with into the desired sizes (85 mm.55 mm). The thickness of the pressboards
other carrier oils, such as vegetable oils, because of different molecular was 2 mm applied in flashover tests. The prepared pressboards were
forms/structures. put into an oven at 105 °C for almost 48 h to carry out the hot air drying
operation. Then they were put in the vacuum drying at 85 °C for 48 h.
2.1.2. Preparation of nanoparticles The impregnated pressboards were prepared by impregnating them
The Aluminum Nitrate Nonahydrate (Al(NO3)3.9H2O) is put into a into dried oil and NF under vacuum beneath 1 kPa at 80 °C for almost
flask and stirred and the base solution is continuously dropped into 48 h to obtain the oil-impregnated pressboard (OIP) and nanofluid-
the solution during all stirring operation of 2 to 3 h. After stirring oper- impregnated pressboard (NIP) respectively.
ation, the pH value is measured by pH meter. The pH value decides the
shape of the NPs. After it, the solution is put into chemical reactor at 200 2.3. Thermal aging treatment process
°C for almost 12 h and then the solution is cooled down until room tem-
perature. After cooling, the product is then centrifuged three times by The life of the oil-paper insulation system is determined by its ther-
the application of ethanol. The product obtained is AlOOH and it is mal aging properties. When the operating temperature of the oil-paper
dried at 70 °C for almost 2 h. The resulting product is grinded and insulation system is in the range of 80 °C to 140 °C, the life loss can be
then sintered at 600 °C for 3 h. The final product obtained is Al2O3 nano- estimated by a factor of 6, and the insulation life is reduced by 6 °C
particle. After it, surface modification is applied. Then sonication is car- [43]. Therefore, we selected to age at 130 °C for 30 days to simulate
ried out at 60 °C for 2 h. After sonication, NPs are washed with ethanol the aging of the transformer insulation for almost 27 years. The samples
through centrifuge machine. In the end, surface modified NPs are prepared were exposed to nitrogen for almost 5 min to remove the air
obtained. from the samples and then covered them tightly by the cap on the bot-
tles. The AC and impulse breakdown measurements were carried out
2.1.3. Preparation of nanofluids after every 6 days for MO and NFs. The creeping flashover voltages
A definite quantity of prepared NPs was weighed in accordance to were measured for OIP and NIP after every 6 days.
the required concentration. The NFs were developed by suspension of
weighed NPs into MO by applying ultrasonic route. All the developed 2.4. AC and lightning impulse breakdown voltage measurement test
NFs samples were preserved into vacuum drying box at less than kPa
for almost 24 h to expel the influence of bubbles created during the son- The most essential and significant perquisite for an insulation liquid
ication action. The morphology of nanomaterials prepared by observing is AC breakdown voltage (BDV), which can be defined as the value of ap-
JEOL JEM-2100F high-resolution transmission electron microscopy plied AC voltage at which disruptive discharge is initiated in the liquid. A
(TEM) is shown in Fig. 1. The nanoplate has 100 nm length and 40 nm number of test methods are usually employed in which a small amount
width whereas the nanorods have diameter of 15 nm and length of of insulating liquid is exposed to an almost homogenous electric field
150 nm. between two electrodes dipped in insulation fluid. The voltage is in-
creased until breakdown occurs. The IEC 60156 standard test is nor-
2.2. Preparation of transformer oil based nanofluids-impregnated mally used with a gap of 2.5 mm and voltage rise rate of 2.5 kV/s [44].
pressboard The AC breakdown test is significantly influenced by impurities in the
insulating liquid, such as small particles, moisture, and air or gas bub-
The oil was refined to discard the contaminants to fulfill the require- bles. Therefore, the measured AC breakdown voltage of an insulating
ments set by CIGRE. Al2O3 NPs were prepared according to the process liquid mainly indicates the oil quality rather than oil properties itself.

Fig. 1. HRTEM images for nanoparticles: (left) Al2O3 nanoplates (right) Al2O3 nanorods.
 M. Rafiq et al. / Journal of Molecular Liquids 284 (2019) 700–708 703

Lightning impulse (LI) BD test was conducted to measure the BDV of

liquid under standard LI (1.2 μs/50 μs). The impulse BDV is stimulating
lightning strikes and typically uses 1.2 microseconds rise for the wave
to attain 90% amplitude and then drops down to 50% amplitude after
50 microseconds. The standard Lightning impulses were generated by
a 10-stage impulse voltage generator. The LI BD voltages of NFs and oil
were measured by employing a test cell and needle to sphere geometry
(Fig. 2). The schematic diagram for lighting impulse measurement of oil
samples with needle to sphere configuration is shown in Fig. 5. The test
was carried out according to IEC 60897-1987 standard with an electrode
gap distance of 25 mm. An initial chosen value of voltages is applied to
the sample and then elevated step by step until a BD occurs. After every
BD, the needle electrode is renewed to guarantee the identical experi-
mental conditions. For every sample, six BD values are obtained to en-
sure repeatability. The value of impulse voltages at which BD happens
Fig. 3. Test cell geometry for flashover voltage measurement.
and time to BD are noted from the oscilloscope and the average values
are considered as experimental results [45]. Meanwhile, for the efficacy
of results, three impulse waves must be applied at each voltage level capacitor is 800 pF. The frequency test platform is shown in Fig. 8.
and each sample should endure at least three voltage levels prior the DobleLemke LDS-6 PD equipment was used to detect PD signals. The
happening of BD. The test cell and electrodes cleansed off with the test cell without the needle electrode was found to be PD free up to
same fluid. To avoid the effect of bubbles, it was ensured that there 60 kV with the lab background noise less than 3 pC. Partial discharge in-
are no bubbles during the filling up of the test cell with the developed ception Voltage (PDIV) here is defined as the voltage under which the
NFs. The initial settling/standing time was 300 s. The experiment was discharge signals up to 19 pC are initially noticed.
conducted at room temperature. The frequency boosting system used in the test consists of the step-
up transformer, the protective impedance, the coupling capacitor, the
2.5. Flashover voltage measurement test cable and the power frequency control console. The rated voltage of
the step-up transformer is 120 kV, the rated power is 10KVA, the protec-
The creeping flashover test was conducted by using an organic glass tion resistance is 10 k, and the capacitance of the coupling capacitor is
cell with the size of 120 mm. 80 mm. 50 mm as indicated in Fig. 3. The 800 pF.
needle to plane set up was used and the pressboard was kept horizon- The needle in the sample is adjusted to the required gap distance.
tally between the electrodes. The needle electrode with a tip radius of The voltages were enhanced at a rate of almost 0.5 kV/s, till the oscillo-
50 to 70 μm was put 30° to horizontal. The needle length was adjusted scope trigger display the PD signal and the applied voltage at this time
to allow the tip to be contacted pressboard firmly. The gap distance of as the PD start value. The PD detection device is disconnected, the volt-
20 mm and 30 mm was used between needle and L-shaped plane ages are raised at the same rate until a flashover occurs, and the applied
electrode. voltage at this time is recorded as surface flashover voltage. Every sam-
Two kinds of flashover tests were conducted. The first test was fo- ple was tested 6 times and the mean of these 6 values is calculated as the
cused to study the flashover traits of OIPs and NIPs under 50 Hz AC elec- PD voltage. In order to avoid the interface of the gas and impurities pro-
tric stresses. In this test, the HV electrode of the test cell is attached to duced by the surface flashover on the subsequent test, oil sample, press-
the AC supply transformer via 10 kΩ resistor to reduce the overcurrent board and the needle electrode were replaced after each test and
resulted by breakdowns. The rated voltage of the step-up transformer is standing time of 5 min was allowed to diminish the bubbles in the
120 kV, the rated power is 10kVA and the capacitance of coupling sample.
The second test was focused to study the flashover traits of OIPs and
NIPs under lightning impulse electric stresses. The experiment was car-
ried out under standard lightning impulse voltages 1.2 μs/50 μs. The im-
pulse producer was applied to generate the required standard lighting
impulses and to apply to the needle electrode. The oscilloscope was
used to record the current signals. The step by step voltage rise method
was used in the experiment. In the beginning of the test, a chosen value
of impulse voltages applied and then increased by a step of 2.5 kV until
flashover happens. The voltages were applied three times and if the
creeping flashover did not occur, the applied voltages were further in-
creased with a step of 2.5 kV until a creeping flashover did happen. Six
values of creeping flashover were obtained for every sample. The exper-
imental setup for this test is given in Fig. 4.

3. Experimental results and discussion

3.1. Color variation of the samples

There was not a noticeable variation in the color of samples after

aging at 55 °C but the significant changes in the colors of oil samples
were noticed aging at 130 °C. The color of the liquid sample turned
deeper and less transparent with the degree of aging. It can be found
that the color of the two types of fluids (pure oil and NFs) gradually
turn darker with the enhancement of aging duration. During early
Fig. 2. Test cell for LI BDV measurement [46]. stages of aging, color of samples was transparent and clear but it turns
704 M. Rafiq et al. / Journal of Molecular Liquids 284 (2019) 700–708

Protecon the aging of 30 days at 130 °C, the AC BD voltages of alumina NFs were
resistance 8% higher than pure oil.
Z
3.3. Positive LI breakdown voltage measurement of mineral oil and
Capacitor
nanofluids
voltage
divider
The LI BD voltages are tested according to IEC 60897 standard. The
gap distance used for positive LI BDV was 25 mm. The LI BDVs of aged
High voltage
oil and aged NFs were determined by utilizing the oil tank and the nee-
dle to sphere electrode arrangement. The positive LI BD voltages were
measured for oils and NFs aged for 0 days, 6 days, 18 days, 24 days
and 30 days. The oil tank and electrodes are immersed with the same
Z NFs. The oil tank is repleted with the NFs to confirm that there are no
bubbles. The needle electrode is replaced post every BD to assure iden-
tical exploratory conditions. Six BD values were collected for each NF to
validate repeatability. The mean value is considered as LI BDV. The re-
sults for positive LI BDVs for aged oil and NFs are presented in Fig. 8.
The positive LI voltage of NFs was lower than pure oil only after the
30 days of aging. This may be caused by agglomeration of NPs.
Workstaon
3.4. Lightning impulse creeping flashover voltages of OIP and NIP
Fig. 4. Experimental geometry for creeping flashover measurement.
The creeping flashover test for aged sample was conducted by using
into reddish brown after 30 days of aging (Fig. 5). The color variation in an organic glass cell with the size of 120 mm. 80 mm. 50 mm. The nee-
OIPs and NIPs after aging at 130 °C is also shown in Fig. 6. The color of dle to plane set up was used and pressboard was kept horizontally be-
the pressboard samples turned into darker after 30 days of aging. tween the electrodes. The needle electrode with a tip radius of 50 to
70 μm was put 30° to the horizontal. The needle length was adjusted
3.2. AC breakdown strength of aged mineral oil and nanofluid to allow the tip to be contacted by pressboard firmly. The gap distance
of 30 mm was used between needle and plane electrodes. The experi-
The AC BD voltage of two aged liquid samples (oil and nanofluid) ment was carried out under standard lightning impulse voltages 1.2
was measured during aging process by using 6801 AC BD tester. The μs/50 μs. The impulse producer was used to generate the standard light-
AC BD voltages were measured for oils and NFs aged for 0 days, ing impulses and applied to the needle electrode. The oscilloscope is
6 days, 18 days and 30 days. The AC BD voltage of the aged samples used to record the current signals. The step by step voltage rise method
was tested according to IEC 60156. The experiment uses brass ball is applied in the experiment. In the beginning of the test, a chosen value
caped electrodes with gap distance of 2.5 mm. The voltage elevating of impulse voltages applied and then enhanced by a step of 2.5 kV until
rate was 2 kV/s. The 5 min initial relieve time was used. A total 60 BD flashover happens. The voltages were applied three times and if the
values were attained for every sample and the average value is taken creeping flashover does not happen, the applied voltages are increased
as AC BDVs. The moisture content of the samples was 12–15 ppm. with a step of 2.5 kV until a creeping flashover did happen. Six values of
The AC breakdown voltage results are shown in Fig. 7.The AC BDVs creeping flashover were obtained for every sample. The results of posi-
for aged oil and aged NF gradually decreases as the aging time increases, tive creeping flashover for aged OIP and NIP are shown in Fig. 9. The re-
which is in accordance with the results presented in ref. [47]. The AC BD sults of aged NFs were higher than the aged MO throughout the aging
voltages of NFs were always higher throughout the aging process. After process at 130 °C.

Fig. 5. Aging of pure oil (left) and nanofluids (right) at 130 °C (after aging 0 days, 6 days, 18 days and 30 days).
 M. Rafiq et al. / Journal of Molecular Liquids 284 (2019) 700–708 705

Fig. 6. Aging of OIPs (upper) and NIPs (lower) at 130 °C (after aging 0 days, 6 days, 18 days and 30 days).

The results for negative creeping flashover for both OIP and NIP 3.5. AC creeping flashover voltages of OIP and NIP
aged at 130 °C are shown in Fig. 10. Both the positive and negative
flashover voltages for NIPs were higher than the OIPs through the The AC creeping flashover test was carried out to study the flashover
aging process. After the aging of 30 days, the negative creeping features of OIPs and NIPs under 50 Hz AC electric stress. In the test, the
flashover voltage of NIPs was almost 8% higher than OIPs at 130 °C HV electrode of the test cell is attached to the AC supply transformer via
after 30 days. 10 kΩ resistor to reduce the overcurrent resulted by breakdowns. The

Fig. 8. Positive lightning impulse breakdown strength of aged oil and aged NFs aged at 130
Fig. 7. AC breakdown strength of aged oil and aged NFs aged at 130 °C °C.
706 M. Rafiq et al. / Journal of Molecular Liquids 284 (2019) 700–708

voltage. In order to avoid the interface of the gas and impurities pro-
duced by the surface flashover on the subsequent test, oil sample, press-
board and the needle electrode were replaced after each test and
standing time of 5 min is allowed to diminish bubbles in the sample.
The AC creeping flashover for aged OIPs and NIPs is shown in Fig. 11.
The PDIV for NIPs was always higher than the PDIV of OIP throughout
the aging time duration of 30 days. The average AC creeping flashover
voltages of NIPs were 8% higher than the OIPs at 130 °C. After 30 days
of aging at 130 °C, the PDIV for aged NIPs were less than the OIPs
which may be caused by saturation of moisture absorption of NPs or ag-
glomeration of NPs.
The results for partial discharge inception voltage are shown in
Fig. 12. The results indicate that the PDIV for NIPs are always higher
than OIPs for all the aging duration range. The significant drop in PDIV
was seen after aging of 18 days at 130 °C, the PDIV for OIP was dropped
to 14.9 kV and PDIV for NIPs dropped to 16.2.

4. Modification mechanism

The Gouy-Chapman-Stern layer between oil and NPs is considered

the pathway where the charge would be transferred, which performs
a vital role in enhancing the dielectric strength of the NFs and NIPs.
The aging by-products are produced in the sample as a result of aging
at high temperature. The fraction of aging bi-product molecules will
Fig. 9. Positive creeping flashover voltage for OIP and NIP aged at 130 °C. be bound to the NP/oil interface and produce an ordered layer [48,49],
which will increase the lateral conductive path along the interface,
resulting in the creation of high local conductivity. This modification
rated voltage of the step-up transformer is 120 kV and the rated power of interfaces may be the primary factor for higher shallow trap density
is 10 kVA and the capacitance of coupling capacitor is 800 pF. LDS-6 PD in NFs with rise in aging time duration. It could be considered that shal-
equipment was applied to detect partial discharge (PD) signals. low traps are related to a charge hopping transport process in the sam-
The test cell without the needle electrode was found to be PD free up ple and charge carriers generated at high field in the NFs may be
to 60 kV with the lab background noise less than 3 pC. The needle in the captured and released briskly by these shallow traps. This brisk transfer
sample is adjusted to the required gap distance. The voltages were en- of charge carriers in the shallower traps lead to higher mobility in NFs
hanced at a rate of almost 0.5 kV/s, until the oscilloscope trigger display than oil under high electric field, which effectively reduce the accumu-
the PD signal and the applied voltage at this time is considered as the PD lation of space charge and electric field distortion produced by aging bi-
start value. The PD detection device is disconnected, the voltages are products and aging periods.
raised at the same rate until a flashover occurs, and the applied voltage The enhancement in creeping flashover voltages of aged NIPs as
at this time is recorded as surface flashover voltage. Every sample was compared to aged OIPs can be attributed to high shallow trap density in-
tested 6 times and the mean of these 6 values is calculated as the PD troduced by Al2O3 NPs. The interface around NPs contains of the bonded
layer and transition layer, which supply deep and shallow traps

Fig. 10. Negative creeping flashover voltage for OIP and NIP aged at 130 °C. Fig. 11. AC creeping flashover voltages of aged oil and NFs aged at 130 °C.
 M. Rafiq et al. / Journal of Molecular Liquids 284 (2019) 700–708 707

References
[1] M. Kohtoh, S. Kaneko, S. Okabe, T. Amimoto, Aging effect on electrical characteristics
of insulating oil in field transformer, IEEE Trans. Dielectr. Electr. Insul. 16 (2009).
[2] D. Peterchuck, A. Pahwa, Sensitivity of transformer's hottest-spot and equivalent
aging to selected parameters, IEEE Trans. Power Deliv. 17 (2002) 698–1706.
[3] T.A. Prevost, T.V. Oommen, Cellulose insulation in oil-filled power transformers: part
I—history and development, IEEE Electr. Insul. Mag. 22 (2006) 28–35.
[4] W.H. Bartley, Investigating transformer failure, Weidmann-ACTI Fifth Annual Tech-
nical Conference on New Diagnostic Concepts for Better Asset Management, No-
vember 13–15 (2006).
[5] EPRI Portfolio 2007- Transmission reliability and performance: 37.002, transformer
life extension. http://www.epri.com/portfolio/
[6] R. Karthik, T. Sree Renga Raja, R. Madavan, Enhancement of critical characteristics of
transformer oil using nanomaterials, Journal of Arab Science and Engineering 38
(2013) 2725–2733.
[7] Kanagesan S. Jesurani, R. Velmurugan, C. Kumar, T. Kalavani, “SOLGEL combustion
synthesis of giant dielectric CaCu3Ti4O12 nanopowder”, Journal of, Manufacturing
Engineering 5 (2) (2010) 124–128.
[8] Y.Z. Lv, Y.F. Du1, J.Q. Zhou, X.X. Li, M.T. Chen, C.R. Li, G.L. Wang, “Nanoparticle Effect
on Electrical Properties of Aged Mineral Oil Based Nanofluids” d1-106,cigre, 2012.
[9] Karthik.R, Sree Renga Raja. T, “Investigations of transformer oil characteristics”, IEEJ
Transactions Electr.Electron.7(4), 369–374, 2012.
[10] Huifei Jin, Thomas Andritsch, Ioannis A. Tsekmes, Roman Kochetov, Peter H.F.
Morshuis, Johan J. Smit, “Properties of mineral oil based silica nanofluids”, IEEE
Transactions on Dielectrics and Electrical Insulation, Vol. 21, no. 3, pp. 1100–1108,
2014.
[11] S. Han, Q. Li, C. Li, J. Yan, Electrical and mechanical properties of the oil-paper insu-
lation under stress of the hot spot temperature, IEEE Trans. Dielectr. Electr. Insul. 21
(1) (2014) 179–185.
[12] M. Rafiq, Y.Z. Lv, Y. Zhou, K.B. Ma, W. Wang, C.R. Li, Q. Wang, Use of vegetable oils as
transformer oils—A review, Renew. Sustain. Energy Rev. 52 (2015) 308–324.
[13] D.J.T. Hill, T.T. Le, M. Darveniza, T. Saha, A study of the degradation of cellulosic in-
sulation materials in a power transformer. Part III: degradation products of cellulose
Fig. 12. PDIV of aged oil and NFs aged at 130 °C. insulation paper, Polymer Degradation Stability 51 (2) (1996) 211–218.
[14] M. Rafiq, Y.Z. Lv, C.R. Li, A review on properties, opportunities, and challenges of
transformer oil-based nanofluids, J. Nanomater. 2016 (2016).
[15] Yuzhen Lv, Muhammad Rafiq, Tommaso Auletta, Li Chengrong, Bingliang Shan,
separately [50,51]. Taking into account the porous and loose structure of Study of Dielectric Breakdown Performance of Transformer Oil Based Magnetic
pressboards, the NPs primarily exist in free volume of cellulose matrix. Nanofluids, Energies 10 (7) (2017) 1025.
[16] S.V. Kulkarni, S.A. Khaparde, “Transformer Engineering, Design, Technology and Di-
The aging will produce the moisture and aging bi-products which will agnostics,” 2nd Edition, CRC press, Taylor and Francis group, New York, 2012.
deteriorate the insulation strength of the insulating system. The aging [17] S.A. Stigant, A.C. Franklin, The J and P Transformer Book: A Practical Technology of
bi-products will modify the interface and enhance the shallow trap den- the Power Transformer, 10th edition Wiley, New York, 1973.
[18] M. Rafiq et al, “Insulating and aging properties of transformer oil-based TiO2
sity in NIPs. It could be considered that shallow traps are associated to nanofluids,” 2014 Annual Report Conference on Electrical Insulation and Dielectric
charge hopping transport process in the NIPs and the fast electrons pro- Phenomena.
duced at high electric field in NIP could be captured by shallow traps [19] Hu-Zhi feng et al., “Thermal aging properties of transformer oil-based Ti02
nanofluids,” 2014 IEEE International Conference on Liquid Dielectrics, Bled,
and released quickly from these shallow traps. The brisk transfer of
Slovenia, June 30–July 3, 2014.
charge carriers in the shallower traps result in their higher mobility in [20] J. Li, Z.M. He, S. Grzybowski, Electrical aging lifetime model of oil-impregnated paper
NIPs as compared to OIPs in high electric field, which will largely dimin- under pulsating DC voltage influenced by temperature, IEEE Trans. Dielectr. Electr.
Insul 20 (6) (2013) 1992–1997.
ish the accumulation of space charge and electric field distortion. In the
[21] S.M. Strachan, S. Rudd, S. McArthur, M.D. Judd, S. Meijer, E. Gulski, Knowledge-based
meantime, the fast electrons may be transformed into slow moving diagnosis of partial discharges in power transformers, IEEE Trans. Dielectr. Electr.
electrons by frequent trapping and de-trapping in the shallow traps. Insul 15 (1) (2008) 259–268.
Therefore, charge transport in NIP is significantly enhanced and the ac- [22] S.-Q. Wang, G.-J. Zhang, J.L. Wei, S.-S. Yang, M. Dong, X.-B. Huang, Investigation on
dielectric response characteristics of thermally aged insulating pressboard in vac-
cumulation of space charge is restrained. uum and oilimpregnated ambient, IEEE Trans. Dielectr. Electr. Insul. 17 (6) (2010)
1853–1862.
5. Conclusion [23] A.K. Lokhanian, T.I. Morozova, G.Y. Shneider, V.V. Sokolov, V.M. Chornogotsky, “In-
ternal Insulation Failure Mechanisms of HV Equipment under Service Conditions”,
CIGRE Report 15-201, France, Paris, 2002.
In this paper, the same amount of insulating paperboard and copper [24] G. Yali, Research on power transformer aging analysis and character, Chongqing
plate were added to the pure oil and Al2O3 NFs for every sample. The ac- University. Chongqing, 2007.
[25] F. Yun, Characteristics and Mechanisms of Aging of Oil-paper Insulation in Power
celerated heat aging of the two kinds of transformer oil and paper com- Transformers, Chongqing University. Chongqing, 2007.
posite insulation was carried out for 30 days under the temperature of [26] H. Provencher, B. Noirhomme, E. David, Effect of aging on the dielectric properties of
130 °C. There was an obvious aging effect on the samples after aging insulation paper of power transformers, IEEE Electr. Insul. Conf. (2009) 473–477.
[27] S. Han, Q. Li, C. Li, J. Yan, Electrical and mechanical properties of the oil-paper insu-
at 130 °C temperature. The results showed that nanometer aluminum
lation under stress of the hot spot temperature, IEEE Trans. Dielectr. Electr. Insul. 21
dioxide can slow down the aging of transformer oil and paper compos- (1) (2014) 179–185.
ite insulation. The average AC and positive LI BDVs of NFs is 11% higher [28] D.J.T. Hill, T.T. Le, M. Darveniza, T. Saha, A study of the degradation of cellulosic in-
sulation materials in a power transformer. Part III: degradation products of cellulose
than pure MO. The average positive and negative creeping FOVs of NIPs
insulation paper, Polymer Degradation Stability 51 (2) (1996) 211–218.
were 6% higher than OIPs. Moreover, the average AC creeping flashover [29] L. Jian, Z. Huaiyu, J. Huachun, Power transformer insulation paper: its properties and
voltages of NIPs were 8% higher than the OIPs. insulation aging, China Pulp & Paper. Beijing 27 (5) (2008) 54–58.
[30] J.H. Li, W.R. Si, X. Yao, Y.M. Li, Measurement and simulation of partial discharge in oil
impregnated pressboard with an electrical aging process, Meas. Sci. Technol. 20
Acknowledgment (2009) 1–7.
[31] K. Temmen, Evaluation of surface changes in flat cavitiesdue to ageing by means of
The authors acknowledge financial support for this work from the phase-angle resolved partial discharge measurement, J. Phys. D: Appl. Phys. 33
(2000) 603–608.
National Natural Science Foundation of China under Contract Number [32] T. R. Blackburn, Z. Liu, R. Morrow, B. T. Phung, "Partial discharge development in a
51337003, and from the Chinese Ministry of Education Fundamental void and its effect on the material surface", IEEE 6th Int'l. Conf. Properties and Appli-
Research Funds for the Central Universities. cations of Dielectric Materials, 2000.
708 M. Rafiq et al. / Journal of Molecular Liquids 284 (2019) 700–708

[33] C. George, Z. Q. Xu, "Charge trapping and detrapping in polymeric materials", J. Appl. [44] IEC 156 International standard, insulating liquids – Determination of the breakdown
Phys., vol. 106, pp. 123707–1-5, 2009. voltage at power frequency – Test method, Second edition; 1995.
[34] V. Sokolov, Z. Berler, V. Rashkes, Effective methods of assessment of insulation sys- [45] “Methods for the determination of the lightning impulse breakdown voltage of in-
tem conditions in power transformers: a view based on practical experience, IEEE sulating liquids,” IEC 60897 (1987).
Electr. Insul. Conf. and Electr. Manufacturing & Coil Winding Conf., Cincinnati, OH, [46] Yuzhen Lv, Muhammad Rafiq, Chengrong Li and Bingliang Shan, “Study of dielectric
USA, pp. (1999) 659–667. breakdown performance of oil based magnetic nanofluids,”, Energies 10 (7) (2017).
[35] Y.Z. Lv, et al., Nanoparticle effects on creeping flashover characteristics of oil/press- [47] N. Yamagata, K. Miyagi, E. Oe, Diagnosis of thermal degradation for thermally
board interface, IEEE Transactions on Dielectrics and Electrical Insulation 21 (2) upgraded paper in mineral oil, International Conference on Condition Monitoring
(April 2014). and Diagnosis 2008, pp. 1000–1004.
[36] M. Chiesa, S.K. Das, Experimental investigation of the dielectric and cooling perfor- [48] T.M. Truskett, The subtleties of water in small spaces, PNAS. 100 (2003) 10139.
mance of colloidal suspensions in insulating media, Colloids and Surfaces A: Physi- [49] Yuzhen Lv, Yuefan Du, Chengrong Li, et.al. “TiO2 nanoparticle induced space charge
cochemical and Engineering Aspects 335 (2009) 88–97. decay in thermal aged transformer oil,” Applied Physics Letters, 2013,102:132902.
[37] Yang Xu, et al., Oxidation stability assessment of a vegetable transformer oil under [50] T. Tanaka, M. Kozako, N. Fuse, Y. Ohki, Proposal of a multi-core model for polymer
thermal aging, IEEE Trans. Dielectr. Electr. Insul. 21 (2) (2014) 683–692. nanocomposite dielectrics, IEEE Trans. Dielectr. Electr. Insul. 12 (2005) 669–681.
[38] Rongsheng Liu, Leif A.A. Pettersson, Tommaso Auletta, Olof Hjortstam, Fundamental [51] S.T. Li, G.L. Yin, G. Chen, J.Y. Li, S.N. Bai, L.S. Zhong, Y.X. Zhang, Q.Q. Lei, Short-term
research on the application of nano dielectrics to transformers, Annual Report Con- breakdown and long-term failure in nanodielectrics: a review, IEEE. Trans. Dielectr.
ference on Electrical Insulation and Dielectric Phenomena (2011) 423–427. Electr. Insul. 17 (2010) 1523–1535.
[39] E.C. Nsofor, Recent patents on NFs (nanoparticles in liquids) heat transfer, Recent
Patents on Mechanical Eng. 1 (3) (2008) 190–197.
Muhammad Rafiq was born in Gujrat, Pakistan. He earned his B. Sc. in electrical engineer-
[40] B. Berkovski, V. Bashtovoy, Magnetic Fluids and Applications Handbook, Begall
ing from University of Engineering and Technology, Pakistan in 2007. He received the MSc.
House Inc, New York, 1996.
degree in Electric Power Engineering from Chalmers University of Technology, Sweden in
[41] R.E. Hebner, Measurement of electrical breakdown in liquids, The Liquid State and
2010. He has been lecturer at University of Gujrat, Pakistan from 2010–2012. He received
Its Electrical Properties, Springer US 193 (1988) 519–537.
the Ph.D. degree in electrical engineering from North China Electric Power University in
[42] IEC60641-1:2007, “Specification for pressboard and presspaper for electrical pur-
2017. His research interests include high-voltage insulation and materials.
pose – part 1: Definition and general requirements”, 2007.
[43] IEC Technical Committee. IEC354-1991 “Loading guide for oil-immersed power
transformers[S]”.Gneva: IEC Publication, 2003.