Comparative Study on Power Transformers
Vegetable and Mineral Oil Ageing
Alexandra Ciuriuc*, Mirela Stefania Vihacencu*, Laurentiu Marius Dumitran* and Petru V. Notingher*
Laboratory of Electrical Materials, *University Politehnica of Bucharest,
aciuriuc@elmat.pub.ro
Abstract Due to the increasing need to replace the petro-
leum based products, theres an interest in replacing the
mineral insulating oil for electrical equipment with biode-
gradable vegetable oil. This paper presents a comparative
experimental study concerning the ageing of mineral and
vegetable oils used in power transformers. In this view,
samples of mineral and vegetable oil were accelerated ther-
mally aged, and their dielectric properties (real part of the
complex relative permittivity, loss factor and resistivity)
were determined at different ageing times. The evolution of
the degradation process in time was studied by determining
the dielectric properties variation rate. The results show
that the values of the relative permittivity and loss factor
determined for new oil are higher in the case of vegetable oil
than for the mineral oil, and vegetable oils resistivity is
lower than that of mineral oil. It has been determined that
during the ageing process the color of the mineral oil sam-
ples visibly changes with the ageing time, while in the case of
vegetable oil samples, these changes are not significant. Al-
so, the variation rate of vegetable oils real part of relative
permittivity, loss factor and resistivity is decreasing in time,
while that of mineral oil has an increasing tendency.
I. INTRODUCTION
The power transformers are some of the most important
and expensive equipments which ensure the operation of
the power grids. It is well established that the lifetime of
power transformers is mostly determined by the insula-
tion systems lifetime [1]. Actually, the large majority of
transformers have their insulation systems made of cellu-
lose paper and mineral oil (which is also the cooling flu-
id). Due to the growing need for friendly environment
products and to the fact that petroleum resources are lim-
ited, theres a great interest in replacing the mineral oil
with vegetable biodegradable oils even in the case of the
power transformers insulation systems [2]. In such a case
it is very important to evaluate in a correct way the main
properties and degradation process characteristics of the
new proposed vegetable oils taking as reference the min-
eral oil data.
The mineral oil was used in power transformers for
over 100 years until now mostly because the fact that it
has appropriate insulating properties and compatibility
with cellulose paper but also due to its relatively low cost.
From chemical point of view, the mineral oil is a mixture
of naphthenic, paraffinic and aromatic hydrocarbons. The
fire point is about 160 oC and for this reason the risk of an
explosion of a mineral oil based power transformer is
higher, and if spills occur, the water and soil are contami-
978-1-4673-1810-5/12/$31.00 2012 IEEE
nated [2]. Moreover, due to the chemical composition, the
mineral oil is not biodegradable and in the case of an ac-
cident, the environment is affected for a long time period.
On the other hand, the vegetable oils consist of triglyc-
eride esters of saturated fatty acids with 8 to 22 carbon
atoms and also fatty acids with one unsaturated bond
have carbon atoms ranging from 10 to 22 and fatty acids
with di- and tri-unsaturation which mostly contain 18
carbon atoms (linoleic and linolenic acids) [3]. Due to
this chemical structure, vegetable oil not only has a high
biodegradability rate but its fire point (360C) is signifi-
cantly higher compared to the mineral oil. From this point
of view, the use of vegetable oils increases the power
transformers operation safety [2].
The vegetable oils are more susceptible to oxidation
and more hygroscopic. For this reason, the power trans-
former must have an appropriate structure so it wont be
contaminated with oxygen and water from the environ-
ment. However, their higher hygroscopicity also repre-
sents an advantage through the fact that in the sealed
transformers, a higher amount of the water which results
from the degradation reactions of the insulation system is
retained in oil; by consequence the paper degrades slower
than in mineral oil [4, 5].
Generally, the effect of moisture content in vegetable
oil on the value of the real part of relative permittivity is
that the r of oil increased with increasing moisture con-
tent and decreased with frequency [6]. In [7] the values of
the real part of relative permittivity, loss factor (both
measured at a frequency of 1mHz) and resistivity of min-
eral oil were determined. It was observed a higher in-
crease of r and tan of mineral oil aged at 155C for
1000 hours, and a decrease in the value of the resistivity
after the same ageing time.
In a previous paper [8], the electrical charging tendency
of the mineral and vegetable oil has been studied. The
results showed a relation between the electrical properties
of the oils and their electrical charging tendencies. It was
also observed that vegetable oil has a higher electrical
charging tendency than the mineral oil.
In this paper, in order to evaluate the main properties
and degradation process characteristics of the vegetable
and mineral oil, samples of mineral and vegetable oil
were accelerated thermally aged, and the real part of the
complex relative permittivity, loss factor and resistivity
were determined at different ageing times. The evolution
of the degradation process in time was studied by deter-
mining the dielectric properties variation rate.
 II. AGEING OF OILS
During transformers operation, the oil-cellulosic paper
insulation is subject to thermal, electrical and environ-
mental stress factors. In time, these factors lead to the
insulation systems degradation, therefore to worsening
its electrical properties. The most important of these is
the temperature. As a consequence of high temperature,
chemical reactions take place and lead to oils degrada-
tion. The most important reactions that lead to the degra-
dation of mineral oil are those who lead to C-C bonding
fracture, with the formation of an alkane, alkene (crack-
ing reaction) and dehydrogenation [9]:
CnH2n+2 CnH2n + H2
2)
Cm+nH2(m+n)+2 CmH2m + CnH2n+2.
The main reactions that cause the degradation of the
vegetable oil are oxidation and hydrolysis, from which
result decomposition products like CO, CO2, water and
acids [3].
In addition, the cellulosic paper from the insulation sys-
tem is subject to the temperatures action and therefore
chemical reactions take place from which degradation
products result, like CO, CO2, H2O, H2, CH4 and furans.
These products are localized in the oil and can determine
a change in its main properties [9].
Oxygen constitutes one of the most influential factors in
the oil ageing. It can be naturally present or may ingress
into oil from the environment. Moreover, high tempera-
ture accelerates the oil oxidation and some metals like
copper act as a catalyst.
In the case of mineral oil, the oxygen can degrade the
hydrocarbons through processes based on free radical
reactions, which generate hydro peroxides [10]. The hy-
dro peroxides are unstable and decompose to free radicals
finally leading to the appearance of water, ketones and
other oxidation products [11]. The oxidation of mineral
oil takes the following steps [9]:
Initiation: RH + O2 RO2
Propagation: RO2 + RH ROOH + R
R + O2 RO2
Termination: RO2 + R ROOR
The oxidation rate and the concentration of transient
products depend on the chemical structure of the oil and
the oxidation conditions, respectively, temperature, pres-
sure, environment air contact etc. [9].
On the other hand, the oxidation reaction of vegetable
oils is more intense as the content of unsaturated fatty
acids increases, due to the fact that the oxygen attacks the
carbon double bonds and leads to the breakage of the
fatty acid chain (Fig. 1) [12]. In fact, this explains the
biodegradability of vegetable oils [2].
Fig. 1. Disruption of a fatty acid by oxidation [12].
Oxidation of vegetable oil presents the following reac-
tion paths [13]:
RH + O2 ROOH (1)
ROOH + RH 2R + H2O2 (2)
R + O2 ROO (3)
ROO + RH ROOH + R (4)
2ROOH ROO + RO +H2O (5)
Oxygen molecule existing as biradical reacts with the
oil to form hydroperoxide (eq. (1)) which reacts further
again with the oil forming hydrocarbon free radicals by
hydrogen attraction from the oil (eq. (2)). According to
eq. (3), the oil free radical readily reacts with oxygen to
generate alkyl peroxide radical with low activation en-
ergy. This peroxy radical reacts again with the oil produc-
ing hydroperoxide (eq. (4)) which finally dissociates into
the peroxy radical, alkoxy radical and water, according to
eq. (5). The steps (4) and (5) are the chain propagation
steps. If the phenolic antioxidant is present in the reaction
system, peroxy radical will attract hydrogen to produce
hydroperoxide and phenoxy radical. Reactivity of the
phenoxy radical is lower than that of ROO because of
the steric effect of phenol and better stability [13].
The hydrolysis of vegetable oil represents the reverse of
the esterification reaction. Under the influence of water,
incomplete ester molecules and free fatty acids emerge.
The ester molecules are disconnected at their C-O-bonds.
Since the free fatty acids themselves accelerate the proc-
ess of the hydrolysis, it is also called an autocatalytic
reaction. Further acceleration occurs when the water con-
tent, the oil temperature, or percentage of the dissolved
metals increases [12].
This paper presents an experimental study concerning
the thermal ageing of mineral and vegetable oil used in
power transformers.
III. EXPERIMENTAL SET-UP AND SAMPLES
The experiments were performed on two different types
of oils. In order to perform the ageing of the oils, 15 con-
trolled clean stainless steel recipients (ageing cells) were
used. The first five of these were filled with vegetable oil,
the next five cells were filled with vegetable oil and 10
samples of cellulose paper with 0.25 mm thickness (5
samples of 55x55 mm2 and 5 samples of 40x40 mm2) and
the last five cells were filled with mineral oil. Each age-
ing cell contained 9 grams of copper and 23 grams of
iron. The experiments were performed on three types of
samples:
x Sample A: vegetable oil aged in the presence of
copper and iron;
x Sample B: vegetable oil aged in the presence of
paper, copper and iron;
x Sample C: mineral oil aged in the presence of
copper and iron;
All ageing cells, fully sealed, were introduced in the
drying oven at a temperature of 155C for a total of 1250
h. Three ageing cells containing each type of sample (A,
B and C) were extracted every 250 h and the following
dielectric properties of the oils were determined: the real
and imaginary parts of the complex relative permittivity,
the loss factor and the resistivity.
In order to determine the dielectric loss factor and the
components of the complex relative permittivity (Hr and
Hr), a Novocontrol spectrometer was used. The voltage
applied to the samples had the value of 1V and the elec-
tric field frequency was in the range of 10-3 103 Hz [14].
The temperature of the measurement was 28C.
In order to determine the values of absorp-
tion/resorption currents it was used a standard liquids cell
Irlab CL-1 type, and the oils volume resistivity U was
calculated based on the absorption ia(t) and resorption
ir(t) currents measured with a Keithley 6517 electrometer
connected to this cell [7]:
trodes), and S is the medium area between the two elec-
trodes calculated with the formula:
S 2SRmed L , (7)
d int  l
where Rmed (8)
2
IV. RESULTS AND DISCUSSIONS
In Figs. 3-5 it can be seen that during the accelerated
thermal ageing process the color and transparency of the
mineral oil samples drastically changes with the ageing
time (Fig. 3), while in the case of vegetable oil samples,
these changes are not significant (Figs. 4 and 5). This
may indicate the fact that the degradation process differs
from the mineral oil to vegetable oil and this is due to the
different nature of resulting chemical by-products.
In Figs. 6-9 the variation curves of the real part of
relative permittivity and loss factor with the ageing time
are presented. It can be seen that both quantities increase
with the ageing time for all types of samples. This could
be due to the increase in degradation reactions products
caused by the high temperature.

S U0 (6)

l ia (3600)  ir (3600)

where ia(3600) and ir(3600) represent the values of ab-

sorption and, respectively resorption currents measured Fig. 3. Mineral oil samples: a - unaged; b - thermally aged for 1000 h at
after 3600 seconds from the moment the voltage was ap- T = 155C; c - thermally aged for 1250 h at T =155C.
plied, respective disconnected, U0 = 40 V is the applied
DC voltage, S - the active electrodes surface and l - the
distance between the electrodes.
The measuring cell for absorption/resorption currents
consists of two cylindrical electrodes (Fig. 2) immersed
in a recipient filled with the oil sample.

Fig.4. Vegetable oil samples: a - unaged; b - thermally aged for 1000 h

at T = 155C; c - thermally aged for 1250 h at T =155C.

Fig. 2: Scheme of the liquid measuring cells electrodes for ab-

sorbtion/resorbtion currents: interior electrode (1), exterior electrode
(2), dint diameter of the interior electrode, l the distance between the
electrodes, L the cylinders height.

In order to calculate the resistivity, it was considered l

Fig. 5. Vegetable oil samples: a - unaged; b - thermally aged in the
as the sample thickness (the distance between the elec-
presence of paper for 1000 h at T= 155 C; c - thermally aged in the
presence of paper for 1250 h at T =155C.
 The initial values of r and tan determined for vegeta-
ble oil are, due to its polar nature, significantly higher
than those of mineral oil.
After 1250 hours of thermally ageing, r of vegetable
oil (Sample A) presents a small increase, by 6.1%. (Table
I). In the case of mineral oil (sample C) it can be seen a
much higher increase of the real part of relative permittiv-
ity, by 72.4 %. The relative permittivity of vegetable oil
aged in the presence of paper (sample B) increases by
33%. The difference between the samples A and B could
be due to the increasing of the dipoles concentration in oil
resulted from the paper degradation products, like furans
and water.
Fig. 6. Variation of mineral oils (sample C) real part of relative permit-
tivity with ageing time ( f = 1mHz).
Fig. 7. Variation of vegetable oils (sample A) and vegetable oils aged
in the presence of paper (sample B) real part of relative permittivity
with ageing time (f = 1mHz).
Fig. 8. Variation of mineral oils (sample C) loss factor with ageing time
(f = 1 mHz).
Fig. 9. Variation of vegetable oils (sample A) and vegetable oils aged
in the presence of paper (sample B) loss factor with ageing time
(f = 1 mHz);
TABLE I. NEW AND AGED OILS DIELECTRIC PROPERTIES VALUES
State Property r Tan
[m]
of oil Sample (f = 1 mHz) (f = 1 mHz)
A 3.75 30.71 3.091011
New
C 2.90 2.69 8.611012
A 3.98 133.05 2.891010
Aged 5.131010
(1250 B 4.99 198.97
hours) 5.131010
C 5 123.77
Also, the loss factor of mineral oil (sample C). in-
creases approximately 45 times. The loss factor for aged
vegetable oil (sample A) increases approximately three
times, and that of vegetable oil aged in the presence of
paper (sample B) presents an increase of about five times.
Tan of vegetable oil aged in the presence of paper (sam-
ple B) is significantly higher than that of vegetable oil
(sample A). The difference between the loss factor of the
samples A and B could have the following explanation:
the products resulted from the paper degradation process
are transferred in the oil, which increase the charge carri-
ers and dipoles concentration. This leads to an increase in
conductivity and relative permittivity, and as a conse-
quence, to an increase in conduction and polarization
losses. Also, new mineral oil has a higher value of the
resistivity than the vegetable oil, but after 1250h of age-
ing, for all three types of samples were determined closed
values of .
Figs. 10 and 11 show the variation of the resistivity of
mineral and vegetable oils as a function of ageing time. It
results that the value of resistivity decreases with the age-
ing time. Also, it can be seen that the resistivity of the
vegetable oil decreases more rapidly in the presence of
the paper (Fig. 11). Most likely, this is due to the result-
ing by-products of paper degradation which pass to the
oil volume. Moreover, even if the initial value of the
mineral oil resistivity is about three times higher than that
of vegetable oil, at the end of the ageing period, their val-
ues become very close (5,131010 :m, respectively
2.91010 :m).

Fig. 10. Variation of the mineral oils (sample C) resistivity with ageing
time.
Fig. 11. Variation of the vegetable oils (sample A) and vegetable oils
aged in the presence of paper (sample B) resistivity with ageing time.
Based on the experimental variations presented in Figs.
6-11, the variation rate (i.e. the derivative of variation
curve) of each property has been calculated for all meas-
ured points. As it can be seen in Fig. 12, the variation
rates of the mineral oils real part of the relative permit-
tivity, loss factor and resistivity have an increasing ten-
dency.
On the other hand, the real part of relative permittivity
variation rate of the vegetable oil is decreasing in time. In
the case of the vegetable oil aged in the presence of pa-
per,
Fig. 12. Variation of the real part of the relative permittivity variation
rate with ageing time.
it can be seen that the r variation rate decreases during
the first part of ageing (W < 500 h) and after this period, it
starts increasing. Even if the r variation rate has an in-
creasing tendency, in the second part of the ageing period
its value is still lower than that of mineral oil.
Fig. 13 shows that both vegetable oil with or without
cellulosic paper, present a decreasing loss factor variation
rate. However, in the case of mineral oil the loss factor
variation rate increases with ageing time.
Also, it can be observed that the loss factor variation
rate is higher for vegetable oil aged in the presence of
paper compared to that of simple vegetable oil. Due to the
ageing process, the vegetable oil can hydrolyze in the
presence of moisture, which releases fatty acids and glyc-
erol, hence to an increase in polar compounds, which
leads to higher conduction and polarization losses. How-
ever, there is no hydrolysis in mineral oil [15]. In the case
of vegetable oil aged in the presence of paper, the water
present in the paper is absorbed by the oil, which leads to
an increase in the hydrolysis reaction, and in consequence
to a higher concentration of free fatty acids. The more
polar compounds present in the oil, leads to higher losses,
hence this could explain the higher value of the loss fac-
tor variation rate of vegetable oil aged in the presence of
paper compared to that of vegetable oil.
In Fig. 14 it can be seen that the increase in the resistiv-
ity variation rate of the vegetable oil with and without
paper is barely visible compared to that of mineral oil.
Fig. 13. Variation of the loss factor variation rate with ageing time.
Fig. 14. Variation of the resistivity variation rate with ageing time.
 V. CONCLUSIONS
For this study, mineral and vegetable oil samples were
accelerated thermally aged at a temperature T = 155C
during 1250 hours and the real part of the complex rela-
tive permittivity, loss factor and resistivity were deter-
mined at different ageing times. Also, the variation rate of
these dielectric properties was calculated.
It has been determined that the values of r and tan
determined for new oil are higher in the case of vegetable
oil than for the mineral oil, and vegetable oils resistivity
is lower than that of mineral oil.
The variation rate of vegetable oils real part of relative
permittivity, loss factor and resistivity is decreasing in
time, while that of mineral oil has an increasing tendency.
An empirical model based on the variation of dipoles
and charge carriers concentration which explains the
dielectric proprieties variation rate in time will be pre-
sented in a future paper.
ACKNOWLEDGMENT
This work has been supported by the Sectoral Opera-
tional Programme Human Resources Development 2007-
2013 of the Romanian Ministry of Labour, Family and
Social Protection through the Financial Agreement
POSDRU/89/1.5/S/62557, POSDRU/88/1.5/S/60203,
POSDRU/107/1.5/S/76903.
REFERENCES
[1] R. Liao, S. Liang, L. Yang, J. Hao, J. Li, Comparison of Ageing
Results for Transformer Oil-paper Insulation Subjected to Thermal
Ageing in Mineral Oil and Ageing in Retardant Oil, IEEE Trans-
actions on Dielectrics and Electrical Insulation, Vol. 19, No. 1,
2012, pp.225-232.
[2] C. Perrier, A. Beroual, Experimental Investigations on Insulating
Liquids for Power Transformers: Mineral, Ester, and Silicone
Oils, IEEE Electrical Insulation Magazine, Vol. 25, No. 6, 2009,
pp. 6-13.
[3] T.V. Oommen, Vegetable Oils for Liquid-Filled Transformers,
IEEE Electrical Insulation Magazine, Vol. 18, No. 1, 2002, pp. 6-
11.
[4] C. P. McShane, Kevin J. Rapp, Jerry L. Corkran, G. A. Gauger, J.
Lukisch , Ageing of Paper Insulation in Natural Ester Dielectric
Fluid, 2001 IEEE/PES Transmission & Distribution Conference
& Exposition, Oct. 28 Nov. 02, Atlanta GA , Vol. 2, 2001, pp.
675 - 679.
[5] M. A. G. Martins, Vegetable Oils, an Alternative to Mineral Oil
for Power Transformers Experimental Study of Paper Ageing in
Vegetable Oil Versus Mineral Oil, IEEE Electrical Insulation
Magazine, Vol. 26, No. 6, 2010, pp. 7-13.
[6] H. Lizhi, K. Toyoda, I. Ihara, Dielectric properties of edible oils
and fatty acids as a function of frequency, temperature, moisture
and composition, Journal of Food Engineering, Vol. 88, 2008,
pp. 151-158.
[7] L.V. Badicu, Diagnosis and Monitoring of Power Transformers
Insulation Systems, Ph D. Thesis, Bucharest, 2011.
[8] M.S. Vihacencu, A. Ciuriuc, L.M. Dumitran, P. V. Notingher,
Experimental Study of Electrical Properties of Mineral and Vege-
table Transformer Oils, International Conference on Innovation
and Engineering Research, 2-4 July, Bucharest, Romania, in press.
[9] Gh. Radulescu, M. Ilea, Physico-chemistry and lubricant oils
technology, Printed by Editura Tehnica, Bucharest, 1982.
[10] L. V. Badicu, L.M. Dumitran, P. V. Notingher, R. Setnescu, T.
Setnescu, Mineral oil lifetime estimation using activation en-
ergy, IEEE International Conference on Dielectric Liquids
(ICDL), 26-30 June 2011, pp. 1-4.
[11] M. H. Meshkatoddini, Ageing Study and Lifetime Estimation of
Transformer Mineral Oil, American Journal of Engineering Ap-
plied Sciences, Vol. 1, No. 4, 2008, pp. 384-388.
[12] S. Tenbohlen, M. Koch, Ageing Performance and Moisture Solu-
bility of Vegetable Oils for Power Transformers, IEEE Transac-
tions on Power Delivery, Vol. 25, No. 2, 2010, pp. 825-830.
[13] K.W. Lee, Z. Hong, F. Piao, Y.W. Kim, K.W. Chung, Bond
reactivity and antioxidant effect on the autooxidation of soybean
oil, Journal of Industrial and Engineering Chemistry, Vol. 16,
2010, pp. 419-424.
[14] P.V. Notingher, L.V. Badicu, L. M. Dumitran, G. Tanasescu, D.
Popa, Dielectric Losses in Cellulose-Based Insulations , Analele
Universitatii din Craiova, No. 33, 2009, pp. 169-175.
[15] R. Liao, J. Hao, G. Chen, Z. Ma and L. Yang, A comparative
study of physicochemical, dielectric and thermal properties of
pressboard insulation impregnated with natural ester and mineral
oil, IEEE Trans. on Dielectrics and Electrical Insulation, Vol.
18, No. 5, 2011, pp. 1626-1637.