Presented By:
David Jiménez de la Plata
Mónica Pérez
María de Haro
Alfonso de Pablo
TJH2b Latina S.L.
TechCon 2023
Abstract
This paper goes on to describe gas generation in wind farm transformers. (WTTs – Wind Turbine Transformers). Local transformers, those directly linked to the turbine of a wind farm, generate abnormally high amounts of dissolved gases. This is virtually independent of the type of transformer (pad mount or distribution) and the type of fluid (mineral oil or synthetic ester) contained in the transformer.
The causes of this anomalous gas formation are not fully understood yet, although some possibilities have been proposed, such as a breakage or delamination of the core, harmonics, low voltage fault ride through, over-voltage size limitation, variable loading cycle or switching surges and transient overvoltage.
Nor can it be ruled out that the cause of these gases could be the interaction of the transformer with the electrical system: windfarm design. These causes are briefly discussed in the paper.
Introduction
It is an undisputed fact that wind energy has recently attracted a great deal of interest worldwide. According to the Global Wind Report (1), by the end of 2021, there is 743 GW of wind power capacity worldwide of which 136 GW are installed in the USA.
Despite this enormous interest in wind energy, gas generation in windfarm transformers has not been adequately addressed in international standards. IEEE Standard C57.104:2019 only mentions these transformers in Sub-clause 1.2, Limitations to use this document (2):
The user of this guide must consider the internal components of the transformer when choosing to apply the guide to specific transformer types such as wind turbine transformers or network transformers.
Note that the initial analysis of the data received in preparation of this guide clearly indicated that DGA results from transformers identified as wind turbine or network transformers often had excessively elevated typical values, in some case two orders of magnitude higher that the rest of the DGA database. There was no information to indicate if the transformers were healthy or under extreme distress. Therefore, the DGA data subset from wind turbine and network transformers was not included in the study to obtain the norms used in this guide. Consequently, if the user of this guide applies the norms presented here in to such transformer types, the result may often be high DGA status levels.
In the case of wind turbine transformers in particular, they have special operating conditions in that they are closely connected to power electronics and have frequently and widely fluctuating loads. Some of the earlier generation designs of these transformers were not adequately robust for such operating conditions, and hence the often elevated gas values are seen.
IEC Standard 60599:2022 provides a Table (Table A.5) with the ranges of 90 % typical concentration values observed in WTTs. However, table footnote c) reads (3):
Hydrogen values obtained statistically can be due to the high electrical stress (transients and harmonics) to which those units are exposed. For this reason, the values for hydrogen can be considered “typical” for statistical purposes, but they could not be inherently attributable to fault absence. Available experience worldwide on wind turbine units is not sufficiently large, yet, to define which statistical values can be considered as fault-free.
Windfarm Layout
To understand gas formation in wind transformers (WTT), it is necessary to understand what a wind farm looks like: in a wind farm there are two different types of step-up transformers. First, there is a transformer, usually referred to as a “local transformer” directly connected to each turbine. In the old days, this transformer was installed in the nacelle, but nowadays the tendency is to install it at the bottom of the turbine in order to facilitate its maintenance. Generally, in Europe, this is a dry transformer, so it does not generate gases. In the Americas, and sometimes in other countries, this transformer can be a pad mount transformer (switchgear in the transformer oil) or a distribution transformer (switchgear outside). Transformers may be indoor built at the top of the tower (highly thermally stressed due to restricted clearance and proximity of rotating machines and converter), indoor built at the base of tower (medium thermal stress), outdoor installations (standard thermal stress) or connected to asynchronous or synchronous generators (AC/DC converter can generate severe harmonic disturbances with high impact on dielectric performances). Typically, these transformers rise voltage from 600 V to 34.5 kV. In other geographical regions, these local transformers may be filled with mineral oil, silicone or synthetic ester. In general, these transformers operate at 12/20 or 12/33 kV. Second type of transformers in a wind farm are named “collector transformers” because they collect the power of a number of local transformers and rise the voltage to the grid voltage. Since these transformers are not subject to the transients and other electrical disturbances of local transformers, they can be diagnosed following the criteria set out in Tables A.1 and A.2 of IEC 60599.
Dissolved Gas Analyses in WTTs
Pad Mount Transformers
The dissolved gas content in the pad mount transformers of 6 wind farms installed in three different American countries of an electricity company has been analyzed. One of these wind farms was built in 2008, three in 2009, one in 2011 and one in 2014. In total, there are 1046 analyses of 537, 12 kV/34.5 kV, 1700 kVA transformers, all containing mineral oil Type I according to ASTM D3487 (4).
The vast majority of transformers are of the same design and belong to the same manufacturer. Statistical evaluation, considering only the latest result, of these analyses shows that the H2, CH4 and C2H6 content is extremely high, as shown in Table 1.
As for the concentrations of other gases, ethylene appears in concentrations above what is considered normal, according to IEEE standard C57.104;2019, in very few cases (1 case in windfarm #1, 3 cases in windfarm #2, 1 case in windfarm #3, 8 cases in windfarm #4, none at the windfarms #5 and #6). Acetylene was found in concentrations higher than 10 ppm in 4 windfarms (2 in windfarm #4 and 1 in windfarms #2 and #3). However, not all transformers in all wind farms behave in the same way, nor is there a relationship between the year of installation and the gases generated, as shown in Table 1.


Unfortunately, we do not have the equivalent farm hours for these facilities, but it can be seen that the transformers that generate the highest concentrations of gases are those in wind farm #6, which are the most recently installed.
Due to the lack of normative references, the interpretation of the DGA results is difficult. Being aware of these difficulties, an attempt has been made to interpret them according to Duval’s pentagon (5), with the result shown in Fig. 1 and Table 2.
Most of the defects, according to Duval Pentagon 1, are of the partial discharge or stray gassing type. In wind farms 1 to 4 there are some defects of type T1 (low temperature thermal), in wind farms 2, 3 and 4 there are one or two defects of type T3 (high temperature thermal) and in wind farm 4 there is only one defect of type D2 (high energy arcing).

Distribution transformers
An alternative to pad mount transformers, mainly used in Europe, is to use step-up distribution transformers. In our laboratory, we have compiled the results of the analysis of dissolved gases in the fluids of 165 local distribution transformers, installed in 5 wind farms located in Spain, Turkey, and Peru. DGA results are summarized in Table 3.

As in the previous case of pad mount transformers, not all transformers in all wind farms behave in the same way: statistically, the transformers in Peru can be considered in relatively good condition (there are only two transformers whose gases exceed the values considered normal), the transformers in Spain 1, 2 and 3 have high concentrations of the four gases considered H2, CH4, C2H6 and C2H4 and Turkey’s transformers, although the most recently installed, contain large quantities of all gases except ethylene. As in the case of the pad mounts, no significant amounts of acetylene are observed in any of the wind farms.
The transformers were made by three different manufacturers, containing between 710 and 1775 kg of fluid. Peru transformers contain mineral oil and all other transformers contain Midel 7131 synthetic ester. The voltage of the transformers in Turkey is 690 V / 34.000 V, those in Peru are 12 kV / 33 kV and those of the three windfarms in Spain are 12 KV / 20 kV.
In this case, the interpretation of dissolved gas analyses is more complicated than with pad mount transformers because two different fluids are involved, one wind farm with mineral oil and four with synthetic ester. Remember that the Duval pentagon is specific to mineral oils and, as far as we know, the corresponding pentagon for synthetic esters has not yet been published.
An attempt has been made to use the Duval pentagon with the transformers of the Spain 1, 2 and 3 and Turkey wind farms (Fig 2a)-2d)). We are aware that this diagnostic method has not yet been tested on fluids other than mineral oils, but we believe that synthetic esters, in a similar way than mineral oils, should be able to withstand the stresses of the transformers. Therefore, if a fluid, even if it is not a mineral oil, generates large quantities of gases, this is a symptom that something is wrong inside or around the transformer that contains such fluid.

For the transformers in Peru, with mineral oil, the Duval pentagon indicates stray gassing in all cases except for the two transformers with the highest hydrogen content (9886 and 394 ppm respectively), which indicates partial discharges. Fig. 2e) shows the Duval pentagon for the two transformers whose gas concentrations exceed the values considered normal (the rest are not shown because their gas concentrations are normal).
In summary, in Spain 1 windfarm there is one transformer with stray gassing and nine with high-temperature defects (T3).
In Spain 2 windfarm there is one transformer with a stray gassing defect, one with a low-temperature defect and eight with high-temperature defects.
In Spain 3 windfarm there are two transformers with partial discharges, one with stray gassing, one with a low-temperature thermal defect and six with a high-temperature thermal defect.
In Turkey, 17 out of 33 transformers show stray gassing, with no electrical or thermal defects.
Discussion
Despite the enormous interest in wind energy, relatively few publications on gases generation have been published.
Excessive gassing in wind farm transformers has been observed for quite some time. L. J. Parthemore and R. T. Rason (6) published at the 78th Annual International Doble Client Conference, in 2011, data of more than 6.000 individual mineral oil-filled cabinet-style transformers installed in wind turbine collection systems. This study was extended in 2013 with data of another 11.000 samples (7). These studies have demonstrated that, when applying the IEEE criteria for power transformers, there is a significant number of wind farm transformers exceeding the normal gas concentrations, with a tendency to increase its percentage from one year to the next, indicating that gas concentrations are continuously increasing (7).
Singh and Blackburn compared dissolved gas analyses in WTTs with those of standard oil-filled distribution transformers in Australian wind farms (8). Fernandez discusses gas formation in collector windfarm transformers; therefore, their values are much lower than ours (9).
We do not know the actual reasons why these transformers are generating such enormous amounts of gases. They could be related to inadequate design of the transformers or to the stresses to which this activity subjects them. It should be noted that pad mount transformers have originally been used in underground power systems. Its construction allows installation in locations accessible to the general public without the need for protective fencing or vaults. Therefore, they have not been specifically designed for use in wind generation.
On the other hand, international standards do not help:
DGA results from transformers identified as wind turbine transformers often had excessively elevated typical values … There was no information to indicate if the transformers were healthy or under extreme distress (2)
For this reason, the values for hydrogen can be considered “typical” for statistical purposes, but they could not be inherently attributable to fault absence (3).
One thing that is observed is that there are apparently two default modes. One, which mainly affects pad mount transformers, but is not exclusive to them, is a low-temperature defect (partial discharge, stray gassing or low-temperature thermal defect T1). The other, which affects a larger proportion of distribution transformers, is a high-temperature thermal defect T3.
In transformers other than wind turbine transformers, stray gassing is generally associated with unit cooling difficulties, which does not seem to be the case for WTTs. Type T3 defects are associated with defects located in the core.
Most likely reason for gas generation
The causes of this gas generation are not yet well understood. In a comprehensive study (10), G. Jose and R. Chacko discussed the most likely reasons of wind farm turbine transformer’s failures. They identified and discussed the following potential problems:
Harmonics
Regulation and control equipment attached to asynchronous generators with double winding rotor may produce voltage and current harmonics, specifically their electronic convertors are those that, with low turbine load, in many occasions exhibit an inadmissible THD (Total Harmonic Distortion) index. There is an electrotechnical relationship, referred to a specific point of the installations, called “short circuit ratio”, which is the quotient between the existing short-circuit power at said point of the installation and the power that currently exists or is passing through said point. This ratio, in reference to the MV Busbars in a wind farm, which is the point of delivery of energy from the wind farm to the MV/AT step-up transformer, must have a value greater than 5, since otherwise, the THD would be high or unacceptable. In other words, should in a receiving substation of a windfarm with a MV/HV step-up transformer, the MV had a short-circuit power of 450 MVA, and the windfarm generates a power of 50 MVA at that moment, the short circuit ratio would be:
Short circuit ratio = √(450 MVA/50 MVA) = √9 = 3 <5
therefore, the converter would be generating an unacceptable number of harmonics.
Variable loading cycle
By definition, wind turbines produce electricity when the wind blows, but it should be noted that the wind speed must be within a certain range. If there is no wind or the wind speed is too low, the turbines do not generate electricity. If the wind speed is too high, the turbines must be stopped to avoid damage. In addition, the wind speed changes continuously, causing stresses on the shaft, windings and other internal components of the wind turbine. These thermal cycles can lead to gas bubbles, partial discharges or hot spots that can cause damage to insulation or electrical connections (10).
Low voltage fault ride through
In order to maintain grid stability, the wind farm turbines must remain connected in the event of a fault. This is called “low voltage fault ride through” and represents higher electrical and thermal stress on the wind turbine transformers than in transformers connected to conventional power sources (10).
Switching surges and transient overvoltages
Since the wind speed is not constant and fluctuates continuously, switching operations occur several times a day. This generates transient overvoltages due to the current cut-off that can cause failures in the transformer insulation decreasing its lifetime (10).
Core breakage or delamination
For the specific case of pad mount transformers, it has also been proposed core breakage or delamination as the root of gas production and transformer failure (11).
High-temperature thermal defect
In the case of distribution transformers where high temperatures thermal defects have been observed, CIGRE has proposed the presence of “supra-harmonics” which can increase the hot spot temperatures above the design limits accelerating insulation aging (12).
Interaction with the electrical system
Windfarm design, that is, the connection point where the energy generated by the wind farm is injected into the grid is also very important because, in some cases, the waveform generated by the asynchronous wind generators could distort the voltage waveform at that connection point producing numerous harmonics.
Conclusions
The WTTs generate abnormally high amounts of dissolved gases. However, there are important differences between transformers of different types (pad mount or distribution) and, even, between transformers of the same type and the same wind farm. Unfortunately, current diagnostic tools for transformers are not sufficient to establish a specific caseload.
DGA results suggest that main failure modes in pad mount transformers are either partial discharge, stray gassing and low-temperature thermal. It has been suggested a core breakage or delamination as one possible cause.
The main failure mode for distribution transformers filled with a synthetic ester is high-temperature thermal. It has been suggested that harmonics that increase the hot spot temperature is a possible cause of failure.
1 An extended version of this paper was first presented at AWRtr22 in Bayona (Spain), October 2022. It is also presented here for a wider audience.
References
- Global Wind Report 2021, https://gwec.net/global-wind-report-2021
- “IEEE Guide for the Interpretation of Gases Generated in Mineral Oil-Immersed Transformers“, IEEE Standard C57.104-2019.
- IEC, “Mineral oil-filled electrical equipment in service ʹ Guidance on the interpretation of dissolved and free gases analysis. IEC 60599:2022“.
- ASTM, “Mineral Insulating Oil Used in Electrical Apparatus“, Standard D3487 2016
- M. Duval and L. Lamarre, The Duval Pentagon – A New Complementary Tool for the Interpretation of Dissolved Gas Analysis in Transformers, “IEEE Electrical Insulation Magazine“, Vol 30, No. 6, pp. 9-12, September/October 2014.
- L. K. Parthemore and R. T. Rasor, Gassing in Wind Farm Transformers, “78th Annual International Conference of Doble Clients“, Paper T-02, 2011.
- M. Dickinson, C. Collins and R, T. Rasor, Gassing in Wind Farm Transformers, “80th Annual International Conference of Doble Clients“, Paper T-08, 2013.
- P. Singh and T. Blackburn, Dissolved Gas Analysis Results in Wind Turbine Transformers, “2018 Australasian Universities Power Engineering Conference (AUPEC)“.
- J. Fernández Daher, “Dissolved Gas Analysis, Recent Changes in Standards. Application to Wind Farm Collector Transformers, 2021” IEEE URUCON.
- G. Jose and R. Chazco, A Review on Wind Turbine Transformers, “International Conference on Magnetics, Machines & Drivers” (AICERA-2014 iCMMD).
- P. J. Hopkinson, Wind Power Transformer Design, Doble Engineering Conference 021913, http://docplayer.net/36234055-Wind-power-transformer-design-by-philip-j-hopkinson-pe.html
- 12. C, Krause, A. de Pablo, F. Devaux, H. Ding, V. Katshuna, J. Lukic, L. Melzer, S. Miyakazi, M. Munro, A. Peisoto, M. Scala and D. Walker, “The Condition of Solid Transformer Insulation at End-of-Life“, Electra 321, 36-47 (2022)