In June 2005, ASTM International adopted a new test method for testing transformer oil. The aim of the test is to determine the gassing pattern of an oil subjected to thermal stress under what is considered to be low temperatures, i.e. 120°C. The method is entitled "ASTM D7150, Standard Test Method for the Determination of Gassing Characteristics of Insulating Liquids Under Thermal Stress at Low Temperature".
Interest in the development of this test resulted from research that Doble performed in the early 1990s and more recently the research that the CIGRE Task Force TF 15/12-01-11 (TF11) performed in trying to discern the differences in gassing behavior of some transformer oils at low temperatures. CIGRE has termed the unusual gassing behavior as "stray gassing".
The CIGRE research was mainly looking at the gassing behavior of new oils in which transformers were exhibiting increasing hydrogen levels, yet no apparent cause for the abnormal gassing could be determined. As a result, the research centered on the oil and found that some oils produced more gas (in this case hydrogen) than others when they were aged under thermal conditions in both sealed transformers (gas blanketed, sealed or bladder/diaphragm transformers) and free breathing transformers (open conservators). Doble had also observed a similar pattern and as a result conducted a massive research study into the phenomenon that has lasted over three years and still goes on today. What was also determined was the fact that not only could the refining process be partially responsible for the abnormal production of dissolved gases, but different types of contamination in the oil, incompatible materials in the transformer, and the addition of additives such as metal passivators could also cause abnormal gassing to occur.
The reason that this information is so important is that it can significantly affect the results of the dissolved gas-in-oil (DGA) test. DGA is one of the most widely used diagnostic tools for assessing the condition of electrical transformers and in more recent years load tap-changers and bulk oil circuit breakers. The test is very sensitive to a wide range of problems and is used to characterize incipient fault conditions. Categories of abnormal conditions such as overheating, partial discharge, and arcing can be further identified by the insulating materials involved (paper or oil) or by the energy (high, moderate, or low temperatures). DGA is ideally suited for condition-based maintenance programs as it allows early detection of problems and can then be used to follow most deterioration processes as they evolve. Then, the decision can be made as to when to perform more definitive tests to identify the specific problem or take remedial action. Indeed, DGA is sensitive enough that it can be used to detect problems during factory testing of new transformers helping manufacturers avoid shipping defective units to the field.
It is imperative that DGA results are interpreted accurately to ensure the correct action is taken concerning any incipient faults within the transformer. It is well known that oils can age differently, particularly where there is ample oxygen such as may be found in older units, or free-breathing transformers. A 1991 study revealed that oils exposed to relatively low temperatures could have different gassing behaviors and oxygen consumption rates under the same controlled conditions. Gassing rates were different for two oils tested if the oxygen content was fairly low at 3000 ppm or high at 30,000 ppm and this is one of the reasons why the new ASTM D7150 test is performed using air and nitrogen sparged samples. It was also shown that passing the dissolved gas-in-oil test for a factory heat run could depend on the oil chosen.
Early Studies on Gassing Behavior
In 1994, the Doble Oil Committee performed a study on the low temperature gassing behavior of an oil manufactured in France that was being used for the first time by a French transformer manufacturer. The transformer did not pass the dissolved gas-in-oil limits used by the purchaser. The manufacturer performed some tests which showed that the oil used (Shell Diala F) gassed much more than the oils they used historically. To evaluate the gassing behavior of the Shell Diala F oil in comparison to other products, air-saturated samples and some samples that had been degassed (all gases removed except for small amounts of oxygen and nitrogen) were aged in ground-glass matched barrel and plunger syringes at 95ºC for 168 hours. The following oils were tested in the study:
(1) Shell Diala F, uninhibited, manufactured in France
(2) Esso Univolt 52, uninhibited, manufactured in France
(3) Exxon Univolt N61, inhibited, manufactured in US
(4) Shell Diala A, uninhibited, manufactured in US
The results are provided in Table 1 (air saturated) and Table 2 after the oil had been vacuum processed.
The results listed in Tables 1 and 2 show clearly that the Shell Diala F product produced larger amounts of combustible gases than the other oils tested. Even when oxygen is removed or reduced in the test samples, the overall concentration of gases is reduced but the same oil still produces the largest quantity of dissolved gases. Table 2 reveals that degassing the two oils to reduce the oxygen content had a much greater effect on the Univolt 52 oil than the Diala F product. This type of experiment shows that the gassing characteristics of new oils under thermal stress can be drastically different. In the case of a transformer heat run, one oil would cause the transformer to fail the test while the other oil would not be of concern.
Recent Studies on Gassing Behavior
Doble, with the inspiration of the CIGRE work and its own previous work, undertook a study to evaluate the gassing characteristics of oils commercially available in 2001 and 2002 and in many cases still sold today. The study was to look at a wide range of oils to determine the amount of variation.
Testing was performed under the following conditions:
(1) Air purged, 16 hours at 120°C
(2) Air purged, 164 hours at 120°C
(3) Nitrogen Purged, 16 hours at 120°C
(4) Nitrogen Purged, 164 hours at 120°C
The aging was performed for differing times to assess initial gassing rate and when a plateau or equilibrium rate was reached. The samples were either air or nitrogen purged to have a range of oxygen contents representative of service conditions. The aging times were as follows:
(1) 16 hours - provides indication of initial generation of gases and would be similar to the time for a factory heat run test.
(2) 164 hours - indicates if the gases reach a plateau or a constant rate of generation
The oils were assigned a sequential number based on the total amount of combustible gases formed during the experiment with air saturated oils aged for 16 hours. For example, number 1 had the least amount of total combustible gas (TCG) and number 30 the greatest amount of TCG. In some cases there was more than one product from a refiner. Only the results of the 164-hour testing are presented here as this is the time that was decided upon for the ASTM method.
Air Saturated Samples Aged for 164 Hours at 120°C For the samples saturated with air and aged for 164 hours, the data shown in Figure 1 reveals that there is a large difference in the total combustible gas (TCG) content between some of the oils and that very significant amounts of combustible gases are formed in some cases. There were four gassing patterns that developed:
(1) Mostly carbon monoxide
(2) Hydrogen and carbon monoxide
(3) Mostly hydrogen
(4) Mixture of hydrogen, carbon monoxide, methane and ethane
The predominant gas formed was hydrogen. However, other gases such as methane, ethane, ethylene, carbon monoxide, and carbon dioxide were also formed. In general, if the oxygen content remained high, the first three gassing patterns occurred (as listed above). If the oxygen was depleted, the methane and ethane were generated in greater concentrations. Ethane was not generated in significant amounts without methane. The oxygen consumption rate was quite different for the various products.
Nitrogen Purged Samples Aged 164 Hours at 120°C
Samples purged with nitrogen and then aged 164 hours did not contain as much combustible gas as those saturated with air before aging for the same amount of time; however, there were five oils that still exhibited high concentrations of hydrogen and total combustible gases. In these cases the hydrogen made up a large percentage of the composition of the TCG.
What is interesting in nitrogen-sparged samples is that the methane and ethane values are consistently low but make up a greater percentage of the TCG for more samples than was the case when there was ample oxygen at the start of aging. This is due to the fact that the hydrogen produced in nitrogen-sparged samples is not as high in concentration. Similarly, the carbon oxide gases are consistently lower for the nitrogen-purged samples than when saturated with air.
Conclusions
(1) There is a new ASTM Method D7150 that provides the protocol for determining the gassing characteristics of oils under air and nitrogen (sealed) conditions and at low temperature (120°C).
(2) It is not meant to be a routine test method in the sense that DGA currently is. However, it can be applied in those situations where there is a question of whether the oil is involved in producing the gassing that is being observed.
(3) DGA is a very important test, but care needs to be taken in the interpretation of results to understand the low temperature gassing behavior of the oil when certain types of problems are detected. As hydrogen is often the predominant gas formed for the oils with the highest combustible gas generation rates, it is possible that low temperature thermal problems could be confused with low energy partial discharge activity. Stray gassing may also be mistaken for excessive gassing at modest temperatures.
(4) Some of the "breaking in" characteristics that are observed for the gassing behavior of new transformers or newly processed transformers may be attributed to the oil and will stabilize over time.
(5) Stray gassing can also be due to the presence of contamination in the oil, the presence of incompatible materials, and the presence of additives.
(6) It is clearly important to know the properties and characteristics of the oil in individual transformers, especially when performing DGA. Similarly, it is important to make sure the gassing characteristics of oils used to "top off" transformers are also well understood.