IET Generation, Transmission & Distribution

Review Article

Dissolved gas analysis differences between ISSN 1751-8687

Received on 3rd August 2018

natural esters and mineral oils used in power

Revised 28th December 2018
Accepted on 5th September 2019
E-First on 18th November 2019
transformers: a review doi: 10.1049/iet-gtd.2018.6318
www.ietdl.org

Matias Meira1 , Cristian Ruschetti1, Raúl Álvarez2, Leonardo Catalano2, Carlos Verucchi1
1INTELYMEC (UNCPBA) and CIFICEN (UNCPBA-CICPBA-CONICET), Av. Del Valle 5737, (B7400JWI) Olavarría, Argentina
2IITREE-LAT, Facultad de Ingeniería, Universidad Nacional de La Plata, La Plata, Argentina
E-mail: [email protected]

Abstract: Typically and historically mineral oil has been used as liquid insulation in power transformers. Its functions are to
ensure the isolation between active parts, absorb heat transmitting it to the outer surfaces and protect other insulators (such as
paper) from moisture. Moreover, the analysis of insulating oil provides a diagnosis of the state of the transformer. In the last 10
years, there has been resurgence in the use of natural ester (vegetable oil) because of their ‘green’ credentials. Their
biodegradability and high fire point compared to mineral oils are some of its advantages. However, its high cost and low
evaluation in service (performance) still limit its application. For mineral oils there is a large database and knowledge obtained
from numerous transformers studied over many years. For this reason, certain gas patterns can be related to a specific fault.
Conversely, the limited field data related to natural ester makes the analysis of dissolved gases unreliable. Some recent studies
attempt to determine the differences between each type of oil when a fault takes place. This study presents a review of the
difference between the generated gases in mineral oil and natural ester as a result of the most common faults in transformers.

1 Introduction modifications. From the above it can be concluded that, through a

correct maintenance of the oil, it is possible to prolong the
Currently, in most power transformers, solid insulation (paper) has expected lifetime of the entire insulating system, including the
been used in combination with a liquid one (oil) to make up its paper [6].
insulating system. The paper is used to ensure the insulation The insulating capacity of the oil–paper combination is affected
between different turns of a winding and between primary and by various stresses (electrical, thermal and environmental) that
secondary windings. The oil is also required to make an important result from the operation of the transformer. The degradation of the
contribution to the efficiency of the solid insulation by penetrating properties of the oil–paper insulation can cause faults in the
into and filling the spaces between layers of paper insulation and transformer and, consequently, in the electrical network of which it
by impregnating, after they have been dried and deaerated by is a part. For this reason, permanent monitoring of the insulation
exposure to vacuum [1]. The oil protects the solid insulation by condition of the transformer is essential to ensure the power
acting as a barrier between the paper and the damaging effects of delivery [7].
oxygen and moisture [2]. Furthermore, the oil is the most efficient Historically mineral oils were used in most transformers and
medium for absorbing heat from the core and the windings and electrical equipment. However, some years ago, the use of non-
transmitting it, sometimes aided by forced circulation, to the mineral or vegetable oils (such as natural and synthetic esters and
naturally or artificially cooled outer surfaces of the transformer [3]. silicone oils) began to gain notoriety. The advantage of vegetable
There are other transformers that use completely solid oils is that their temperature of inflammation is significantly higher
insulation instead of oil. These transformers are called dry-type than that of mineral oils. Therefore, they have advantages in
transformers. Its advantage lies in a lower risk of inflammation and applications where it is necessary to reduce the fire risk (such as in
fire, e.g. in indoor substations where the possibility of an oil spill is indoors, in urban areas or in railway transformers) [3, 8, 9].
unacceptable [4]. Its use is restricted for economic reasons since So, alternative fluids are now widely used in a variety of
dry-type transformers are more expensive than oil-immersed transformer applications, covering transformers for distribution,
transformers. For this reason, this alternative is not available for power and traction, but not yet including instrument transformers,
large power transformers [5]. as these have only recently been introduced as prototypes. One
The lifetime of an oil-immersed transformer is limited by the reason why non-mineral oil is not used in a greater number of
lifetime of the insulating paper (cellulose). Unlike oil, the applications is due to the lack of information on their behaviour in
degradation of paper is irreversible and its replacement requires the face of failure modes. In effect, the criteria for the diagnosis of
removal of the transformer's tank, operation that involves large transformers based on physical and chemical analysis of the oil
costs. To extend the lifetime of the paper it is essential to maintain should be reviewed in the case of non-mineral oils. Anyway,
the characteristics of the oil in which it is immersed without currently the knowledge database of alternative fluids is growing
rapidly which may encourage end users to look at these fluids, as
Table 1 Use of insulating liquids [10] more data become available [10].
Table 1 briefly summarises the up-to-date distribution of the use
Mineral oil Vegetable oils (natural of insulating mineral oil and natural ester. It is evident that
ester) currently the use of natural esters in power transformers is not
power transformers A B widely spread.
traction transformers A X An important question for vegetable oils still unresolved lies in
distribution transformers A A the chemical processes that occur in the presence of certain faults.
instrument transformers A X When there are electrical arcs, partial discharges or high-
A, largely used; B, used but less common; X, not used.
temperature spots, chemical reactions occur that generate gases in
the oil. These gases are shown in Table 2. For mineral oils, there

IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 24, pp. 5441-5448 5441
© The Institution of Engineering and Technology 2019
are precise and widely tested procedures that allow deducing the possible to identify the type and severity of the fault that produced
characteristics and severity of a fault based on the content of gases them [13–15].
present in an oil sample by dissolved gas analysis (DGA). In the The most commonly used techniques for evaluating the health
case of vegetable oils, such diagnostic procedures are based on a of power transformers based on DGA are Rogers ratio,
limited amount of in-service DGA results and several laboratory- Doernenburg ratio [16], IEC ratio [17], Key Gas [17] and Duval
based experimental datasets [11, 12]. Therefore, the application of triangles [18]. Rogers ratio, Doernenburg ratio and Key Gas
these diagnostic techniques for transformers immersed in vegetable techniques are included in IEEE Std C57.104 [19].
oil is being studied. Although the literature presents some The methods based on ‘gas ratio’ (Rogers, Doernenburg and
preliminary results, it is still difficult to draw definitive IEC) analyse the proportion of different gases. If the gas ratios are
conclusions. Faults that can occur in a transformer are listed in within certain values it is possible to establish that a fault occurred
Table 3. in a transformer [17, 19]. As an example for Rogers ratio, if ratios
In this paper, a general review of the studies carried out to date C2H2/C2H4, CH4/H2 and C2H4/C2H6 are below 0.1, low-energy
in relation to the new oils is presented. The focus is on the prospect density arcing – partial discharge (PD) – is suggested as possible
of fault diagnosis through DGA. Based on an exhaustive literature fault type.
review, conclusions are reached about the possibility of applying On the other hand, Key Gas technique is based on associating
DGA in vegetable oils. In addition, the studies that should still be each type of fault with a particular gas. If the relative percentage of
done are mentioned. one of the typical or predominant gas exceeds a limit value, it is
concluded that the transformer has a fault associated with these gas
2 Techniques for the interpretation of gases [11]. Duval triangles use the concentration of three dissolved gases
generated and based on the relative proportions of each one of them
distinguishes between faults. In general terms, the involved gases
DGA has been used for many years as an effective and reliable tool vary according to the triangle used. There are seven Duval triangles
to detect incipient faults in mineral oil-filled transformers. The with different purposes. The Duval triangle 1 is for mineral oil
information provided by DGA analysis is extremely important to filled transformers, the Duval triangle 2 is for transformers with
the asset managers with electricity supply companies. Fault gases on-load tap changer (OLTC) and mineral oil filled, the Duval
are formed when sufficient energy breaks the chemical bonds to triangle 3 is for non-mineral oils filled transformers, the Duval
form reactive fragments. These fragments combine to form the triangles 4 and 5 are for low-temperature faults and mineral oil
range of fault gases. Although the chemical structures of mineral filled and the Duval triangles 6 and 7 for the low temperature faults
oil, ester fluids and silicone fluid are fundamentally different, they in the transformers filled with FR3 [14, 18, 20–22].
all contain similar bond types, for e.g. C–H bonds, such that the There are also other alternatives that propose the combination
types of fault gases are similar from each of the different fluids. of the different techniques listed above in order to obtain a more
For a given fault, the amounts and ratios of the gases may differ reliable diagnosis. Thus, the possibility of diagnostic error is
from fluid to fluid. reduced [23–25]. The condition diagnosis criterion of transformers
Currently, the results of the application of the DGA technique is to integrate the results of the different interpretations techniques
allow identifying different types of faults in a transformer. Based and thus achieve a single diagnosis. Table 4 shows as an example
on the gases present in an oil sample and at its concentration, it is the correct predictions made with the use of the different
techniques in an individual and combined way [25]. For this, 30
samples with a known cause of fault taken from [26] were
Table 2 Chemical names and symbols considered. The authors carried out this analysis for mineral oils
Name Symbol and propose to develop it, as future work, for vegetable oils.
hydrogen H2 In [23], an algorithm is also presented where the criteria of IEC
[17] and IEEE [19] were incorporated for the analysis of DGA
carbon monoxide CO
results. The paper provides an easy-to-implement solution that
carbon dioxide CO2 allows obtaining a diagnosis that simultaneously contemplates the
methane CH4 results of the two mentioned criteria. Currently, these combined
ethane C2H6 methods have been applied only for mineral oils. It is expected that
in future work they will be extended to natural ester oils. Further
ethylene C2H4
details on the application of combined techniques of interpretation
acetylene C2H2 of DGA and a comparison of their performance can be found in
[27, 28].
Finally, it should be noted that in most power transformers, the
Table 3 Types of faults OLTC and the transformer's windings and core are often placed in
PD Partial discharges separate tanks. So, the changer oil and the transformer oil are
D1 discharges of low energy separated. However, in some transformers the OLTC can be found
directly inside the main tank. In these cases, the commutation arcs
D2 discharges of high energy
could be identified by DGA techniques as a fault. In order to
T1 thermal faults of temperature <300°C overcome this limitation, the standards already propose two limits
T2 thermal faults of temperature 300°C<T<700°C for C2H2 (for transformers with and without OLTC) both for
T3 thermal faults of temperature >700°C typical gas concentration and for typical rates of gas increase [17].
DT mixtures of electrical and thermal faults In addition to the traditional DGA interpretation techniques
mentioned, and their combination to improve the transformer
diagnosis, there are other proposals and complementary tools for
Table 4 Prediction percentages of DGA techniques [25] DGA interpretation.
Method Correct predictions Percentage of In [29], a new technique based on a graphical heptagon made
prediction, % with seven gases produced in faulty transformer oil decomposition
is presented. It is found that the proposed heptagon technique
key gas 13/30 43
improves the accuracy of fault classification in power transformers
IEC ratio 19/30 63 compared with other techniques, reaching a precision of 89.41%
ROGERS ratio 23/30 76 while the Duval triangle 1, Doernenburg ratio, IEC ratio and
Doernenburg ratio 22/30 73 Rogers ratio reached an accuracy of 64.67, 38.48, 54.67 and
Duval triangle 25/30 83 46.43%, respectively. However, these precisions were self-
combined methods 28/30 93 evaluated only on 452 test samples, so it cannot be generalised.

5442 IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 24, pp. 5441-5448
© The Institution of Engineering and Technology 2019
Evaluations with other oil samples from different sources must be The pentagons use the five combustible gases (H2, C2H2, C2H4,
performed to validate the method. CH4 and C2H6). The various pentagons available are presented for
As a complement to the Duval triangle1, a low energy applications in mineral oils (pentagons 1 and 2) and ester oils
degradation triangle (LEDT) with a degradation parameter R is (pentagons 3 for FR3, Rapeseed, Sunflower, Midel). The use of the
proposed in [30]. The LEDT and the degradation parameter R are pentagons allows the identification of additional subtypes of
sensitive indicators only with low energy degradation. Therefore, it thermal faults. These faults are ‘stray gassing of oil at 120 and
should be used in conjunction with other interpretation methods. 200°C’, ‘overheating (<250°C)’, ‘thermal fault T3 or T2 in oil
The LEDT has been applied to DGA records from over 200 only’ and ‘possible carbonisation of paper’.
transformers in service, giving an early prediction of the fault of
over 90%.
More recently, Duval developed pentagons for transformers and 3 General differences between mineral oil and
similar equipment, e.g. bushings and cables, for both mineral and natural ester
non-mineral oils (natural and synthetic esters, silicones) [31, 32]. While both mineral and natural oils are suitable for insulation and
cooling, their atomic structures are significantly different [33]. The
Table 5 General comparison of the characteristics of chemical differences between mineral and natural oils must be
different origin oils taken into account for asset operators to understand how to monitor
and maintain these liquids [34].
Characteristic Mineral oil Vegetable oil
There are several motivations for using dielectric fluids made
use highly used in highly used in from natural esters as a replacement for mineral oil in transformers.
power, distribution, and Natural esters are increasingly used because they are less
distribution and little used in flammable or more environmentally friendly (after one month, its
measurement power degradation is over 90% while mineral oil reaches 10%). Besides,
origin petroleum seeds and flowers natural esters are produced from vegetable oils, which are
derivatives themselves manufactured from renewable (sustainable) plant crops.
Is it renewable? no yes – also On the other hand, even small amounts of water in mineral oil
biodegradable cause a rapid deterioration in breakdown voltage. In contrast,
susceptibility to oxidation low high natural esters maintain high breakdown voltages even with
stability to up to very good acceptable but significantly larger amounts of dissolved moisture. The higher
oxidation 100°C (emergency lower than mineral viscosity of the natural ester fluids is partially compensated by the
conditions would oil more favourable thermal conductivity. Flash and fire points are
allow up to both significantly higher for esters than mineral oils. Esters have a
115°C) lower calorific value than mineral oil, so in the event of a fire they
greater very bad maintain stability provide less energy per unit weight than mineral oil. Natural esters
than 100°C (insulating up to 160°C have higher temperature stability than mineral oil, meaning esters
properties of the can be exposed to a higher temperature for longer periods with less
oil are degraded degradation than would be expected when using mineral oil. Use of
quickly) esters has been demonstrated to extend transformer asset life [10].
Data on the physical, chemical and electrical behaviour of
oxidation main effect carboxylic acids viscosity increase
natural esters compared to mineral oils have been analysed in
formation (very in long-term
numerous works [35–39]. However, the gas generation behaviour
corrosive)
is still under study. This topic will be discussed in the following
viscosity at 40°C, 9 30 to 42 section.
mm2/s Table 5 presents a general comparison of the characteristics of
at 100°C, 2.5 5.25 to 9 both oils.
mm2/s
thermal conductivity at 25°C, 0.126 0.144–0.17 3.1 Differences in diagnosis based on DGA
W/mK
DGA has been used for several decades in transformers immersed
combustion point, °C 170–180 >350 in mineral oil as an effective diagnostic tool. Therefore, there is a
flash point, °C 160–170 >300 large knowledge database available [3, 6, 17, 19, 44]. For natural
resistivity, GΩm 90–130 12–15 esters, this database is in an incipient stage of development. So, it
moisture content for dielectric 50 ppm 600 ppm approx. is important to know the way in which DGA can be applied to
strength = 30 kV/mm natural esters.
breakdown voltage (IEC 60156 > 70 kV [41] > 75 kV [42] In the literature, there is a small number of works on the
2.5 mm [40]) differences between mineral oils and natural esters in terms of gas
formation for different faults.
thermal capacity of paper–oil minor major
The equipment necessary to carry out the tests is not easily
insulation
obtained. The gas detection must be done with chromatography
sludge formation yes, affecting almost inexistent equipment in order to distinguish the different gases generated and
cooling their respective concentrations/proportions.
neutralisation number varies according varies according The same treatment equipment of mineral oils can be used for
to oxidation to moisture natural esters. Although natural ester and mineral oil are
dissipation power factor at ≤0.1 0.8 completely miscible, the mixture between them is undesirable and
90°C, % should be avoided in transformers and oil treatment equipment
DGA large database more [45]. The mixture with insulating mineral oil changes the
investigations are dissipation factor, decreases the flash point and can reduce the
needed combustion point of the natural ester. Therefore, to decontaminate
cost minor 2 to 3 times the treatment equipment, the oil filters should be removed,
higher recirculated with vegetable oil for 30 min and discard this volume
Is it hygroscopic? [43] no (less than highly (more than of used oil. After the installation of a new filter, the equipment is
cellulose) cellulose) ready for use [45].
A review of the methods used for producing vegetable oil is
Many of the data in Table 5 are extracted from [10].
presented in [46]. Furthermore, various experimental works carried

IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 24, pp. 5441-5448 5443
© The Institution of Engineering and Technology 2019
Fig. 1 Hydrocarbon gas distribution in natural ester after 2800 h of ageing [47]
(a) Sunflower-based natural ester, (b) Soy-based natural ester

out in different countries are described, giving an overview of the correspondence gas as given in the IEC 60567 [49]. Therefore, IEC
current research carried out on the vegetable oils. Electrical, gas ratio and Duval triangle methods seem to be applicable for
physical and environmental properties of natural esters and mineral interpretation of DGA results of natural esters. However, more
oil are compared as well as its performance. Merits, demerits and research is needed for identifying their applicability on various
challenges faced by vegetable oil are mentioned. types of natural esters.
In [47], the ageing behaviour of the dry pressboard insulation In [50, 51], a comparative study between mineral and various
immersed in a commercially available natural ester (soy oil based) ester oils based on DGA is reported. Electrical (breakdown of low
and a synthetic ester (MIDEL 7131) through accelerated ageing energy) and thermal (stray gassing) faults were realised at
experiment is evaluated. At different stages of ageing, oil laboratory scale, and then gas formation as well as diagnosis
conditions were analysed by measuring acidity value, 2- methods were compared: IEC gas ratio, Duval's triangle and key
furaldehyde (2-FAL) concentration and kinematic viscosity. It is gases. Two extraction methods based on gas chromatography
revealed that pressboard in both natural and synthetic esters have a described in IEC 60567:2011 [49] were used: total vacuum
similar ageing pattern. Similar to the behaviour of mineral oils, degassing and stripping method. For electrical faults and especially
dissolved 2-FAL and carbon oxide concentrations are closely for low energy discharges, the same main gases (H2 and C2H2) are
associated with pressboard degradation. DGA is also performed to created in relatively same ratios, thus confirming the efficiency of
assess the degradation of both pressboard and oil. Overall, total key gases method for ester oils. It is also shown that classic Duval's
combustible gases (TCG) in natural ester are many times greater triangle (triangle 1) is well applicable for this type of fault
than in synthetic ester. For example, after 1960 h ageing, the total confirming the results reported in literature [13]. For thermal faults
hydrocarbon gas dissolved in natural ester is almost 20 times and especially for stray gassing issue (low thermal faults), C2H6
greater than synthetic ester. C2H6 remains as the main stray gas in (associated with H2) constitutes the key gas of natural esters. The
natural ester for thermal fault followed by CH4. There is no new Duval's triangles can be applied: triangle 4 for mineral oil and
evidence of stray gas in synthetic esters. Duval triangle technique triangle 6 for ester oils. The extraction methods and results are
successfully identifies the faults in both types of oil as T1. The comparable for mineral oil whereas the higher viscosity of ester
breakdown voltage of new natural and synthetic ester is almost oils could interfere with the extraction by stripping method.
similar to that of new mineral oil. However, the mixing of natural In [52], the concentration of dissolved gases under electrical
and synthetic ester with mineral oil (new and aged) results in performance in vegetable oils – rice bran oil and palm oil – were
decreasing the breakdown voltage. Compared to DGA results of investigated and compared with mineral oils. From six
natural ester used in [47] with sunflower oil-based natural ester concentrations of gases contained in the samples (H2, CH4, C2H6,
used in [48] (at similar ageing conditions), the two natural esters C2H4, C2H2 and CO), C2H2 and H2 were the highest gases
have different gassing behaviour for thermal fault. Fig. 1 shows generated due to electrical stress for all the oils. Based on the
hydrocarbon gas distribution for both natural esters after 2800 h of theory, these gases are representative of arcing fault condition.
ageing. Moreover, soy-based natural ester produces quantitatively However, the prediction of Doernenburg, Rogers ratio and IEC
more stray gases than sunflower oil-based natural ester (1383 and ratio for mineral oil are not applicable to vegetable oils sample.
176 ppm, respectively). Some DGA results may fall outside the gas ratio methods and no
Experimental results of [48] show that mineral oil (Shell Diala) diagnosis can be given. Nevertheless, all samples were correctly
and natural ester (sunflower oil based) exhibit a similar spectral interpreted with the Duval triangle method.
pattern in frequency-domain dielectric responses. That is, as the In [53, 54], experimental studies on performances of a mineral
frequency increases, the dielectric losses decrease. However, oil (Nynas Gemini X) and a natural ester (Cargill FR3) in terms of
natural ester exhibits higher capacitance, loss and polarisation fault gas generation were performed. Laboratory simulated faults
current. DGA results indicate that C2H2 remains as the main gas include electrical sparks, electrical PD and high-temperature
generated due to low energy arc and followed by H2. The thermal hotspot types were simulated. Moreover, DGA in oil due to
individual gas concentration of all considered gases (H2, CH4, the faults were analysed by an on-line gas monitoring system
C2H4, C 2H6 and C 2H2) in mineral oil is greater than that of natural (TM8) and by a laboratory analysis. The results obtained with both
ester. For thermal fault (120°C), C2H6, CH4 and CO remain as the methods are comparable, within a deviation of 30% under all the
main combustible gases in mineral oil. In the case of the natural faults. The main findings of this work are summarised in Table 6.
ester, CH4 is very low. CO concentration in oil samples aged with Anyway, during the studies, some new questions were found and
pressboard is larger than that of oil aged standalone. Significant therefore further studies are required. For example, the sealing state
increase of the CO2 concentration with ageing indicates the severe of the system needs to be improved to provide no or less gas
leakage for the PD test (the test duration needs to be extended to
degradation of cellulose, because CO2 is mainly produced by
more than 4 days because the gas generation rate is too small).
cellulose degradation. Compared to natural ester, mineral oil Furthermore, a more accurate temperature measurement method is
produces more gases in volume due to low energy discharges and needed in order to evaluate the thermal fault using average
overheating of oil. Duval triangle method correctly diagnosed low temperature. On the other hand, degradation effect of the insulating
energy electrical discharges in mineral and natural ester. However, paper in the oil was not analysed.
for thermal faults in region T1 for both mineral oil and natural ester In [55], experimental results of the application of DGA
accurately diagnoses the fault, except for sample on natural ester. technique to natural ester different moisture levels were presented.
IEC gas ratio method accurately diagnoses the faults as D1 and T1 The covered faults were discharge (in accordance with ASTM
of each sample for each type of oil. For the ratio calculation, zero D877) and overheating in four different commercial insulating oils,
values are replaced with a value just below the detection limit of
5444 IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 24, pp. 5441-5448
© The Institution of Engineering and Technology 2019
Table 6 Gases generated in mineral oil and natural ester larger than those from mineral oil, 25–280 times and 2 times,
under different type of faults respectively. The amount of CH4 from mineral oil is 4 times larger
Electrical fault Thermal fault than from natural ester as well as C2H4 (1.2 times) and CO (4
Spark PD Hot spot times). On the other hand, for thermal deterioration of insulating
Gemini X — gas formation key gases: CH4 and oil added with transformer materials, the amount of C2H6 in natural
rates slightly C2H4 ester is 2 times larger than mineral oil. When the insulating oil
lower added with transformer materials was overheated to 150°C, the
Duval triangle method
results obtained were similar. However, due to only thermal stress
correctly recognises
is analysed in both cases, it is necessary to study electrical stress
all thermal faults
such as partial and arc discharges in order to establish a more
FR3 25% fault gases at the same PD generates less precise criterion.
higher amplitude greater amount of gases at In [58], experimental results of DGA generated in a
considering the gases are higher temperatures conventional naphthenic-based mineral oil, a conventional natural
same sparking generated (due to (>300°C) ester (FR3) and a camellia-based natural ester under typical
energy higher PD thermal faults at temperatures from 90 to 800°C were presented.
repetition rate) The characteristic dissolved gases in natural ester differ from those
recognised in the key gases: CH4, in mineral oils under thermal faults. In the mineral oil, the
adjusted Duval C2H4, C2H6 and CO characteristic gas is CH4, in FR3, the characteristic gas is C2H6,
triangle and in camellia oil, the characteristic gases are H2 and C2H6. The
gas generation a little revision should traditional IEC ratio criterion for DGA is not suitable for fault
under PD faults be made for the Duval diagnosis of transformers filled with natural ester. Hence,
are determined by triangle in order to traditional diagnosis methods for mineral oil filled transformers are
energy instead of recognise the thermal needed to be reconsidered when they are applied to natural ester
PD amplitude or faults fluid filled transformer.
PD numbers Gas generation of natural monoesters (derived from palm and
common similar amount gas generation gas generation rates rice bran oils) with and without C–C double bond in their
feature of gases for a rates increase for are mainly molecular structures under low thermal faults (ranging from 200 to
fault PD with higher determined by the hot 300°C) was analysed in [59]. The diagnosis of thermal or electrical
Duval triangle amplitudes spot temperature faults in mineral oil [60] and their applicability for natural ester oils
sparking faults rather than the were evaluated [55], concluding that Duval triangle 3 shows the
as low energy average temperature best performance based on the number of correct predictions.
discharges However, under low thermal faults where the amount of generated
gases is very low, the conventional gas ratio methods cannot be
applied. Instead, an alternative gas ratio is used, e.g. the ratio of
one mineral and three natural esters (corn oil, sunflower oil and C2H6/CH4 that was used in [61] to estimate the overheating
FR3 – soy oil). Five key dissolved hydrocarbon gases (H2, CH4, temperature within the range of 100–200°C. The proposed ratio of
C2H6, C2H4 and C2H2) were analysed and compared with mineral experimental result in this work is CO/CH4 within the range of
oil fault gases. The classic Duval triangle 1, modified Duval 200–300°C. Both alternative gas ratios estimate the fault
triangle 3, Rogers ratio, Doernenburg and IEC 60599 methods temperature. However, further investigations are required to obtain
were applied to predict the fault type. However, the direct a more comprehensive understanding of how the ratio of CO/CH4
application of classic DGA fault interpretation methods, developed varies to the change in the overheating temperature. Other
for mineral oils, to non-mineral oils needs some modifications. contributions of [59] are:
Moreover, as suggested in [14], some zone boundaries of the
modified Duval triangles may need further adjustments for natural • Types of gases produced by natural monoesters without C–C
esters. Some conclusions reached were: double bond are more similar to those produced by mineral oil,
rather than those produced by natural monoesters having C–C
• Under thermal stress, CH4 in natural ester was produced at double bond.
significant lower concentration than mineral oil, furthermore the • Natural monoesters tend to produce a higher amount of CO, but
H2 generation is not really significant. Compared to mineral oil, less amount of CH4 compared to mineral oils.
natural esters generated much more C2H6 during thermal • Natural monoesters having C–C double bond tend to generate
degradation. Moreover, in mineral oil more CH4 was generated noticeable amount of C2H6, accompanied by small amount of
than the other key gases. H2, whereas those without C–C double bond and mineral oil do
• Under electrical stress, higher quantities of C2H2 and H2 are not. Therefore, C2H6 could be used as the key gas for the low
generated in natural esters compared to other gases. The thermal faults in natural ester oils (only applies for the oils
concentration of these gases increased with increasing water having C–C double bond).
content in each oil sample. For natural esters, C2H2 is the key
gas that will indicate a defect or fault by electrical discharge. In [62], the experimental results on the effects of thermal ageing on
This gas also indicates electrical faults in power transformers the properties of natural ester from palm oil and Kraft paper
insulated with mineral oil [56]. C2H6, which was the gas formed composite insulation were reported. All samples were subjected to
in the highest concentration in thermal stress, is formed in very thermal ageing condition for duration of 336–672 h at 120 and
low concentration when compared to the other gases generated 150°C in a controllable oven. DGA was conducted, indicating that
under electrical stress. thermal ageing in ester-paper composite insulation released H2,
CH4, C2H2, C2H4, C2H6 and CO. Fig. 2 shows the concentration of
In [57], five generated gases (H2, CH4, C2H4, C2H6, C2H2) in relevant combustible gases dissolved in the ester from palm
mineral oil and natural ester under thermal stress with and without samples aged with paper at different ageing temperature and time.
added transformer materials were compared. Both oils were In this particular ageing test, it is clear that CO appeared in all of
deteriorated for 1600 h at 150°C and collected every 200 h to the samples. CH4 was detected in the first hours of ageing and
analyse changes in the contents of dissolved gases. As a result, began to decrease rapidly to zero. C2H6 was detected at all ageing
almost no H2 and C 2H2 were detected and the amount of TCG was temperature and time.
almost the same. Thermal deterioration of insulating oil shows that
the concentrations of C2H6 and TCG from natural ester fluid were

IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 24, pp. 5441-5448 5445
© The Institution of Engineering and Technology 2019
4 Current status of the normative described in Method C of IEC 61125:2018 [64]. Anyway, there is
an IEC working group that is currently working on the
The main current standards for dielectric fluids and the development of natural esters guidelines for maintenance and use
interpretation of generated gases in transformers come from IEEE in electrical equipment [65].
and IEC. TB 436 of CIGRE [10] covers the in-service experience of
IEEE Std C57.147-2018 [11] makes reference to IEEE Std synthetic esters, natural esters and silicone fluids (alternative
C57.155-2014 [12], where it addresses the interpretation of the fluids) and their performance. Mention the physical and chemical
DGA results and the recommended actions based on the differences between alternative fluids and mineral oil, as well as
interpretation of DGA results for natural ester and synthetic ester- research and laboratory tests on the electrical properties of
immersed transformers. However, this guide is based on a limited alternative fluids. It also gives a complete list of the IEC, IEEE and
amount of in-service DGA results and several laboratory-based ASTM standards to determine the test method to be applied to each
experimental datasets. Experimental laboratory data can give fluid.
insight into the insulating liquid's behaviour, but it does not
replicate actual in-service conditions and fault conditions found in
an operating transformer. Thus, more field data relative to natural 5 Discussion
esters are warranted. This standard establishes that the key ASTM Although up to the present, there is only one IEEE standard
tests related to gases generated in liquid-insulated transformers for exclusive for the interpretation of gases generated in transformers
natural ester liquids are the same methods used for mineral oil. immersed in vegetable oils, it is based on a limited amount of in-
Furthermore, the same fault gases are produced in both natural service DGA results and several laboratory-based experimental
esters and mineral oil, although the circumstances and quantities in datasets. Moreover, IEC is working on a guideline related to the
which they are produced are sometimes different. There are three maintenance and use of vegetable oils in electrical equipment.
principal differences between the fluids caused by the difference in The investigations carried out so far converge that for these
chemical structure [11]. After the initial operation of a transformer fluids to be fully accepted and used in transformers of any capacity,
under normal conditions, the ethane levels can rise to a few they must prove to be reliable, economical (in contrast to other
hundred ppm. This generation is considered ‘stray gassing’ and not market alternatives) and offer high thermal and electrical
a fault condition. Unlike mineral oil, carbon oxides can be performance throughout their lifetime. There are some new
produced from the pyrolysis of natural ester liquid in amounts and insulating compositions, based on modified vegetable oils, which
proportions that can mask or confuse the carbon oxide production has been developed to present very similar properties to those of
from the pyrolysis of cellulose. Moreover, the same hydrocarbon the standard insulating mineral oils and to meet the specific needs
gases are produced from heating but the proportions are different of current distribution transformers [66].
than those produced from mineral oil. Results of accelerated ageing studies at the same temperature
IEC 62770:2013 standard [63] does not mention the gas and time exposure conditions [67] show that natural esters reduce
formation in natural esters with respect to mineral oils when a fault cellulose ageing rates [68]. Thus, it is possible to achieve a longer
occurs. On the contrary, it is limited to establishing general lifetime compared to that obtained in transformers immersed in
specifications and limit values for different properties (physical, mineral oil. Moreover, retro filling of mineral oil filled
chemical, electrical, among others) of natural esters and their test transformers by natural esters promises to be the most financially
methods. This standard is applicable only to unused natural esters. viable technique for enhancing transformer capacity and life with
Reclaimed natural esters and natural esters blended with non- added advantages of reduced environmental impact, reduced
natural esters fluids are beyond the scope of this standard. As in footprint and improved safety. This is due to the fact that natural
other documents, IEC does not recommend the use of natural esters esters have an inherent property of absorbing moisture from the
for electrical equipment that is open to the atmosphere because paper insulation that is impregnated with. Life of the transformer is
these fluids are prone to rapid oxidation. Oxidation stability of enhanced owing to its unique ability of continuously drying the
these fluids can be assessed with modifications to the procedure as insulating material through the hydrolysis process [69].
DGA interpretation techniques, used for many decades in
transformers immersed in mineral oil, must be adapted for
vegetable oils. New criteria for diagnosis must be established
through gas ratios, key gases and Duval triangles. Also,
complementary research on ageing and lifetime expectancy in
transformers in service (real scale) are necessary [70].
Table 7 compares the amount of generated gases in oils of
different origin under different types of fault. Table 7 shows that, in
five out of seven presented works, differences in the generated
gases under different faults were found between oils of different
origin. Table 8 analyses the applicability of the different DGA
interpretation methods in natural esters. Note that different authors
prove the applicability of Duval triangle 1 which is for mineral oils
Fig. 2 DGA results on ester from palm oil for faults interpretation on natural esters. Correct predictions can
be interpreted as the limits that define the different faults zones of

Table 7 Amount of generated gases in oil of different origin for different type of faults
Ref. Compares With Under Generated gases
Similar Different
[47] impregnated paper in natural ester impregnated paper in synthetic ester thermal ageing •
[47] soy-based natural ester synthetic ester thermal fault •
[47] soy-based natural ester sunflower-based natural ester thermal ageing •
[47] mineral oil natural ester low energy discharge and •
overheating
[58] conventional natural ester (FR3) camellia-based natural ester thermal fault •
[59] natural monoesters without C–C double bond mineral oil low thermal fault •
[59] natural monoesters without C–C double bond natural monoesters with C–C double low thermal fault •
bond

5446 IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 24, pp. 5441-5448
© The Institution of Engineering and Technology 2019
Table 8 Applicability of different DGA interpretation methods in natural esters
Ref. Oil type Under Method that applies
Duval triangle 1 Duval triangle 3 IEC gas ratio Key gas
[47] soy-based natural ester thermal fault (T1) •
synthetic ester
[48] sunflower-based natural ester low energy electrical discharge • •
[48] sunflower-based natural ester thermal fault (T1) • •
for some samples
[50, 51] natural ester electrical fault (low energy discharge) • •
[55] natural ester discharge and overheating •
[59] natural monoesters low thermal fault •

Table 9 Key gases for a fault in natural esters

Ref. Oil type Under Key gas
H2 CH4 C2H6 C2H4 C2H2
[48] sunflower-based natural ester thermal ageing • • •
[48] soy-based natural ester thermal ageing • •
[50, 51] natural ester electrical fault (low energy discharge) • •
[50, 51] natural ester low thermal fault • •
[57] natural ester thermal ageing •
[58] natural ester (FR3) thermal fault •
[58] camellia-based natural ester thermal fault • •
[59] natural monoesters with C–C double bond low thermal fault • •

the triangle do not require modifications and vice versa. Table 9 gases present in natural esters at times of faults to form a
shows the key gases produced in natural esters before a specific knowledge base that helps decision making, in an accurate and
type of fault. However, further analysis is necessary in order to reliable way.
establish a more precise criterion. The diversity of results obtained
and variations between oils make insufficient the conclusions 7 Acknowledgments
reached so far. For example, in [47] the IEC gas ratio method
identifies thermal faults, being applicable for both mineral oils and This research was supported by the ANPCYT, CONICET,
natural esters. However, in the case presented in [58], the Universidad Nacional del Centro de la Provincia de Buenos Aires
traditional criteria of IEC gas ratio were not suitable to identify a and Universidad Nacional de La Plata (Argentina).
thermal fault in natural esters.
8 References
6 Conclusion [1] Heathcote, M.J.: ‘J & P transformer book’ (Newnes, Oxford, UK, 2007, 13th
edn.)
Vegetable oils are a possible substitute for the traditionally used [2] Horning, M., Kelly, J., Myers, S., et al.: ‘Transformer maintenance guide’
mineral oil, mainly when high flash point characteristics and (Transformer Maintenance Institute, Akron, OH, USA, 2004, Third edition)
greater environmental compatibility are required. Their behaviour [3] CIGRE Technical Brochure 393: ‘Thermal Performance of Transformers’.
and performance during different types of faults are being Working Group D1.31, 2010
[4] ‘IEEE Recommended Practice for Installation, Application, Operation, and
analysed. Maintenance of Dry-Type Distribution and Power Transformers’. IEEE Std
From the gathered data in the present overview, it can be C57.94-2015 (Revision of IEEE Std C57.94-1982), 2016, pp. 1–29
concluded that natural esters generate the same gases under fault [5] Nunn, T.: ‘A comparison of liquid-filled and dry-type transformer
conditions as mineral oils. Furthermore, at the proposed test technologies’. 2000 IEEE-IAS/PCA Cement Industry Technical Conf. Conf.
Record (Cat. No.00CH37047) 2000 IEEE-IAS/PCA Cement Industry
conditions by different authors, certain faults have a direct relation Technical Conf. Conf. Record (Cat. No.00CH37047), Salt Lake City, UT,
with the DGA interpretation techniques applied to mineral oil. USA, 2000, pp. 105–112
However, other faults need a limits revision because they cause a [6] CIGRE Technical Brochure 254: ‘Dielectric response methods for diagnostics
wrong diagnosis. of power transformers’. Task Force D1.01.09 CIGRE, 2004
[7] CIGRE Technical Brochure 227: ‘Life Management Techniques For Power
The main conclusions of the differences between natural esters Transformers’. Working Group A2.18, CIGRE, 2003
and mineral oils in terms of gas generation are: [8] Lashbrook, M., Kuhn, M.: ‘The use of ester transformer fluids for increased
fire safety and reduced costs’. CIGRE Session, 2012
• For very similar faults, there is a great diversity of results. [9] Gomez, N.A., Abonia, R., Cadavid, H., et al.: ‘Chemical and spectroscopic
characterization of a vegetable oil used as dielectric coolant in distribution
Results obtained are strongly dependent of the set-up used, so transformers’, J. Braz. Chem. Soc., 2011, 22, (12), pp. 2292–2303
laboratory tests must be considered as comparative between [10] CIGRE Technical Brochure 436: ‘Experiences in service with new insulating
types of oils rather than considering absolute values. liquids’. Working Group A2.35 CIGRE, 2010, pp. 1–95
[11] ‘IEEE Guide for Acceptance and Maintenance of Natural Ester Insulating
• For low thermal fault, CH4, C2H6 and C2H4 are produced in Liquid in Transformers’. IEEE Std C57.147-2018 (Revision of IEEE Std
higher concentrations and different proportions. C57.147-2008), 2018, pp. 1–47
• CO and CO2 are produced from the natural ester overheating, [12] ‘IEEE Guide for Interpretation of Gases Generated in Natural Ester and
Synthetic Ester-Immersed Transformers’. IEEE Std C57.155-2014, 2014, pp.
generating imprecise diagnosis of the state of solid insulation. 1–52
• Key gases are the same for discharge faults and partial [13] Khan, I.U., Wang, Z., Cotton, I., et al.: ‘Dissolved gas analysis of alternative
discharges and differ for thermal faults, being C2H6 for natural fluids for power transformers’, IEEE Electr. Insul. Mag., 2007, 23, (5), pp. 5–
14
ester and C2H4 for mineral oil. [14] Duval, M.: ‘The Duval triangle for load tap changers, non-mineral oils and
low temperature faults in transformers’, IEEE Electr. Insul. Mag., 2008, 24,
The DGA has proven to be applicable to vegetable oils. Historical (6), pp. 22–29
[15] Duval, M.: ‘Calculation of DGA limit values and sampling intervals in
information with this type of technique must begin to be transformers in service’, IEEE Electr. Insul. Mag., 2008, 24, (5), pp. 7–13
constructed. Further real field values are needed to determine the

IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 24, pp. 5441-5448 5447
© The Institution of Engineering and Technology 2019
[16] Dornenburg, E., Strittmatter, W.: ‘Monitoring oil-cooled transformers by gas- Éster natural (Aceite Vegetal Aislante)’. 2015 XVI Encuentro Regional
analysis’, Brown Boveri Rev., 1974, 61, (5), pp. 238–247 Iberoamericano de CIGRÉ, Puerto Iguazú, 17 al 21 de mayo de 2015
[17] IEC 60599:2015: ‘Mineral oil-filled electrical equipment in service - [46] Rafiq, M., Lv, Y.Z., Zhou, Y., et al.: ‘Use of vegetable oils as transformer
Guidance on the interpretation of dissolved and free gases analysis’ oils–a review’, Renew. Sust. Energy Rev., 2015, 52, pp. 308–324
[18] Duval, M.: ‘A review of faults detectable by gas-in-oil analysis in [47] Bandara, K., Ekanayake, C., Saha, T.K.: ‘Compare the performance of natural
transformers’, IEEE Electr. Insul. Mag., 2002, 18, (3), pp. 8–17 ester with synthetic ester as transformer insulating oil’. 2015 IEEE 11th Int.
[19] ‘IEEE Guide for the Interpretation of Gases Generated in Oil-Immersed Conf. on the Properties and Applications of Dielectric Materials (ICPADM),
Transformers’. IEEE Std C57.104-2008 (Revision of IEEE Std Sydney, NSW, Australia, 2015, pp. 975–978
C57.104-1991) [48] Bandara, K., Ekanayake, C., Saha, T.K.: ‘Comparative study for
[20] Duval, M.: ‘Dissolved gas analysis: it can save your transformer’, IEEE understanding the behaviour of natural ester with mineral oil as a transformer
Electr. Insul. Mag., 1989, 5, (6), pp. 22–27 insulating liquid’. 2014 IEEE Conf. on Electrical Insulation and Dielectric
[21] Duval, M., Baldyga, R.: ‘Stray gassing of FR3 oils in transformers in service’. Phenomena (CEIDP), Des Moines, IA, USA, 2014, pp. 792–795
76th Doble Int. Client, Conf., Boston, MA, USA, 2009 [49] IEC 60567:2011, Edition 4.0: ‘Oil-filled electrical equipment – Sampling of
[22] Duval, M.: ‘New frontiers of DGA interpretation for power transformers and gases and analysis of free and dissolved gases – Guidance’
their accessories’ (Society of Electrical and Electronics Engineers (SEEEI), [50] Perrier, C., Marugan, M., Saravolac, M., et al.: ‘DGA comparison between
Israel, 2012), pp. 1–8 ester and mineral oils’, in ‘2011 IEEE international conference on dielectric
[23] Piegari, J.S., Álvarez, R.E., Catalano, L.J.: ‘Herramienta de gestión y análisis liquids’. 2011 IEEE Int. Conf. on Dielectric Liquids, Trondheim, Norway,
para los ensayos de DGA en transformadores de potencia inmersos en aceite 2011, pp. 1–4
mineral’. 2015 XVI Encuentro Regional Iberoamericano de CIGRÉ, Puerto [51] Perrier, C., Marugan, M., Beroual, A.: ‘DGA comparison between ester and
Iguazú, 17 al 21 de mayo de 2015 mineral oils’, IEEE Trans. Dielectr. Electr. Insul., 2012, 19, (5), pp. 1609–
[24] Singh, S., Joshi, D., Bandyopadhyay, M.N.: ‘Software implementation of 1614
duval triangle technique for DGA in power transformers’, Int. J. Electr. Eng., [52] Hamid, M.H.A., Ishak, M.T., Ariffin, M.M., et al.: ‘Dissolved gas analysis
2011, 4, (5), pp. 529–540 (DGA) of vegetable oils under electrical stress’. 2017 Int. Conf. on High
[25] Reddy, N.V.: ‘Dissolved Gas Analysis–An Early Identification of Faults in Voltage Engineering and Power Systems (ICHVEPS), Sanur, Indonesia, 2017,
High Voltage Power Equipment using MATLAB GUI’. MTech thesis, pp. 29–34
National Institute of Technology, Rourkela, 2014 [53] Li, S.: ‘Study of Dissolved Gas Analysis under Electrical and Thermal
[26] Siva Sarma, D.V.S.S., Kalyani, G.N.S.: ‘ANN approach for condition Stresses for Natural Esters used in Power Transformers’. MPhil thesis, The
monitoring of power transformers using DGA’. 2004 IEEE Region 10 Conf. University of Manchester, 2012
(TENCON 2004), Chiang Mai, Thailand, 2004, vol. 3, pp. 444–447 [54] Wang, Z., Wang, X., Yi, X., et al.: ‘Gas generation in natural ester and
[27] Sarria-Arias, J.T., Guerrero-Bello, N.A., Rivas-Trujillo, E.: ‘Estado del arte mineral oil under partial discharge and sparking faults’, IEEE Electr. Insul.
del análisis de gases disueltos en transformadores de potencia’, Revista Mag., 2013, 29, (5), pp. 62–70
Facultad de Ingeniería (Fac. Ing.), 2014, 23, (36), pp. 105–122 [55] Wilhelm, H., Santos, C., Stocco, G.: ‘Dissolved gas analysis (DGA) of natural
[28] Faiz, J., Soleimani, M.: ‘Dissolved gas analysis evaluation in electric power ester insulating fluids with different chemical compositions’, IEEE Trans.
transformers using conventional methods a review’, IEEE Trans. Dielectr. Dielectr. Electr. Insul., 2014, 21, (3), pp. 1071–1078
Electr. Insul., 2017, 24, (2), pp. 1239–1248 [56] Muhamad, N.A., Phung, B.T., Blackburn, T.R.: ‘Dissolved gas analysis for
[29] Gouda, O.E., El-Hoshy, S.H., El-Tamaly, H.H.: ‘Proposed heptagon graph for common transformer faults in soy seed-based oil’, IET Electr. Power Appl.,
DGA interpretation of oil transformers’, IET Gener. Transm. Distrib., 2018, 2011, 5, (1), pp. 133–142
12, (2), pp. 490–498 [57] Dongjin, K., Kyosun, K.: ‘The criteria of dissolved gas analysis in natural
[30] Moodley, N., Gaunt, C.T.: ‘Low energy degradation triangle for power ester fluid’. 2012 Int. Conf. on Condition Monitoring and Diagnosis (CMD),
transformer health assessment’, IEEE Trans. Dielectr. Electr. Insul., 2017, 24, Bali, Indonesia, 2012, pp. 901–904
(1), pp. 639–646 [58] Liu, Y., Li, J., Zhang, Z.: ‘Fault gases dissolved in vegetable insulating oil
[31] Duval, M., Lamarre, L.: ‘The Duval pentagon-a new complementary tool for under electrical faults’. 2013 IEEE Conf. on Electrical Insulation and
the interpretation of dissolved gas analysis in transformers’, IEEE Electr. Dielectric Phenomena (CEIDP), Shenzhen, People's Republic of China, 2013,
Insul. Mag., 2014, 30, (6), pp. 9–12 pp. 198–201
[32] Duval, M., Lamarre, L.: ‘The new Duval Pentagons available for DGA [59] Rajab, A., Tsuchie, M., Kozako, M., et al.: ‘Low thermal fault gases of
diagnosis in transformers filled with mineral and ester oils’. 2017 IEEE various natural monoesters and comparison with mineral oil’, IEEE Trans.
Electrical Insulation Conf., (EIC), Baltimore, MD, USA, 2017, pp. 279–281 Dielectr. Electr. Insul., 2016, 23, (6), pp. 3421–3428
[33] Tenbohlen, S., Höhlein, I., Lukas, M., et al.: ‘Dynamic behaviour of fault [60] Muhamad, N.A., Phung, B.T., Blackburn, T.R., et al.: ‘Comparative study and
gases and online gas sensors’. CIGRE Session, 2016 analysis of DGA methods for transformer mineral oil’. IEEE Power Tech
[34] Lashbrook, M., Al-Amin, H., Martin, R.: ‘Natural ester and synthetic ester 2007, Lausanne, 2007, pp. 45–50
fluids, applications and maintenance’. 2017 10th Jordanian Int. Electrical and [61] Tsuchie, M., Kozako, M., Hikita, M., et al.: ‘Modeling of early stage partial
Electronics Engineering Conf. (JIEEEC), Amman, Jordan, 2017, pp. 1–6 discharge and overheating degradation of paper-oil insulation’, IEEE Trans.
[35] Tenbohlen, S., Koch, M., Vukovic, D., et al.: ‘Application of vegetable oil- Dielectr. Electr. Insul., 2014, 21, (3), pp. 1342–1349
based insulating fluids to hermetically sealed power transformers’. CIGRE [62] Suwarno, A.S: ‘Investigation on thermal aging of ester from palm oil and
Session, 2008, pp. 24–29 kraft paper composite insulation system for high voltage transformer’,
[36] Gockenbach, E., Borsi, H.: ‘Natural and synthetic ester liquids as alternative WSEAS Trans. Environ. Dev., 2017, 13, pp. 75–84
to mineral oil for power transformers’. 2008 Annual Report Conf. on [63] IEC 62770:2013: ‘Fluids for electrotechnical applications - Unused natural
Electrical Insulation and Dielectric Phenomena (CEIDP 2008), Quebec, QC, esters for transformers and similar electrical equipment’
Canada, 2008), pp. 521–524 [64] IEC 61125:2018: ‘Insulating liquids - Test methods for oxidation stability -
[37] Marci, M., Kolcunová, I., Kurimský, J.: ‘Dielectric properties of natural Test method for evaluating the oxidation stability of insulating liquids in the
esters’. 2011 10th Int. Conf. on Environment and Electrical Engineering delivered state’
(EEEIC), Rome, Italy, 2011, pp. 1–4 [65] IEC 62975 ED1: ‘Natural esters - Guidelines for maintenance and use in
[38] Perrier, C., Beroual, A.: ‘Experimental investigations on mineral and ester oils electrical equipment’
for power transformers’. Conf. Record of the 2008 IEEE Int. Symp. on [66] Bertrand, Y., Lauzevis, P.: ‘Development of a low viscosity insulating liquid
Electrical Insulation, 2008 (ISEI 2008), Vancouver, BC, Canada, 2008, pp. based on natural esters for distribution transformers’. 22nd Int. Conf. and
178–181 Exhibition on Electricity Distribution (CIRED 2013), Stockholm, Sweden,
[39] Reffas, A., Idir, O., Ziani, A., et al.: ‘Influence of thermal ageing and 2013, pp. 1–4
electrical discharges on uninhibited olive oil properties’, IET Sci. Meas. [67] Hopkinson, P.: ‘Panel discussion to focus on natural ester fluids’, The Line,
Technol., 2016, 10, (7), pp. 711–718 2006, 240, pp. 8–11
[40] IEC 60156:1995: ‘Insulating liquids - Determination of the breakdown [68] Luksich, J.: ‘Loading guide A and B factors for envirotemp FR3 fluid and
voltage at power frequency - Test method’ thermally upgraded kraft insulation’. Report TP03-DR-009, Cooper Power
[41] ‘Cargill, Envirotemp ‘FR3TM Fluid, Data Sheet’ System, 2003
[42] ‘M&I Materials MIDEL® eN TDS 5 Product Overview’ [69] Pillai, R., Havaldar, F., Chitnis, C.: ‘Natural esters for life and capacity
[43] Suleiman, A.A., Muhamad, N.A., Bashir, N., et al.: ‘Moisture effect on enhancement of distribution transformers’, CIRED - Open Access Proc. J.,
conductivity of kraft paper immersed in power transformer vegetable-based 2017, 2017, (1), pp. 355–358
insulation oils’, IET Gener. Transm. Distrib., 2017, 11, (9), pp. 2269–2274 [70] Fernando-Navas, D., Cadavid-Ramírez, H., Echeverry-Ibarra, D.F.:
[44] CIGRE Technical Brochure 296: ‘Recent developments in DGA ‘Aplicación del aceite dieléctrico de origen vegetal en transformadores
interpretation’. Joint Task Force D1.01/A2.11 CIGRE, 2004 eléctricos’, Ing. Universidad, 2012, 16, (1), pp. 201–223
[45] Sbravati, A., Martins, M.N., Gomes, J.C., et al.: ‘Límites de investigación,
tratamiento y regeneración de transformadores llenos con fluidos basados en

5448 IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 24, pp. 5441-5448
© The Institution of Engineering and Technology 2019