246 K. Bandara et al.

: Understanding the Ageing Aspects of Natural Ester Based Insulation liquid in Power Transformer

Understanding the Ageing Aspects of Natural Ester

Based Insulation liquid in Power Transformer
Kapila Bandara, Chandima Ekanayake, Tapan Kumar Saha
School of Information Technology and Electrical Engineering
The University of Queensland
Brisbane, Australia, 4072
and Pratheep Kumar Annamalai
Australian Institute of Bioengineering & Nanotechnology
The University of Queensland
Brisbane, Australia, 4072

ABSTRACT
Cellulose based insulation materials and mineral oils have widely been used in liquid
filled transformers. Environmental and fire-safety concerns have accelerated the use of
Natural Esters (NE) as a sustainable alternative to mineral oil. There is a rapid growth
of their application for distribution level transformers. To ensure a safe and long-term
operation of the vegetable oil filled transformers, it is necessary to clarify the
knowledge gaps with respect to the ageing of vegetable oil impregnated cellulose and
insulation diagnostic methods. The condition monitoring of mineral oil filled
transformers has been well characterised, due to the numerous laboratory and field
studies conducted during the past several decades. The application of these available
condition monitoring techniques for NE filled transformers is yet to be validated as the
use of NE for power transformers is in the evaluation stage. In this paper the
performance of NE as insulation in transformers by evaluating the ageing of NE
impregnated cellulose pressboard is presented. For this purpose, an accelerated ageing
experiment has been carried out in sealed tubes at 120 ⁰C. The ageing of oil
impregnated pressboard has been characterised by decrease in the degree of
polymerisation (DP) of pressboard, and the formation of both 2-furfuraldehyde and
dissolved carbon dioxide (CO2) in the oil. The structural changes of cellulose have been
supported by Fourier Transform Infrared (FTIR) spectra. Moreover, the results have
been compared with cellulose insulation samples aged under similar conditions in
conventional mineral oil. Acidity value, dissipation factor and viscosity measurements
have been used to compare the stability of NE in a transformer operating environment
to mineral oil.
Index Terms — Cellulose, degree of polymerisation (DP), dissipation factor dissolved
gas analysis (DGA), Fourier transform infrared spectroscopy (FTIRS), 2-
furfuraldehyde, natural ester (NE).

of cellulosic materials, such as hemp Kraft paper, crepe Kraft

1 INTRODUCTION
CELLULOSE, which is a natural polymer from wood,
provides outstanding insulating properties at a reasonable cost
when impregnated with insulating oil. Therefore, cellulose
based insulation materials are commonly used in many high
voltage assets including power transformers. Different types
Manuscript received on 11 April 2015, in final form 27 August 2015, 
accepted 27 August 2015.  insulation. It is already accepted that the ageing rate of
paper and pressboard are used in liquid filled transformers. In
power transformers, up to several tonnes of cellulose is used
in form of pressboard and paper [1]. During operation,
cellulose insulation is subjected to irreversible degradation. It
reduces the ability to withstand high mechanical stresses;
which are the results of inrush currents through windings
under faulty conditions [2]. Heat, moisture and oxygen
accelerate the degradation/ageing process of cellulose

DOI 10.1109/TDEI.2015.004744
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 23, No. 1; February 2016
mineral oil impregnated cellulose doubles for every 6-8 ºC
increase in temperature [3]. During ageing, cellulose is
decomposed into several substances including furanic
compounds, acids, CO2 and CO. These by-products are partly
soluble in oil and are usually characterised as indicators for
cellulose ageing.
Oil in a transformer plays a dual role as insulation and as a
coolant. Mineral oils have been used in liquid immersed
transformers for more than 100 years. However, mineral oil
inherently has several disadvantages such as, its limited fire
resistance and its dependence on petroleum sources [4].
Moreover, it can contaminate soil and water-ways in a case of
leakage. In contrast, vegetable oils (NE) can overcome some
of these major drawbacks of mineral oils. Thus, the
application of vegetable oils in transformers is becoming more
popular among the utilities, especially due to the escalating
price of conventional oils and increasing the environmental
constraints from authorities. In Germany, NEs are considered
as non-hazard to water and an ultimately biodegradable
substance according to aquatic biodegradation tests [5]. The
fire point of NE based insulating liquids is greater than 300 ºC
[6]. Therefore, from a fire resistance perspective NE filled
transformers offer better performance than conventional
mineral oil filled transformers. According to section 450.23 of
U.S. National Electrical Code, NEs are defined as less
flammable fluids [5].
The chemical structure of NE is different from that of
mineral oil. Therefore, performance of cellulose based
insulation in NE filled transformers may be different from that
of the conventional mineral oil filled transformers. Liao et al.
[4] and Lijun et al. [7] have performed accelerated ageing
experiments to investigate the ageing behaviour of Kraft paper
in NE. In addition to the general techniques such as
determining of Degree of Polymerisation (DP) and analysing
of 2-furfuraldehyde, they have used scanning electron
microscopic analysis, FTIR spectrum analysis and X-ray
electron microscopic analysis to investigate the structural
change of cellulose paper after ageing. However, they have
not used these techniques at regular intervals during the
ageing process.
This study is aimed at better understanding of the ageing of
cellulose insulation in NE filled transformers. Ageing of
cellulose in NE is also compared with that in mineral oil.
Ageing status of pressboard is determined using DP. Further,
the concentrations of 2-furfuraldehyde and dissolved CO2 are
also examined to understand the relationship between these
parameters and the degradation of cellulose. In this study, the
application of FTIR spectra for characterising the ageing of
cellulose pressboard and dissolved ageing by-products in the
oil is extensively studied. Under similar ageing conditions,
dissolved gases in both NE and mineral oil are also analysed.
Acidity, dissipation factor and viscosity of oil samples are
measured at different ageing levels.
The chemistry of insulation materials including chemical
structure and ageing mechanism is broadly discussed in
section 2. It is followed by a description of detailed
experimental procedure. In section 4, the ageing performance
247
of NE impregnated pressboard insulation is determined with
DP and furanic compounds analysis. Applicability of FTIR
technique to investigate the structural changes of cellulose and
identify the ageing by-products are also examined in this
section. In section 5, ageing of NE and mineral oil is
compared using acidity value, viscosity and dissipation factor
measurement results. The results of DGA have also been
discussed in section 6. Moisture analysis of pressboard over
thermal ageing is presented in section 7.
2 INSULATION MATERIAL CHEMISTRY
2.1 CELLULOSE CHEMISTRY AND AGEING
MECHANISM
Kraft pulping process is used to produce electrically graded
Kraft paper with high concentration of cellulose [1].
Electrically graded Kraft paper or pressboard comprises of
about 90% of cellulose, 6-7% of hemicellulose and 3-4% of
lignin [8, 9]. Cellulose is a homopolymer of D-anhydrogluco-
se units (AGU) bonded together via C1-C4 glycosidic oxygen
linkage [10]. It is often characterised by the degree of
polymerisation (DP) which is average number of glucose
rings in a polymeric chain. Intramolecular hydrogen bonding
between adjacent glucose units imparts linearity in cellulose
polymeric chains. Intermolecular hydrogen bonding further
forms hierarchically arranged highly crystalline microfibrils in
the matrix of lignin and hemicellulose, which ultimately
develops into a fibre [10, 11]. Physically cellulose is a semi-
crystalline material in nature, which has both highly
crystalline nano-scale domains and amorphous regions. The
crystalline domains are formed by arranging cellulose chains
into a monoclinic lattice structure. Moreover, amorphous
domains in the cellulose micro-fibrils provide room for
moisture absorption. For insulation application, cellulose
produced via Kraft process is highly preferred because of its
higher DP and crystallinity of cellulose.
The ageing of insulation paper is mainly due to the
combined effect of pyrolysis, hydrolysis and oxidation.
Temperature, moisture and oxygen are the key factors, which
determine the rate of ageing. Pyrolysis is thermal
decomposition of carbonaceous materials in absence of
oxygen. In case of pyrolysis reaction at higher temperature,
cellulose molecules firstly decompose into glucose units
followed by formation of levoglucosan, which is the key
intermediate product. The levoglucosan further decomposes
into 2-furfuraldehyde through dehydration [11, 12].
Hydrolysis of cellulose paper is catalysed by H+ ions,
which is derived by the dissociation of carboxyl acid in water
[3]. In acid catalysed hydrolysis, one molecule of water is
consumed in every linkage cleave and three water molecules
are released. The primary products of hydrolysis are furanic
compounds namely 2-furfuraldehyde (2-FAL), 5-
Hydroxymethyl 2-furfural (5-HMF) and 2-Furfuryl alcohol
(2-FOL) [2, 11, 13]. Among them, 2-FAL is the most stable
component. Therefore, it is considered as one of chemical
indicators of paper degradation. 5-HMF and 2-FOL are
unstable and decompose further into laevulinic and formic
acids [2]. These acids enhance the acidic hydrolytic
248 K. Bandara et al.: Understanding the Ageing Aspects of Natural Ester Based Insulation liquid in Power Transformer
degradation of the paper. Oxidative degradation of cellulose
paper mainly yields CO2 and water. Moreover, furanic
compounds, and carboxyl acids are also formed through
oxidative degradation of cellulose paper [14, 15]. In oil,
oxidation of cellulose paper can also be catalysed by the
hydroxyl radicals, which are produced by the decomposition
of hydrogen peroxide.
Rupturing of 1, 4-β glycosidic bonds in cellulose molecules
results in formation of methanol (CH3OH) [16, 17]. Jalbert et
al have identified the existence of strong correlation between
the rupturing of 1, 4-β glycosidic bonds of cellulose polymer
and the production of CH3OH from chopped chains [16, 18].
In the case of oil-paper system in a transformer, some portion
of methanol produced by the degradation of paper based
insulation dissolves in oil. Jalbert et al. [18] have
experimentally confirmed that CH3OH is a stable substance in
a transformer operating environment. Dissolved CH3OH in
transformer oil can be assessed using headspace gas
chromatography/mass spectrometry (HSGC/MS) technique.
This technique is generally used to identify the volatile by-
products dissolved in oil. This implies that CH3OH dissolved
in oil is possible to use as an indicator of paper degradation in
transformers.
2.2 CHEMISTRY OF MINERAL OIL AND
OXIDATION
Chemically, mineral oil is a complex mixture of several
hydrocarbon molecules mainly having paraffinic, aromatic
and naphthenic structures [19]. Even trace quantities of
hydrocarbon molecules, which have olefinic structures, are
not preferred for insulating liquid, since they are prone to
oxidation. It is generally accepted that molecules with
unsaturated carbon double bonds show inferior chemical
stability (oxidative susceptible). Aromatic molecules also
have alternating configuration of carbon double bond in their
cyclic structure. However, mono- and di-aromatic compounds
show surprisingly better oxidation stability [19]. On the other
hand, by-products of oxidation of some aromatic compounds
accelerate the oxidation of oil. Thus, an optimum
concentration of residual aromatic compounds (5-10%) are
recommended to achieve optimal oxidation stability [19, 20].
Figure 1. Mechanism of oil oxidation [19].
Oxidation is the major degradation mechanism of
hydrocarbon oils and the rate of oxidation exponentially
increases with temperature (the reaction velocity doubles for
every 8-9 ⁰C rise in temperature) [19, 20]. Oxidation
decomposes the hydrocarbon molecules into other substances
including hydro-peroxide, alcohol, aldehyde, ketones and
esters. This is considered as a progressive reaction, which
starts by the formation of radicals due to the ejection of a
hydrogen atom from a healthy hydrocarbon molecule. These
radicals further react with oxygen and produce highly reactive
peroxide molecules. Metallic soaps, which are produced
through corrosive oxidation of Copper, Zinc and ferrous
materials catalyse the oxidation of oil [19]. Indeed, combining
of large free radicals results in the formation of non-soluble
substances, which have molecular weights of approximately
500-600. These intermediate products are deposited on
winding insulation as sludge due to the high electromagnetic
fields at the windings [21]. Figure 1 illustrates the typical
oxidation reaction of mineral oil.
2.3 CHEMISTRY OF NE AND AGEING
NEs are mainly derived from vegetable oils, which
primarily consist of triglyceride. A triglyceride molecule has
three fatty acid molecules, which are chemically bonded to a
glycerine backbone. These fatty acid groups may have 4-22
carbon atoms and 0-4 unsaturated carbon double bonds. A
high concentration of saturated fatty acids causes the oil to
freeze below room temperature. Moreover, a high
unsaturation level may help to keep the oil in liquid form at
low temperatures (i.e. -15 ⁰C to -30 ⁰C) [22]. However,
unsaturated fatty acids lead to a higher oxidation of oil. The
relative oxidation susceptibility of saturated: mono: di: tri:
unsaturated C-18 (18 carbon atoms) triglycerides is roughly
1:10:100:200 [22]. Therefore, maintaining a good balance
between the degree of saturation and unsaturation is
important, when vegetable oil is used as an insulating liquid.
NE used in this study is made of high oleic sunflower oil.
High oleic sunflower oil contains 5% of saturated stearic acid
(C:18:0), 86% of monounsaturated oleic acids (C18:1), 6%
linoleic acid (C:18:2) and 0.1% of linolenic (C18:3) [23]. The
high percentage of oleic acid composition in the based fluid
provides better oxidation stability to NE.
Similar to mineral oil, ageing of vegetable oil is mainly due
to oxidation and this is a progressive reaction as shown in
Figure 1. The initial step of vegetable oil oxidation is the
formation of free radicals by removing a weakly bonded
hydrogen atom located next to a carbon double bond of a fatty
acid group [24]. Therefore, carbon double bonds sites in
vegetable oil provide a more susceptible environment for
oxidation.
In the case of NEs, hydrolysis is another degradation
reaction, which concurrently occurs with oxidation. This is an
autocatalytic reaction, because free fatty acid molecules
themselves accelerate the hydrolysis reaction [25]. There are
three major stages in the hydrolysis reaction in vegetable oil,
which are also reversible. The corresponding chemical
reactions are given below [26].
IEEE Transactions on Dielectrics and Electrical Insulation
Figure 2. Hydrolysis reaction of NE.
3 EXPERIMENTAL PROCEDURE
3.1 MATERIALS
One type of sunflower oil based NE and non-inhibited
naphthenic based mineral oil were used in this experiment.
The pressboard samples (150x15x1.5 mm) used for this study
were normal transformer insulation pressboard from
Weidmann Electrical technology.
3.2 EXPERIMENTAL SETUP AND AGEING
PROCEDURE
Firstly, the pressboard samples were dried under a very
high vacuum level at 65 ⁰C for 72 h. Then the temperature
was increased up to 95 ⁰C and vacuum drying was continued
for another 24 h. Secondly, the degassed oil (mineral or NE)
was infused into the container, while keeping it under
vacuum. The pressboard samples were then impregnated
under vacuum condition at 60 ⁰C for 7 days.
For the accelerated ageing experiments, oil impregnated
pressboard samples, copper and degassed mineral oil or NE
were inserted into a set of specially designed glass tubes. The
mass ratio between oil and pressboard in these ageing tubes
was 10:1 (175 cm3 of oil and 16 g of pressboard). Air in the
tube headspace was pumped out until the pressure decreased
to a level between 10-20 kPa. Then, each test tube was sealed
with dry nitrogen headspace (pressure ≈110 kPa). A similar
number of glass bottles were prepared with degassed oil only.
The initial conditions including the average DP value of
pressboard, moisture in pressboard and moisture content in oil
are reported in Table 1. All samples with the same oil were
prepared under similar conditions to maintain similar initial
conditions of pressboard and oil in all tubes.
Table 1. Initial condition of pressboard and oil samples used for ageing.
Impregnated oil Moisture in Moisture in
Pressboard (%) oil(PPM)
Mineral oil 0.5 1 1281
NE 0.7 35
Finally, for accelerated ageing all the bottles were placed in
an aluminium heater block as shown in Figure 3.Temperature
of the heater block was adjusted to 120 ⁰C. Temperature
hysteresis of the heater block was ±2 ⁰C. Oil and pressboard
samples were taken for analysis after 384 h, 840 h, 1280 h,
1960 h and 2800 h of ageing. The oil samples were indexed as
Mxxh, MPxxh, Nxxh and NPxxh, where M and N represent
oil type (mineral or NE), P represents sample aged with
pressboard and xxh indicates the number of hours aged.
Further MIN_P and NE_P are used to indicate the group of
mineral oil and NE oil samples aged with pressboard
respectively.
Vol. 23, No. 1; February 2016 249
The ageing of the pressboard insulation was mainly
determined by measuring the DP value. Furanic compounds
and dissolved gases in oil were also analysed to verify the
ageing condition of pressboard further. FTIR spectra were
used to determine the structural changes of cellulose due to
ageing. The Attenuated Total Reflection (ATR) mode
technique was used in FTIR measurements.
Figure 3. Glass tubes in aluminium heater block during ageing.
Degradation of oil was mainly determined by the
dissipation factor (tan δ), viscosity and acidity value. A
frequency response analyser (IDA 200) was used to measure
the dissipation factor.
4 DEGRADATION OF PRESSBOARD
4.1 AGEING PERFORMANCE OF CELLULOSE
Figure 4a shows the percentage reduction of measured
average DP value of pressboard insulation samples aged in
mineral oil and NE over the ageing period. In this study, the
test method given in IEC standard 60450 – 2004 is used to
measure the average viscometric degree of polymerisation of
new and aged pressboard. In this method after the pressboard
has been cleaned, it is stripped off to provide enough sample
size for the DP analysis. Then, the analysis is performed for
two different pressboard samples from the same batch and the
average is reported.
Initial
DP value
1360
Figure 4. (a) Measured DP of pressboard samples aged in mineral oil and
NE (b) Modelled and measured DP (MIN_IP and NE_IP denotes the paper
samples impregnated in mineral oil and natural ester respectively).
As shown in Figure 4a, in the first 384 h of ageing the DP
of the pressboard has drastically decreased regardless to the
type of oil. Weak links existing in polymer chains are chopped
easily due thermal stress at a high temperature, which would
account for the fast initial drop in DP [4, 11, 27]. Further, the
amorphous region of cellulose degrades more rapidly than the
250 K. Bandara et al.: Understanding the Ageing Aspects of Natural Ester Based Insulation liquid in Power Transformer

crystalline region, which would also support the rapid initial Table 2. Modelled parameters of equation (1) for ageing at 120 ºC.
ageing [11]. This may be due to fact that the majority of water Oil type k2*10-4(h-1) k10*10-7(h) R2
in cellulose sits in the amorphous region [28] and it reduces Mineral 1.1116 4.49 0.972
the degree or ordering in the amorphous region. It has been NE 5.348 7.4 0.966
noticed that pressboard samples immersed in NE have aged
slightly faster than the pressboard immersed in mineral oil 4.2 ANALYSIS OF FURANIC COMPOUNDS
during the first 1280 h of ageing. During this time period, the
Figure 5 depicts the variation of dissolved 2-FAL
percentage reduction of DP of pressboard in NE is 53%
concentration in both NE and mineral oil over thermal ageing.
whereas in mineral oil it is only 42%. On the other hand, from
The 2-FAL concentration in NE increases sharply during the
1280 h to 2800 h of ageing the DP value of the pressboard
first 1280 h and then remains fairly constant. On the other
sample in NE decreased only by 6%. During the same time
hand, the 2-FAL concentration in mineral oil continuously
period, the DP of pressboard in mineral oil has dropped by
increases throughout the ageing period. Overall, dissolved 2-
18%.
FAL content found in this study is fairly low. It is accepted
The initial moisture contents of the pressboard sample aged that significant increase in 2-FAL concentration occurs for DP
in mineral oil and NE were 0.5% and 0.7% respectively. This value lower than 400 [32]. DP value of pressboard either aged
would partly account for the fast initial ageing rate in NE. in mineral or NE has not reached level of 400 and this could
Further, it has been reported in previous research that be a reason for having low 2-FAL concentration.
pressboard immersed in NE ages slightly faster than that of in
Distribution of 2-FAL between pressboard and oil is a
mineral oil during the first stage of ageing [4]. After 2800 h,
complex phenomenon. Moisture, level of oil degradation
the percentage reduction in DP of pressboard aged in both
(acidity and dissipation factor) and temperature have shown
types of oils is similar. On the other hand, between 1280 h and
influence on the formation and partition of 2-FAL between
2800 h the average rate of ageing for pressboard in mineral oil
pressboard and oil [33-36]. Moreover, significantly low
is three times greater than that of in NE. Therefore, one can
concentration of 2-FAL can be found in inhibited oil
assume that time taken by the pressboard in NE to reach the
compared to non-inhibited oil. However, besides to above
end of life DP will be higher than that of pressboard in
facts, it is accepted that dissolved 2-FAL concentration in oil
mineral oil.
provides a clear indication about paper degradation kinetics in
It is very clear that after rapid degradation during the initial transformers. Further, IEC 61198 provides basic guidelines
period of ageing (0-1280 h), pressboard in NE degrades much for analysing the 2-FAL measured in transformer oil.
slower than that in mineral oil. This advantage of NE over Different empirical formulae have also been proposed to
conventional mineral oil is mainly due to hydrophilic nature establish a relationship between dissolved 2-FAL and DP
of NE [29] and hydrolysis reaction occurs in NE [4]. The value of paper insulation in transformers [37, 38].
hydrophilic characteristics and hydrolysis reaction in NE
attract some of the water molecules in the pressboard and
eventually reduce the potential damage to cellulose from
water.
The Emsley model given in equation (1) [30] is widely used
to characterise the ageing of cellulose with time [4, 7, 30, 31].
Ageing behaviour of the pressboard samples in both mineral
oil and NE is modelled using Emsley model and results are
shown in Figure 4b. Table 2 provides the corresponding
modelled parameters.

1 1 k
  10 (1  exp( k 2 t ) (1)
DPt DP0 k2

Where, k2 and k10 are temperature dependent constants and t
represents the ageing time. DP0 and DPt represent the DP
value before and after ageing respectively.
In both cases, the modelled curves show exactly fitting with
measured data with R2> 0.95. The value of parameter k10 of
NE is higher than that of mineral oil. It indicates fast initial
ageing rate of pressboard in NE. On the other hand, higher
value of k2 of NE designates the slower ageing rate of
pressboard in NE than that in mineral oil during the last stage
of ageing experiment. It implies that ageing of pressboard in
NE and mineral oil can be exactly explained with second
order kinetic model as (1).
Figure 5. 2-FAL concentration in aged mineral and NE oil.
Based on above facts, one can assume that the observed
trend of 2-FAL in both type of oils may represent the ageing
behaviour of pressboard. Moreover, fairly constant 2-FAL
concentration found in NE (1280 h-2800 h) may represent the
slow ageing rate of pressboard in NE during this time period.
On the other hand, acidity value of NE markedly increases
during the last 1520 h of ageing from 0.25 mg KOH/g to 0.6
mg KOH/g (refer Figure 10 for more information). During the
same period 2-FAL concentration in NE remains fairly
constant. Thus, one can also assume that the increase of
acidity level of NE may have influence on partition of 2-FAL
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 23, No. 1; February 2016 251

between pressboard and NE. Moreover, slow degradation of
pressboard in NE does not solely contribute to the observed
trend of 2-FAL during the last stage of ageing experiment.
However, for a conclusive interpretation further investigation
will be required to understand the influence of acidity level on
partition of 2-FAL between oil and pressboard.
Table 3. 5-HMF concentration in oil over ageing time.
5-HMF concentration (ppm)
Oil type
384 h 840 h 1280 h 1960 h 2800 h
MIN_P 0 0 0 0.01 0.04
NE_P 0 0.42 0.64 0.38 0 C-H groups. As shown in these figures, there are two peaks at
As shown in Table 3, in addition to 2-FAL a significant
quantity of 5-HMF has also been detected in NP840h,
NP1280h, NP1960h, MP1960h and MP2800h. In fact, the
concentration of 5-HMF in MP2800h is much higher than that
in MP1960h. On the other hand, the concentration of 5-HMF
in NE drops to zero after 2800 h of ageing once reaching its
maximum value during the first 1280 h of ageing. The
presence of 5-HMF may be considered as a factor, which
indicates the contribution of water to degradation of cellulose
[12]. Hence, one can hypothesise that the hydrolytic
degradation of pressboard in mineral oil starts to be
pronounced after the pressboard reaches midlife, whereas in
NE it starts to slow down with ageing.
reduction of molecular weight (DP). The hydrolytic and
oxidative degradation of cellulose can break the ether linkage
and reduce the number of H-bonds per polymer chain. In the
case of NE, possible esterification of fatty acid with cellulose
can also result in the reduction of the H-bond. Even though
water molecules also have hydroxyl groups, the observed
peaks are not attributed to water. The peaks at 3200 cm-1 and
3600 cm-1 in FTIR spectra of cellulose can be assigned for
strongly bonded and loosely bonded water molecules
respectively [40].
Figures 6b and 6e illustrate the variations of FTIR spectra
around 2900 cm-1, which represent the stretching vibration of
2925 cm-1 and 2850 cm-1 ,which increase with ageing. This
may be attributed to the adsorption of both mineral and NE
oils to the pressboard surface during ageing. In addition, upon
ageing the radicals generated on the cellulose surface may
facilitate to physical attachment of low molecular weight oil
molecules. This mechanism leads to a formation of a layer of
hydrocarbon lamination on the cellulose surface and increase
the intensity corresponding to C-H vibration. This process is
pronounced with NE due to esterification. Moreover,
degradation of cellulose may form more reactive sites and
enhances this process.

4.3 ANALYSIS THE STRUCTURAL CHANGES OF

CELLULOSE OVER AGEING WITH FTIR
The change in chemical structure of pressboard during
ageing in both types of oils has been monitored through FTIR
spectroscopy. It provides precise information relevant to
functional groups changes of specific molecular structures.
Here, we have used Attenuated Total Reflectance (ATR)
mode to acquire the FTIR spectra. The penetration depth of
light into the sample is about 0.5-3 µm in this technique.
Therefore, it characterises the surface properties of each
sample under investigation. In this case, FTIR spectra on the
outer surface of new and aged pressboard have been observed.
Small samples required for FTIR measurement are prepared
from bigger pressboard specimens taken from the ageing tube.
Reproducibility of results has been confirmed by taking two
consecutive FTIR measurements on each sample.
Figure 6 compares the FTIR spectra of mineral oil
impregnated pressboard (Figures 6a-6c) with that of NE
impregnated pressboard (Figures 6d-6f) in different spans of
wave number. Figures 6a and 6d are associated with the
hydroxyl group region of the spectra. In general, a broad peak
with the maximum located close to 3340 cm-1 can be assigned
to stretching vibrations of hydroxyl groups. It is a typical
characteristic of cellulose [39]. As shown in Figures 6a and Figure 6. FTIR spectrums of pressboard insulation samples at different
ageing status (a) –(c) in mineral oil (d)-(f) in NE
6d, it is clear that the intensity of absorbance peaks located
close to 3325 cm-1 and 3275 cm-1 have decreased with ageing,
this effect is more significant in NE (Figure 6d). This Figures 6c and 6f depict the FTIR spectra corresponding to
indicates the reduction of inter and intramolecular hydrogen cellulose skeleton oscillation [41]. The peak around 1160 cm-1,
bonding on the cellulose surface due to faster degradation or which corresponds to C-O-C stretching vibration, has a
other chemical changes. Further, it can be attributed to the tendency to increase with ageing in NE, whereas no significant
252 K. Bandara et al.: Understanding the Ageing Aspects of Natural Ester Based Insulation liquid in Power Transformer

change observed in pressboard aged in mineral oil. The

absorbance peak around 1050 cm-1, which is due to the C-O
valence vibration, has decreased after 1280 h of ageing [39].
The drop is more significant for samples aged in NE. This may
be caused by the complex chemical reactions occurring
between the cellulose surface and fatty acid in vegetable oil.
The influence from each layer within a pressboard has been
considered in FTIR measurements. FTIR spectra on the first
three layers of 2800 h aged pressboard samples have been
observed and the results are reported in Figure 7. It has been
noticed that there is no significant difference between FTIR
spectra of each different layer of aged pressboard. It confirms
the homogenous ageing throughout the sample. It is
recommended not to compare these results with FTIR
measurement results of other sample, as this sample has been
stored in an open environment for more than 300 days after
finished the ageing before the measurement. Figure 8. FTIR spectra at carbonyl region upon ageing for (a) Pressboard in
mineral oil, (b) Mineral oil, (c) pressboard in NE (d) NE.

Figures 8b and 8d compare the FTIR spectra of mineral oil

and NE at different ageing conditions. Figure 8b shows two
peaks at 1750 cm-1 and around 1600 cm-1. The former peak is
due to dissolved carboxyl or aldehyde compounds and
aromatic composition of mineral oil may also responsible for
the latter peak. None of these peaks appears in FTIR spectra
of NE, but there is a peak at 1650 cm-1, which is not in FTIR
spectra of mineral oil. This peak is assigned for angle
vibration of water molecules [40]. The presence of this peak
provides an indication about the high moisture affinity of NE
compared to mineral oil.
4.5 CELLULOSE ESTERIFICATION
Figure 7. FTIR spectra of three different layers of 2880 h aged (in NE)
pressboard. Hydrolysis of triglyceride molecules produces long-chain
fatty acids. In general, carboxyl acid molecules (fatty acids in
4.4 THE APPLICABILITY OF FTIR TECHNIQUE this case) can be esterified to C-6 hydroxyl group in cellulose
FOR IDENTIFYING AGEING BY-PRODUCTS
[44, 45]. This phenomenon has been verified with X-ray
Dominant ageing mechanisms of pressboard depend highly
on the working temperature. Around 120 ºC, the most likely photoelectron spectroscopy and FTIR techniques in [4, 44].
ageing mechanisms of cellulose are hydrolysis and oxidation Furthermore, using molecular modelling techniques, it has
[42]. These degradation processes produce several carbonyl been stated that the fatty acid molecules are distributed in
and carboxyl compounds. Peaks corresponding to these by- parallel to the cellulose surface due to multi-hydroxyl
products appear in the wave band (1550-1800 cm-1). construction of cellulose.
Figures 8a and 8c present the FTIR spectra of the
pressboard aged with mineral oil (1800-1500 cm-1) and in NE
(1700-1500 cm-1) respectively (this wave bands were selected
to highlight the less pronounced peaks associated with ageing
by-products). As displayed in Figure 8a, there are three
absorbance peaks around 1750 cm-1, 1650 cm-1 and 1600 cm-1.
The latter two peaks also appear in FTIR spectra shown in
Figure 8c. These two peaks are due to the carbonyl compound
associated with cellulose. Therefore, peaks around 1600 cm-1
and 1650 cm-1 should be increasing with ageing. However,
one can observe that the intensity of the last two peaks
decreases with ageing. Reduction of these two peaks can be
assigned to opening of weak 1, 4-β glycosidic bonds [43]. The
peak at 1750 cm-1 in Figure 8a is evidently due to the
formation of carboxyl or aldehyde compounds. However, it is Figure 9. Development of ester peak due to esterification of pressboard
difficult to correlate this peak with ageing. insulation samples aged in NE.
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 23, No. 1; February 2016 253

In this study, esterification of fatty acids with cellulose has

been characterised by analysing the FTIR spectra of
pressboard samples aged in NE at different ageing status. As
shown in Figure 9, the intensity of the peak at 1746 cm-1,
which can be assigned for ester groups, has increased upon
ageing. In contrast the peak at 1746 cm-1 in the FTIR spectra
of the pressboard in mineral oil (Figure 8a) shows tendency to
decrease upon ageing. Thus, this peak confirms the
esterification of fatty acids with the pressboard aged in NE. In
addition, the rapid reduction of peaks at 3325 and 3275 cm-1
(Figures 6a and 6d) of pressboard in NE compared to that of
pressboard in mineral oil also provides indirect evidence for
the physical and chemical attachments of fatty acids on the
cellulose surface. Esterification of long chain fatty acids with
Figure 10. Change of acidity value of oil over ageing period.
cellulose may protect the cellulose insulation in two ways: by
strengthening the C-O and C-C bonds in cellulose and 5.2 DISSIPATION FACTOR AND VISCOSITY
forming a barrier for water ingress into cellulose [44].
Figure 11a compares the dissipation factor (tan) measured
at 55 °C of mineral and NE oil samples. At higher
5 AGEING OF OIL
temperatures, current for a given voltage is higher than that at
5.1 VARIATION OF ACIDITY OF OIL OVER AGEING room temperature. Therefore the accuracy of the
Figure 10 compares the total acidity value of mineral and measurements is high at high temperature. On the other hand,
NE oil samples measured after different ageing periods. The the influence from impurities comes from outside (oil
total acidity value in mineral oil samples aged with pressboard handling) could be significant at high temperature. Thus, in
and copper increases from 0 to 0.01 during the first 384 h and this study a moderate temperature (55 °C) to measure the
then remains at the same value during the rest of the period. dissipation factor has been selected. Measurements were
The acidity of the oil sample aged only with mineral oil is performed using an IDA 200 insulation analyser under an
zero for all the instances except at 384 h. As a result of having applied voltage of 50 V across 2 mm oil gap in a three
three fatty acid groups in molecular structure of NE, it always electrode test cell.
shows higher acidity value than mineral oil.
According to Figure 10, the acidity value of NE is
continuously increasing with ageing time, particularly in NE
aged with pressboard. Indeed, the increase during 840 h to
2800 h is many times greater than that of during 0-840 h. The
main reason that may cause this behaviour is production of
more free fatty acids in NE through hydrolysis reaction as
explained in section 2.3.
When the trend for acidity value of NE_P is compared with
its dissipation factor shown in Figure 11a, it can clearly be
perceived that dissipation factor remains at a fairly constant
value while the acidity level increases with ageing (from 840- Figure 11. Change of dissipation factor and viscosity of oil over ageing.
2800 h). The high molecular acids derived through hydrolysis
As shown in Figure 11a in all cases the dissipation factor
reaction in NE do not dissociate as low molecular acids. Thus,
has changed significantly between un-aged and 1280 h of
the acidity of an ester increases while the dissipation factor ageing followed by a relatively low variation of it. The initial
remains fairly constant during hydrolytic degradation of NE rise of dissipation factor may be due to the formation of acids
[46]. and other polar substances due to ageing of oil and pressboard
An increase of acidity in oil enhances the paper degradation [21].
via acid catalyst hydrolysis. However, acids which cannot In this study headspace of the test tube was vacuumed only
provide H+ ion through dissociation do not support up to 10-20 kPa before sealed with dry nitrogen. Moreover,
degradation of cellulose via hydrolysis [3]. NE consists of the oil used in test tubes was exposed to air during the
high molecular acids, which are not dissociated as low sampling period (about 5 minutes). Thus, one can expect that
molecular acid produced by mineral oil. Moreover, low a considerable amount of oxygen could have dissolved in oil
molecular acids are readily absorbed by cellulose but high (diffusion coefficient of oxygen in fluid hydrocarbon at 25 ⁰C
molecular acids are not absorbed by cellulose [47]. This may is about 1.8-2.6 cm2/s [48]) and it may contribute to the initial
be a reason for the pressboard in NE aged very slowly after oxidation of oil. This may be one reason for increasing
1280 h even with higher acid concentration in NE. dissipation factor during the first 384 h of ageing. Moreover,
254 K. Bandara et al.: Understanding the Ageing Aspects of Natural Ester Based Insulation liquid in Power Transformer

the overall moisture content in oil has also been increased stages. It is clear that CO2 dissolved in mineral oil and NE
during first 840 h of ageing (particularly oil aged standalone). when they aged alone is relatively small compared to that in
This may also cause the initial rise of dissipation factor. The oil samples aged with pressboard. As displayed in Figure 12a,
increase of the dissipation factor of oil sample aged with the dissolved CO2 level in NE aged with pressboard sharply
pressboard is less than that of oil samples aged stand alone in increases during the first 1280 h and remains at a fairly
first 384 h. The ageing of pressboard may also consume constant level during the rest of the ageing period. In the case
dissolved oxygen in oil and retard the oxidation reaction in the of mineral oil sample aged with pressboard, the CO2 level
oil. This may be a reason for observed different trends of continuously grows over the whole ageing period. During the
dissipation factor between oil aged standalone and with first 1280 h dissolved CO2 in NE_P increases quicker than
pressboard during the initial stage of ageing. that in MIN_P. This information provides indirect evidence on
the fast initial ageing rate of pressboard in NE. Overall CO
Oxidation of oil produces radicals, which can combine with
concentration in both MIN_P and NE_P increased throughout
ionic particles in the oil. This phenomenon causes to increase
the ageing process as shown in Figure 12b.
the physical size of ions; that consequently decreases ionic
mobility. When this effect is dominant, dissipation factor
starts to decrease. This could be a reason for decreasing the
dissipation factor of oil samples aged standalone after the first
384 h of ageing. When pressboards exist in the oil, radicals
and ionic particles are adsorbed by pressboard surfaces. As a
result there is no big change in dissipation factor due to above
mention phenomenon in both MIN_P and NE_P. On the other
hand, since there is no continuous oxygen supply, it can
expect only limited oxidation in the oil until dissolved oxygen
is consumed. Thus, one can expect fairly constant dissipation
factor after initial increment as observed in both MIN_P and
NE_P. In fact, results of viscosity measurements agree with
this hypothesis, because the viscosity has not changed during
the whole ageing period. It confirms that no severe oil
oxidation has occurred during ageing. Figure 11b represents
the viscosity of the same oil samples determined at 40 ⁰C.
Moreover, dissipation factor of tested NE oil is greater than
that of mineral oil under similar ageing conditions. This is due
to a higher dielectric constant associated with NE. There is an
inverse relationship between dielectric constant of continuum
separating ions and Coulomb force attraction between
oppositely charged ions [49]. Therefore, larger dielectric
constant of NE enhances the dissociation of ionic impurities
leading to a higher dissipation factor.

6 DISOLVED GAS ANALYSIS

6.1 CHANGE OF CO2 AND CO CONCENTRATION
OVER AGEING
Thermal and electrical stresses in a transformer operating
environment decompose insulating oil into low molecular
hydrocarbon gases (H2, CH4, C2H4, C2H6, C2H2) and carbon
oxide (CO and CO2). However, it has been reported that
different chemical compositions of NE results in different
gassing behaviour compared to mineral oil [50-52]. In power
transformers, it is accepted that a large quantity of CO and
CO2 dissolved in oil comes from paper degradation.
Moreover, when the dissolved CO and CO2 concentrations in
oil are excessively high, the ratio of CO2/CO is used as an
indicator for thermal degradation of paper based insulation
[53]. Thus, these two gases can be considered as an indicator
for analysing the paper degradation kinetics [17, 54].
Figure 12 compares the CO2 and CO concentrations
dissolved in mineral oil and NE samples at different ageing
Figure 12. Variation of CO2 and CO concentration dissolved in mineral oil
and NE samples with ageing.
In order to understand the relationship between the ageing
of pressboard and dissolved carbon dioxide, CO2
concentration is plotted against the average number of chain
scission of the corresponding pressboard sample. The results
are shown in Figure 13a. The average number of chain
scissions is calculated using equation (2) [18]. In both cases,
during the initial stage of ageing a linear increase of dissolved
CO2 level with a number of chain scissions is observed.
However, in NE observed linear relationship has been
diminished after 1280 h. After 1280 h of ageing acidity level
of NE_P has drastically increased (Figure 10). Thus, there
may be an influence from increase of acidity level on CO2 gas
solubility in NE. However, for conclusive interpretation
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 23, No. 1; February 2016 255

further investigation will be required to understand such

behaviour.
As shown in Figure 13b it is difficult to correlate variation
of dissolved CO concentration in either NE or mineral oil with
the number of chain scission of pressboard aged in particular
oil.
1 1
Avg . number of chain scision   (2)
DPt DP0

(a)

Figure 13. Relationship between number of chain scission and (a) dissolved
CO2, (b) CO.

6.2 HYDROCARBON COMBUSTIBLE GASSES

Figures 14a and 14b compares the variation of dissolved
major hydrocarbon gases in mineral oil and NE over the
ageing period. Overall total hydrocarbon gases dissolved in
mineral oil is always higher than that of in NE. According to
Figure 14, in addition to CO, ethane (C2H6) remains as the
main combustible gas, concentration of which is much higher
than the other hydrocarbon gases, followed by methane (CH4)
in both mineral and NE oils. Moreover, the ageing of oil can
mainly contribute to the increase of these two gases.
It is worthwhile to mention that the total dissolved gas
content in samples MP 1280h, MP1960h, N1280h, NP1280h
and NP1960 is relatively low, which may be due to the loss of
gas occurred during sampling or transport of the gas samples.
However, their proportional gas distribution remains similar.
7 VARIATION OF MOISTURE CONTENT
OVER AGEING
As shown in Figure 15a, one can clearly see that moisture
content of pressboard aged in mineral oil firstly decreases and
then increases to its initial value during the last 840 h of
ageing. On the other hand, that of in pressboard aged in NE
decreases from 0.7% to 0.26% and remains at fairly constant
value during the last 840 h of ageing. Initial reduction of
moisture content may partly be due to migration of moisture
from the paper to oil then to gas space [55], because
headspace is filled with dry nitrogen.
In the presented case, degradation of pressboard primarily
occurs via hydrolysis reaction by consuming moisture
absorbed in it. This behaviour may result in decreasing the
moisture content in pressboard until acid catalysed
dehydration reactions (acids catalysed hydrolysis) become
pronounced in the system [2].
(b)
Figure 14. Dissolved gases in (a) mineral oil (b) NE.
As shown in Figure 15b, one can clearly see that moisture
content of NE aged with pressboard first increases and then
decreases. This behaviour may be due to the fact that moisture
in pressboard migrates into NE (due to the hygroscopic nature
of NE) and then enhances the hydrolytic degradation of NE.
When the hydrolysis reaction is pronounce in NE, moisture
content in NE decreases as shown in Figure 15b while its
acidity is increasing as presented in the Section 5.1.
Therefore, one can also assume that moisture content in
pressboard aged in NE decreases over ageing due to the fact
that moisture from pressboard can be consumed by NE. This
behaviour is further enhanced by hydrolysis reaction occurs in
NE.
Figure 15. Change of moisture content with ageing (a) in pressboard, (b) in
oil.
8 CONCLUSIONS
This study confirmed that the ageing rate of cellulose
material in NE is relatively higher than that in mineral oil up
to 1280 h. On the other hand, during 1280 h to 2800 h, the
256 K. Bandara et al.: Understanding the Ageing Aspects of Natural Ester Based Insulation liquid in Power Transformer
average ageing rate of pressboard in mineral oil is three times
greater than that of pressboard in NE. This trend was clearly
supported by the percentage decrease in DP of the pressboard.
Based on this one could expect that after 2800 h the time to
acquire end of life DP (250) for the pressboard in NE will be
higher than that in mineral oil.
The 2-FAL concentrations found in mineral oil and NE
revealed that it has some correlation with pressboard
ageing kinetic. Therefore, it seems that 2-FAL can also be
used as a chemical indicator for identifying the ageing
condition of solid insulation in NE filled transformer.
However, more research is required to understand whether
there is any relationship with partitioning of 2-FAL
between NE and pressboard insulation under varying
acidity levels.
FTIR results showed that absorbance peaks observed at
3275 cm-1 and 3325 cm-1 corresponding to H-bonds in
cellulose polymers are more sensitive to degradation of
cellulose. The peak observed in the carbonyl range (both in
oil and pressboard) indicates the trace quantity of by-
products upon ageing. However, more quantitative
measurements are still required verify the findings. The
peak observed at 1746 cm-1 in the FTIR spectra of
pressboard samples aged in NE confirms the esterification
of fatty acids with cellulose.
The continuous increase of acidity value of NE oil
samples aged with pressboard and copper is due to
hydrolysis of NE. Indeed, at high temperature moisture
migrates from pressboard to oil and enhances the
hydrolytic degradation of NE. Moreover, one could
observe that dissipation factor is fairly constant with
increasing acidity, which also confirms the dominance of
hydrolytic degradation in NE over oxidation. Moreover,
high molecular acids in NE do not provide H+ through
dissociation leading to relatively constant dissipation factor
of NE while its acidity is increasing.
REFERENCES
[1] T. A. Prevost and T. V. Oommen, "Cellulose insulation in oil-filled
power transformers: Part I - history and development", IEEE Electr.
Insul. Mag., Vol. 22, No. 1, pp. 28-35, 2006.
[2] L. E. Lundgaard, W. Hansen, D. Linhjell, and T. J. Painter, "Aging of oil-
impregnated paper in power transformers", IEEE Trans. Power Delivery,
Vol. 19, pp. 230-239, 2004.
[3] L. E. Lundgaard, W. Hansen, and S. Ingebrigtsen, "Ageing of Mineral
Oil Impregnated Cellulose by Acid Catalysis", IEEE Trans. Dielectr.
Electr. Insul., Vol. 15, pp. 540-546, 2008.
[4] R. Liao, S. Liang; C. Sun, L. Yang and H. Sun, "A comparative study of
thermal aging of transformer insulation paper impregnated in natural
ester and in mineral oil", Euro. Trans. Electr. Power, Vol. 20, pp. 518-
533, 2010.
[5] C. P. McShane, K. J. Rapp, J. L. Corkran, G. A. Gauger, and J. Luksich,
"Aging of Kraft paper in natural ester dielectric fluid", Proceedings of
IEEE 14th Int’l. Conf. Dielectr. Liquids, pp. 173-177, 2002.
[6] T. V. Oommen, C. C. Claiborne, and J. T. Mullen, "Biodegradable
electrical insulation fluids", IEEE Electr. Insul., Electr. Manufacturing
and Coil Winding Conf., pp. 465-468, 1997.
[7] Y. Lijun, L. Ruijin, C. Sun, and Z. Mengzhao, "Influence of vegetable oil
on the thermal aging of transformer paper and its mechanism", IEEE
Trans. Dielectr. Electr. Insul., Vol.18, pp. 692-700, 2011.
[8] S. Levchik, J. Scheirs, G. Camino, W. Tumiatti, and M. Avidano,
"Depolymerization processes in the thermal degradation of cellulosic
paper insulation in electrical transformers", Elsevier J. Polymer
Degradation and Stability, Vol. 61, pp. 507-511, 1998.
[9] A. M. Emsley, X. Xiao, R. J. Heywood, and M. Ali, "Degradation of
cellulosic insulation in power transformers. Part 3: effects of oxygen and
water on ageing in oil", IEE Proc. Sci., Measurement and Technology,
Vol. 147, pp. 115-119, 2000.
[10] A. P. Kumar and R. P. Singh, "Biocomposites of cellulose reinforced
starch: Improvement of properties by photo-induced crosslinking",
Elsevier J. Bioresource Technology, Vol. 99, pp. 8803-8809, 2008.
[11] A. M. Emsley and G. C. Stevens, "Kinetics and mechanisms of the low-
temperature degradation of cellulose", J. Cellulose, Vol. 1, pp. 26-56,
1994.
[12] A. M. Emsley, "The kinetics and mechanisms of degradation of cellulosic
insulation in power transformers," Elsevier J. Polymer Degradation and
Stability, Vol. 44, pp. 343-349, 1994.
[13] D. J. T. Hill, T. T. Le, M. Darveniza, and T. Saha, "A study of the
degradation of cellulosic insulation materials in a power transformer. Part
III: Degradation products of cellulose insulation paper", Elsevier J.
Polymer Degradation and Stability, Vol. 51, pp. 211-218, 1996.
[14] Cigre Working Group T.F. D1.01.10, "Ageing of cellulose in mineral-oil
insulated transformers", Cigre brochure 323, 2007.
[15] J. Unsworth and F. Mitchell, "Degradation of electrical insulating paper
monitored with high performance liquid chromatography", IEEE Trans.
Electr. Insul., Vol. 25, pp. 737-746, 1990.
[16] R. Gilbert, J. Jalbert, P. Tétreault, B. Morin, and Y. Denos, "Kinetics of
the production of chain-end groups and methanol from the
depolymerization of cellulose during the ageing of paper/oil systems. Part
1: Standard wood kraft insulation", J. Cellulose, Vol. 16, pp. 327-338,
2009.
[17] R. Gilbert, J. Jalbert, S. Duchesne, P. Tétreault, B. Morin, and Y. Denos,
"Kinetics of the production of chain-end groups and methanol from the
depolymerization of cellulose during the ageing of paper/oil systems. Part
2: Thermally-upgraded insulating papers", J. Cellulose, Vol. 17, pp. 253-
269, 2010.
[18] J. Jalbert, R. Gilbert, P. Tétreault, B. Morin, and D. Lessard-Déziel,
"Identification of a chemical indicator of the rupture of 1,4-β-glycosidic
bonds of cellulose in an oil-impregnated insulating paper system", J.
Cellulose, Vol. 14, pp. 295-309, 2007.
[19] P.K.B. Hodges, “Mineral base oils and oxidation stabilitty”, in Hydraulic
Fluids, Elsevier, 1996.
[20] M. J. Heathcote, “Basic Materials”, in J and P Transformer Book - A
Practical Technology of the Power Transformer (13th Edition),: Elsevier,
2007.
[21] I. Fofana, A. Bouaicha, M. Farzaneh, J. Sabau, D. Bussieres, and E. B.
Robertson, "Decay products in the liquid insulation of power
transformers", IET Electric Power Applications, Vol. 4, pp. 177-184,
2010.
[22] T. V. Oommen, "Vegetable oils for liquid-filled transformers", IEEE
Electr. Insul. Mag., Vol. 18, No.1, pp. 6-11, 2002.
[23] T. M. P. T. Rob Milledge, “Introduction to BIOTEMP-The superior,
biodegradable, high fire point, dielectric insulating fluid", ABB seminar
presentation, Australia, 2011.
[24] N. J. Fox and G. W. Stachowiak, "Vegetable oil-based lubricants—A
review of oxidation," Tribology Int’l., Vol. 40, pp. 1035-1046, 2007.
[25] S. Tenbohlen and M. Koch, "Aging Performance and Moisture Solubility
of Vegetable Oils for Power Transformers", IEEE Trans. Power Delivery,
Vol. 25, pp. 825-830, 2010.
[26] D. Rooney and L. R. Weatherley, "The effect of reaction conditions upon
lipase catalysed hydrolysis of high oleate sunflower oil in a stirred liquid–
liquid reactor", Elsevier J. Process Biochemistry, Vol. 36, pp. 947-953,
2001.
[27] L. Ruijin, H. Jian, G. Chen, M. Zhiqin, and Y. Lijun, "A comparative
study of physicochemical, dielectric and thermal properties of pressboard
insulation impregnated with natural ester and mineral oil", IEEE Trans.
Dielectr. Electr. Insul., Vol.18, pp. 1626-1637, 2011.
[28] L. A. Dissado and R. M. Hill, "Anomalous low-frequency dispersion.
Near direct current conductivity in disordered low-dimensional
materials", J. Chem. Soc., Faraday Trans. 2: Molecular and Chemical
Physics, Vol. 80, pp. 291-319, 1984.
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 23, No. 1; February 2016
[29] Z. Zhaotao, L. Jian, S. Grzybowski, and Z. Ping, "Moisture equilibrium
in vegetable oil and paper insulation system," IEEE Conf. Electr. Insul.
Dielectr. Phenomena, pp. 428-431, 2011.
[30] A. Emsley, R. Heywood, M. Ali, and C. Eley, "On the kinetics of
degradation of cellulose", J. Cellulose, Vol. 4, pp. 1-5, 1997.
[31] R. J. Heywood, G. C. Stevens, C. Ferguson, and A. M. Emsley, "Life
assessment of cable paper using slow thermal ramp methods", Elsevier J.
Thermochimica Acta, Vol. 332, pp. 189-195, 1999.
[32] A. M. Emsley, X. Xiao, R. J. Heywood, and M. Ali, "Degradation of
cellulosic insulation in power transformers. Part 2: formation of furan
products in insulating oil", IEE Proc. Sci., Measurement and Technology,
Vol. 147, pp. 110-114, 2000.
[33] Cigre working group TF D1.01.13 "Furanic Compounds for Diagnosis",
Cigre Brochure 494, 2012.
[34] I. Hohlein and A. J. Kachler, "Aging of cellulose at transformer service
temperatures. Part 2. Influence of moisture and temperature on degree of
polymerization and formation of furanic compounds in free-breathing
systems", IEEE Electr. Insul. Mag., Vol. 21, No. 5, pp. 20-24, 2005.
[35] Y. Shang, L. Yang, Z. J. Guo, and Z. Yan, "Assessing aging of large
transformers by furfural investigation", IEEE 7th Int’l. Conf. Solid
Dielectr., pp. 272-274, 2001.
[36] J. Jalbert, M. C. Lessard, and M. Ryadi, "Cellulose chemical markers in
transformer oil insulation Part 1: Temperature correction factors", IEEE
Trans. Dielectr. Electr. Insul., Vol. 20, pp. 2287-2291, 2013.
[37] A. J. Kachler and I. Hohlein, "Aging of cellulose at transformer service
temperatures. Part 1: Influence of type of oil and air on the degree of
polymerization of pressboard, dissolved gases, and furanic compounds in
oil", IEEE Electr. Insul. Mag., Vol. 21, No. 2, pp. 15-21, 2005.
[38] A. de Pablo, "Furfural and ageing: how are they related", IEE
Colloquium on Insulating Liquids, pp. 5/1-5/4, 1999.
[39] B. Hinterstoisser, M. Åkerholm, and L. Salmén, "Effect of fiber
orientation in dynamic FTIR study on native cellulose", Elsevier J.
Carbohydrate Research, Vol. 334, pp. 27-37, 2001.
[40] A.-M. Olsson and L. Salmén, "The association of water to cellulose and
hemicellulose in paper examined by FTIR spectroscopy", Elsevier J.
Carbohydrate Research, Vol. 339, pp. 813-818, 2004.
[41] H. Yang, R. Yan, H. Chen, D. H. Lee, and C. Zheng, "Characteristics of
hemicellulose, cellulose and lignin pyrolysis", Elsevier J. Fuel, Vol. 86,
pp. 1781-1788, 2007.
[42] J. Łojewska, P. Miśkowiec, T. Łojewski, and L. M. Proniewicz,
"Cellulose oxidative and hydrolytic degradation: In situ FTIR approach",
Elsevier J. Polymer Degradation and Stability, Vol. 88, pp. 512-520,
2005.
[43] J. Jalbert, E. Rodriguez-Celis, S. Duchesne, B. Morin, M. Ryadi, and R.
Gilbert, "Kinetics of the production of chain-end groups and methanol
from the depolymerization of cellulose during the ageing of paper/oil
systems. Part 3: extension of the study under temperature conditions over
120 °C", J. Cellulose, Vol. 22, pp. 829-848, 2015.
[44] Y. Lijun, L. Ruijin, S. Caixin, Y. Jianguo, and Z. Mengzhao, "Influence
of vegetable oil on the thermal aging rate of kraft paper and its
mechanism", Int’l. Conf. High Voltage Eng. and Application, pp. 381-
384, 2010.
[45] M. H. G. Ese, K. B. Liland, C. Lesaint, and M. Kes, "Esterification of
low molecular weight acids in cellulose", IEEE Trans. Dielectr. Electr.
Insul., Vol. 21, pp. 662-665, 2014.
[46] D. Martin, Z. D. Wang, A. W. Darwin, and I. James, "A Comparative
Study of the Chemical Stability of Esters for Use in Large Power
Transformers", IEEE Conf. Electr. Insul. Dielectr. Phenomena, pp. 493-
496, 2006.
[47] L. E. Lundgaard, W. Hansen, S. Ingebrigtsen, D. Linhjell, and M.
Dahlund, "Aging of Kraft paper by acid catalyzed hydrolysis", IEEE
Int’l. Conf. Dielectr. Liquids, pp. 381-384, 2005.
[48] W. K. Subzynski. J. S. Hyde, "Diffusion of oxygen in water and
hydrocarbons using an electron spin resonance spin-label technique", J.
Biophysical, Vol. 45, pp. 743-748, 1984.
[49] R. Bartnikas, Engineering Dielectric, Electrical Insulating Liquids, Vol.
III: ASTM Publication, 1994.
[50] T. V. Oommen, C. C. Claiborne, E. J. Walsh, and J. P. Baker, "A new
vegetable oil based transformer fluid: development and verification",
IEEE Conf. Electr. Insul. Dielectr. Phenomena, pp. 308-312, 2000.
[51] M. Duval, "The duval triangle for load tap changers, non-mineral oils and
low temperature faults in transformers", IEEE Electr. Insul. Mag., Vol.
24,No. 6, pp. 22-29, 2008.
257
[52] R. B. Michel Duval, "Stray Gassing of FR3 Oils in Transformers In
Service", Int’l. Conf. Doble Clients Bostan USA, 2009.
[53] "IEEE Guide for the Interpretation of Gases Generated in Oil-Immersed
Transformers", IEEE Std C57.104-2008 (Revision of IEEE Std C57.104-
1991), pp. 1-36, 2009.
[54] S. Singh and M. N. Bandyopadhyay, "Dissolved gas analysis technique
for incipient fault diagnosis in power transformers: A bibliographic
survey", IEEE Electr. Insul. Mag., Vol. 26, No. 6, pp. 41-46, 2010.
[55] T. K. Saha, P. Purkait, "Understanding the impacts of moisture and
thermal ageing on transformer's insulation by dielectric response and
molecular weight measurements", IEEE Trans. Dielectr. Electr. Insul.,
Vol. 15, pp. 568-582, 2008.
Kapila Bandara (M’o8) received his B.Sc. Eng.
(Hons) and M.Sc.Eng from the University of
Peradeniya, Sri Lanka in 2004 and 2010 respectively.
He is currently a Ph.D. degree student in Power and
Energy System research group in the University of
Queensland Australia. He is a member of IEEE DEIS
since 2012. His research interest includes condition
monitoring of power transformers and biodegradable
insulation.
Chandima Ekanayake (M’00-SM’15) received his
B. Sc. Eng.(Hons) in 1999 from University of
Peradeniya, Sri Lanka. He obtained his Tech. Lic. and
the Ph.D. degree from Chalmers University of
Technology Sweden in 2003 and 2006 respectively.
Currently he is a lecturer in Griffith University Gold
Coast Australia and an honorary research fellow
attached to power and energy research group of the
University of Queensland. Before moving to
Australia he was with University of Peradeniya Sri Lanka as a Senior lecturer.
His research interests include condition monitoring of power apparatus,
alternatives for insulating oil, insulation ageing and power systems.
Tapan Kumar Saha (M’93-SM’97) was born in
Bangladesh in 1959 and immigrated to Australia in
1989. He received his B.Sc. Engineering (electrical
and electronic) degree in 1982 from the Bangladesh
University of Engineering & Technology, Dhaka,
Bangladesh, the M.Tech (electrical engineering)
in1985 from the Indian Institute of Technology, New
Delhi, India and the Ph.D. degree in 1994 from the
University of Queensland, Brisbane, Australia. Tapan
is currently Professor of Electrical Engineering in the School of Information
Technology and Electrical Engineering, University of Queensland, Australia.
Previously he has had visiting appointments for a semester at both the Royal
Institute of Technology (KTH), Stockholm, Sweden and at the University of
Newcastle (Australia). He is a Fellow of the Institution of Engineers,
Australia. .His research interests include condition monitoring of electrical
plants, power systems and power quality.
Pratheep K Annamalai is a materials engineer at
Australian institute for bioengineering and
nanotechnology (AIBN), the University of
Queensland (UQ). His research interests focus mainly
on the formulation, processing and evaluation of
polymer nanocomposites for industrial applications
and sustainable polymer materials. Before joining UQ,
he studied Chemistry in University of Madras,
received PhD in Chemistry jointly from University of
Pune and National Chemical Laboratory (India), then
went on to work as postdoctoral researcher on polymer membranes at
Université Montpellier II (France) and on smart polymer nanocomposites
based on cellulose at Adolphe Merkle Institute, Université de Fribourg
(Switzerland). He is an emerging researcher of polymer science and
engineering with contribution of 28 research articles, 3 review articles and 4
book chapters with focus on biopolymers. His research activities are currently
funded by Australian Research Council and focus on cellulose based
materials (nanocomposites, carbon fibres and polymer foams)