Dissolved Gas Analysis Of Transformer Oil


Like other equipment, power transformers also require periodic checks for its smooth and trouble free functioning. Preventive inspections must be made at predetermined frequency and corrective measures must be taken for optimum performance of the transformer.

Now the days are gone when the power transformer was considered as requiring minimum preventive measures, because it did not have any moving parts. Reliable power flow is considered necessary for all power generating stations. This makes the power transformers as one of the most important assets. In addition to this, the power transformers at most of the locations have achieved their working lives. Hence, it does not need to over emphasis that transformer condition monitoring and failure analysis becomes a high priority. 

Faults inside a transformer like arcing, sparking, partial discharge and overheating result in evolution of gases which are dissolved in transformer oil. These gases in transformer oil can be analyzed using advanced techniques of dissolved gas analysis (DGA) for determining the type and severity of fault already occurred or a pending fault. 

DGA is an excellent and accurate method for measuring these gases and determining the internal health of a transformer. The use of a DGA diagnosis tool can provide better equipment availability, prevention of transformer failure and capital expenditures. Purpose of this article is to discuss existing DGA techniques and assessment of the results. 

Oil checks on a Power Transformer

Following are oil checks on a power transformer

  • Break down voltage (BDV)
  • Moisture content
  • Resistivity
  • Dissipation factor
  • Acidity
  • Dissolved gas analysis

Functions Of Transformer Oil

  • It contributes to the dielectric strength of transformer insulation system
  • Provides cooling to transformer winding and core.
  • The paper insulation on winding melts down due to heat and aging and settles down at the bottom of the tank. The transformer oil prepares sludge by reacting with these insulation parts. This sludge can be filtered out by regeneration method.
  • It acts as a protective layer and prevents oxidation of metal surfaces and paper insulation.

When the parameters of transformer oil are not within acceptable range of values, the oil is called as degraded. The further use of this oil reduces the life of the transformer. Hence, conditioning monitoring of the oil is essential to ascertain the good health of the oil. Table 1 summarizes acceptable limits of transformer oil parameters as below.

Table 1

Dissolved Gas Analysis (DGA) Of Transformer Oil

The transformer oil in operation is subjected to thermal as well as electrical stresses. As these stresses accumulate, the insulating materials may breakdown or decompose resulting in evolution of several different gases from hydrocarbon mineral oil and paper insulation. Particles of solid insulation present in the oil also result in formation of gases.  A qualitative and quantitative assessment of these gases helps in diagnosis of internal faults in a transformer. 

Gases in oil filled transformers may also be formed due to aging of oil & paper insulation, oxidation and electrolytic action.

In case of an internal fault, assessment of composition of gases and the rate of their formation is often helpful in diagnosis of type and severity of fault. In case of incipient faults, the gases formed due to decomposition of oil have enough time to dissolve in the oil. 

It is recommended to conduct the DGA test of transformer oil at periodic intervals. It helps to maintain a DGA history card of transformer oil. This data is helpful in providing information about occurrence of incipient faults, pending faults and the internal health of a transformer over its lifetime. In a DGA test, the gases in oil are analyzed to determine the quantity of gasses in a particular sample of oil. Hence, DGA is helpful in commencing predictive or preventive maintenance action before the occurrence of a fault or we can say a pending fault. 

Relationship Between Temperature And Formation Of Gases

Some quantity of gas formation is expected during normal operation due to aging of the transformer. Hence, it is important to separate normal gas formation and abnormal gas formation. Quantity of normal gas formation varies with transformer loading and the type of insulation material used. Abnormal formation of gases is used universally for all transformers to analyze abnormal conditions.

Typical gases that are formed in power transformation are hydrogen (H2), methane (CH4), ethane (C2H6), ethylene (C2H4), acetylene (C2H2), methyl ethylene (C3H6), Propane (C3H8), carbon dioxide (CO2), carbon monoxide (CO) and oxygen (O2). Each of these gases begin to form at different values of temperature and dissolve inside the transformer oil. The types and quantity of these gases formed during fault conditions depend upon type and severity of fault. 

Hydrogen and methane commence to evolve in small quantities at 1500C. With increase in temperature, formation of hydrogen continues to increase. At about 2500C, the formation of ethane begins. The formation of ethylene starts at around 3500C. Generation of methane, ethane and ethylene reaches a highest value and then goes down with further increase in temperature.

As per latest studies, formation of trace amounts (a few ppm) of acetylene can be found at 5000C due to overheating fault. Larger amount is found at temperature beyond 7000C by arcing fault. 

Thermal decomposition of winding paper (cellulose) insulation takes place at about 1000C. This process generates oxygen (O2), methane (CH4), hydrogen (H2), carbon dioxide (CO2) and carbon monoxide (CO). It is therefore recommended to operate the transformer at 900C or less.

There are several techniques for fault diagnosis by DGA, like characteristic or key gas analysis method (KGM), Rogers gas ratios method (RRM), Doernenburg Ratio Method (DRM) and Duval Triangle Method (DTM). Out of these methods Duval Triangle Method (DTM) is the most popular and accurate one. Following part describes the characteristic or key gas analysis method (KGM) and Duval Triangle Method (DTM). 

Characteristic Or Key Gas Analysis Method (KGM)

The Key Gas Method is based on the philosophy that on occurrence of a particular type of fault, decomposition of oil and/or solid insulation takes place in addition to increase in temperature of the transformer. This process results in evolution of one or two key gases in major quantities in addition to formation of other gases in smaller quantities. The table – 2 and 3 relates the type of fault with key gas that is mainly formed. Total released gases inside the transformer is given in table – 4.

Table 2

Table 3

Fig – 1 indicates the key gases and their relative proportions to indicate the four general types of transformer faults.

Total released gases inside the transformer are given in following table

Table 4

Characteristic or key gas analysis method provides results for all the fault diagnoses to be done by DGA. But the percentage of correct results is only 42 %. This problem is resolved by Duval Triangle Method (DTM). By this method we get results for all the fault diagnoses to be done by DGA. But the percentage of correct results is as high as 96 %. Hence, we can say that the Duval Triangle Method provides more accurate and correct results compared to any other method available at present time. This method is explained as below.

Duval Triangle Method (DTM)

The Duval triangle was developed using about 200 numbers of transformers diagnosed for one or more number of faults. The triangle is divided in 6 zones. Each zone defines a particular type of fault. The faults indicated by these zones are partial discharges (PD), thermal fault less than 3000C (T1), thermal fault between 3000C and 7000C (T2), thermal fault greater than 7000C (T3), low energy discharge (sparking, D1), high energy discharge (arcing, D2), mix of thermal and electrical faults (DT).

The DTM is based on three key gases, namely CH4, C2H4 and C2H2. These gases are in sequence of increasing energy levels as shown in Fig – 2. The % concentration of the three gases form the three sides of the triangle. 

C:\Users\hp\Desktop\dovel triangle.PNG
Fig – 2 The Duval Triangle

This method is applicable only when it has been ascertained that a fault exists in the transformer, because there is no zone of “no fault” inside the triangle.

Worked Example Of Fault Diagnosis Using Duval Triangle

There will be a sudden increase in concentration of gases when a fault begins. Hence, we shall take two samples for DGA. One before the fault, say DGA 1 and second after the fault begins, say DGA 2. The concentration of the three gases obtained by DGA is illustrated in Table – 5.

Table 5

Using above concentrations two points are plotted inside the triangle i.e. point 1 and 2. 

Steps To Obtain Point 1

Point 1 is obtained using total concentration of three gases by           DGA 2 as shown in Fig – 2. The concentration of these gases in ppm is given by CH4, C2H4 and  C2H2. The concentration of these gases in % is calculated as  below.

% CH4 = { CH4/( CH4 + C2H4 + C2H2)} * 100 ------------------Eq. 1

%  C2H4 = { C2H4/( CH4 + C2H4 + C2H2)} * 100-----------------Eq. 2

%  C2H2 = { C2H2 /( CH4 + C2H4 + C2H2)} * 100----------------Eq. 3

These % concentrations are plotted along the three sides of the Duval triangle starting at % concentration of the gases obtained by Eq. 1 to 3. The three lines intersect at a point. This is point 1. The location of this point in the triangle indicates the type of fault defined by that zone. Following calculations are done using values given in table – 5 and  Eq. 1 to 3.

Total concentration of gases ( CH4 + C2H4 + C2H2) by DGA2 = 369 ppm

% CH4 = 192*100/369 = 52 %

%  C2H4 = 170*100/369 = 46 %

%  C2H2 = 7*100/369 = 2 %

Drawing three lines as aforesaid equivalent to % concentrations of gases, we obtain point 1. Point 1 is located in the T2 zone, which indicates a thermal fault between 3000C and 7000C.

Fig – 3

Steps To Obtain Point 2

Point 2 was obtained using % increase in individual concentration of the three gases by DGA 1 and DGA 2 as shown Fig – 3. Say, the increase in concentration of these gases in ppm is given by CH4, C2H4 and C2H2. Say, the total increase in ppm concentration of these gases is A. The increase in concentration of these gases in % is calculated as given below.

Increase in % CH4 = (CH4/A) * 100 ------------------Eq. 4

Increase in % C2H4 = (C2H4/A) * 100----------------Eq. 5

Increase in % C2H2 = (C2H2/A) * 100----------------Eq. 6

These % increases in concentrations are plotted along the three sides of the Duval triangle starting at an increase in % concentration of gases obtained by Eq. 4 to 6. The three lines intersect at a point. This is         point 2. The location of this point in the triangle indicates the type of fault defined by that zone. Following calculations are done using values given in table – 5 and Eq. 4 to 6.

Total increase in gas concentration = 139 ppm

Increase in % CH4 = 50/139 = 36 %

Increase in % C2H4 = 86/139 = 46 %

Increase in % C2H2 = 3/139 = 2 %

Drawing three lines as aforesaid equivalent to % increase in concentrations of gases, we obtain point 2. Point 2 is located in T3 zone, which indicates a thermal fault greater than 7000C. It must be noted that in both cases % of C2H2 is 2 %, hence both lines are the same.

Many times both the points are located in the same zone. Still, it is recommended to locate both the points for accurate and reliable diagnosis. Here we note that fault at point 2 is more severe than fault at point 1. For all practical diagnostic purposes, if there is an increase in % of gas concentrations, we have to consider more severe faults. Hence, in this case, the actual fault is at point 2 i.e. a thermal fault greater than 7000C. 

From the above diagnosis, it can be concluded that the fault is probably a bad connection on a bushing bottom, a bad contact in the tap changer or a problem with a core ground. These are the areas where the fault will not degrade the paper (cellulose) insulation.

One drawback of ratio methods is that for some instances, no result is obtained (unsolved diagnostic). This is not the case with DTM, because this is a closed system and we obtain results for each diagnosis.

 Applicability Of DTM

  • This is applicable even when two or more numbers of faults are existing simultaneously.
  • This is useful in cases when fault type can not be diagnosed by any other technique of DGA.
  • A rapidly developing fault can always be detected. 

Case Studies Related To Transformer Faults

Case Study – 1 

In this case study, a transformer oil arcing fault was diagnosed on the basis of presence of C2H2 in oil. In view of this, an internal inspection of the transformer was carried out. But no fault was found. This can be explained as below.

In case OLTC is an integral part of the transformer, the gases generated in OLTC are transferred to the main tank. During operation of OLTC, C2H2 is generated in OLTC tank and transferred to the main tank. This resulted in misinterpretation of the existence of arcing fault. 

The degassing of the transformer was done and it was taken back into service.

Case Study – 2 

This case study is also related to arcing fault in the transformer. DGA was carried out. This indicated the presence of C2H2 and CO. Presence of these gases is an indication of arcing fault in paper insulation. Inspection of the transformer reported the development of an arc between HV winding and the tank. Transformer was repaired. After filtration and degassing and ascertaining all parameters being within acceptable limits, the transformer was taken into service.

Acceptable concentration of dissolved gases of a typical healthy transformer for different periods of service are given in table – 6 

Table 6

Acceptable concentration of dissolved gases of a typical healthy transformer as per Russian standard are given in table – 7

Table 7

Acceptable concentration of dissolved gases of a typical healthy transformer as per IEEE recommendations are given in table – 8

Table 8


  • Above Acceptable concentration of dissolved gases is applicable when no oil filtration is done for degassing of oil. It is recommended to carry out oil filtration before and after DGA to create historical data for future analysis.
  • It is recommended to degas the oil at regular intervals as per table – 4. It is due to the fact that gassing of oil results in deterioration of dielectric properties of oil and lowering of flash point from 145 deg. cent. to 50 – 80 deg. cent. 
  • Recommendation for Acceptable concentration of Oxygen

As per the paper of W Lampe and E Spicer, Sweden, 1976, the concentration of O2 in oil should not be more than 200 ppm. The advantages of having Oxygen concentration less than 200 ppm are given as below.

  • Aging of oil is reduced considerably
  • Aging of paper insulation is reduced by a factor 5

Levels at which dissolved gases are significant

For significant presence of gas, concentration should be at least 10 times of sensitivity limit.

The sensitivity limits for gases dissolved in oil are given as below  

  • Hydrogen – 5 ppm
  • Hydrocarbons – 1 ppm
  • Carbon dioxide – 25 ppm
  • Atmospheric gases – 50 ppm

Guidelines For Determination Of Action Based On DGA

  • If C2H2 evolution rate is 10 ppm per day or more, internal inspection of the transformer is required.
  • If evolution of total combustible gasses is 100 ppm per day or more, internal inspection of the transformer is required.

Labeling Of Oil Samples

Oil samples should be labelled properly before sending it to the laboratory along with following information.

  1. Transformer tag no.
  2. Object of test – Transformer tripping or periodic test
  3. Date
  4. Load at the time of sampling
  5. Type of oil
  6. Volume of oil in transformer
  7. Oil temperature
  8. Date of last filtration of oil
  9. Tap changer – Integral/Isolated
  10. Date of last DGA
  11. Point of sampling


  1. In case of any abnormal trend of DGA, take two samples of oil from same point at the same point for three consecutive days.
  2. Volume of oil in a transformer is important for DGA. If the fault energy in two transformers is the same, same amount of gasses shall be generated in the transformer. But the amount (ppm) of individual gasses shall be inversely proportional to transformer oil volume.


In this article, we discussed the prevailing DGA diagnostic tools. These tools are capable of providing information on transformer faults which have already occurred or pending for occurrence. This information is useful to determine suitable decisions and future course of action. In some instances, faults may not be so serious or severe requiring shutdown of the transformer.

If the process of DGA depends on the laboratory then the laboratory decides the diagnosis results. But, if the end user is interested, necessary   understanding, skills and expertise may be developed for fault diagnosis.

The objective of the DGA is to diagnose the fault accurately. There are several methods available and we have to select the most accurate and suitable method. It is recapitulated that DTM has 96 % accuracy value and is capable of providing results for all diagnoses by DGA.

Techniques discussed so far, somewhat depend on technical experience of experts. Researchers are inclined to develop alternative ways to use DGA data more efficiently with highest accuracy and diversified manners. These alternative approaches involve artificial intelligence (AI) which includes Fuzzy logic and neural network techniques.

Recent development of AI models based on a combination of KGM, DRM, RRM and DTM techniques provides future vision. AI enhances the performance of conventional techniques. However, sometimes AI technique fails to distinguish between thermal faults in oil and the same faults in paper insulation, so expert opinion is still required.

Future of DGA as a fault diagnostic tool involves AI to enhance speed, accuracy, reliability and performance of these techniques.

August 13, 2020
Anupam Rastogi
Get great insight from our expert team.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
By signing up you agree to our Terms & Conditions