Dissolved Gas Analysis of Overheating Fault in Bio-degradable Transformer Oil

N.A. Muhamad, B.T. Phung and T.R. Blackburn School of Electrical Engineering and Telecommunications University of New South Wales Sydney, Australia
Abstract- This paper presents results of experiments performed in the laboratory to simulate overheating fault on biodegradable transformer insulation oil. Dissolved hydrocarbon gas products generated by such a fault in three different oil conditions (normal, dry, wet) and with or without presence of pressboard are analysed. For comparison, the tests are also performed on conventional mineral oil. Results obtained indicate that existing DGA techniques which were developed for monitoring overheating faults in mineral oil need to be modified when used for insulation assessment of bio-degradable oil.

investigation. TABLE 1 lists the relevant properties of the two oils. The objective of drying the oil is to remove moisture which is absorbed during atmospheric exposure. To achieve dry oil condition, the oils were circulated in an oil dry-out plant for at least 48 hours to remove moisture in the oil before being used in the experiments. The highest measured value of moisture content after drying for both oils is 25ppm.
TABLE 1. STANDARD APPLIED AND PROPERTIES OF THE TESTED OILS

I.

INTRODUCTION

Overheating is one of the common faults happened to oil immersed power transformers. Dissolved gas in oil analysis (DGA) is one of the techniques that are widely used to diagnose transformer conditions. During faults, transformer insulation oil will release hydrocarbon gases that are dissolved into the oil. The types and quantities of hydrocarbon gases produced give an early indication of inception faults in transformer. There are five common types of hydrocarbon gases dissolved in transformer oil. They are H2 (hydrogen), CH4 (methane), C2H6 (Ethane), C2H4 (ethylene) and C2H2 (acetylene). Common faults that can be detected using the DGA method are partial discharge, overheating, and arcing or breakdown faults. Overheating fault can be detected from the presence of hydrogen together with methane, ethane and ethylene. The key gas for this type of fault is ethylene gas [1]. The guideline for interpretation of these dissolved hydrocarbon gases such as IEEE C57.104.1991 and IEC60567 standards have been established based on conventional transformer insulation oil (mineral oil). In this work, hydrocarbon gases produced by biodegradable transformer insulation oil are investigated. The results of these investigations are presented in this paper and analysed to determine the possibility of applying existing DGA interpretation rules to the case of overheating faults in bio-degradable oil insulated transformers. II. EXPERIMENT SETUP AND PROCEDURES A. Tests Oils A bio-degradable oil and a mineral oil, commercially available and widely used in industries, were tested in this

Moisture was added into the dried oil using measured quantities of water for specified quantities of oil to achieve wet oil condition. In this investigation to add 200ppm moisture into the oils, a 2 liter beaker was used to hold the dried oil and a 3ml syringe was used to inject distilled water into the dried oil. A magnetic stirrer with hotplate was used for mixing water in the oil. The temperature was set to 50C to avoid vaporization of water and the oil was stirred for 24 hours before being used for the test (Figure 1)

2008 Australasian Universities Power Engineering Conference (AUPEC'08)

Paper P-179 page 1

silicone seal. These glass bottles are also capable of being sealed gas tight. Argon gas was added to fill in the empty space inside the bottle. D. Dissolved Gas Analysis (DGA) and Interpretation. Oil samples were sent to a chemical laboratory of a local power utility for gas extraction and moisture. The Duval Triangle method has been found to have better performance compared to others DGA interpretation methods, including those enhanced with the use of artificial intelligent systems [2, 3]. In this paper, the L1 values for the Duval triangle method are chosen as the reference limit[4]. This method and Roger Ratio method are used to interpret the DGA result.

III. RESULTS
Figure 1. Adding moisture into oil process TABLE 2 MOISTURE LEVEL OF TESTI OILS

Oil Conditions

Moisture Level (ppm) Mineral Oil Biodegradable oil 14 31 40 22 170 380

B. Drying and Oil Absorption of Pressboard The procedure follows the IEC60641-2 Standard. The pressboard pieces were placed in the oven at a temperature of 105C for 48 hours. The test oil was placed in an aluminium tray and heated to 90C. The dried pressboards were then transferred to the test oil and left submerged. The pressboards were left in the oil for 6 hours without heating before testing (Figure 2).

Dried Normal Added 200ppm Moisture

Moisture results of the oil samples before testing are shown in Table 2. Mineral oil was found very hard to add or absorb moisture as referred to Table 2, moisture level for added 200ppm moisture and normal oil conditions of mineral oil is not much different and not as high as biodegradable oil. Fig. 2 shows the color changing of the tests oil at after and before overheating test. Mineral oil changed color from pale yellow to golden yellow while for biodegradable oil, there not too much color different between oil before overheating and after overheating. The DGA results obtained from the heating test are summarized in Fig 4 to Fig 7. In general biodegradable oil and mineral oil produced almost same types of fault gases which are hydrogen, methane ethylene and ethane gases for testing with/without pressboard. However, mineral and biodegradable oil produced different kind of key gas during the tests. Ethane gas was found to be the main key gas for biodegradable oil in all oil conditions for tests with or without presence of pressboard. Meanwhile, methane was found to be the key gas in the mineral oil for both tests in all oil conditions.

Figure 2. Pressboards immersed in the insulation oil

C. Overheating Test The overheating test was performed by heating the oils in the oven at temperature maintained between 130oC and 140oC. Special bottles with a temperature resistant screw-cap and matching pouring ring were used. This bottle can be used in temperature up to 260C and is also fitted with a PFTE-coated

2008 Australasian Universities Power Engineering Conference (AUPEC'08)

Paper P-179 page 2

(a)

Figure 5: Faults gases dissolved in mineral oil for the test without pressboard.

(b)

(c)
Figure 3: Oil color after heated more than 96 hours (left: biodegradable oil, right: mineral oil) (a) before testing (b) after overheated without pressboard (c) after overheated with pressboard

Figure 6: Faults gases dissolved in biodegradable oil for the test with pressboard.

Figure 7: Faults gases dissolved in mineral oil for the test with pressboard. Figure 4: Faults gases dissolved in biodegradable oil for the test without pressboard.

2008 Australasian Universities Power Engineering Conference (AUPEC'08)

Paper P-179 page 3

TABLE 3 FAULTS INTERPRETATION

II. Results for tests on biodegradable oil without pressboard (Fig. 4) show that this oil generates only ethane gas that exceeds the L1 value for all oil conditions. The drier the biodegradable oil the higher the amount of ethane will be dissolved in oil when overheated. From test results for mineral oil (Fig. 5), it can be seen that the higher the moisture level, the more gases will exceed L1 level. In the dried oil condition, no faults gases produced exceed L1 level and as the moisture increased in normal oil condition, methane gas was detected to exceed L1 limit. When more moisture was added to the oil, ethane and methane value exceed the normal limit. The pattern of dissolved gases in biodegradable oil for the test with pressboard (Fig.6) is the same as the test without pressboard for all oil conditions. Only the ethane amount is above the normal limit. However, increasing amount of methane gas was detected in dried and normal oil condition. Nevertheless, these values are still well below the L1 limit. Results for test on mineral oil with pressboard (Fig.7) are also almost the same as the test without pressboard. The difference is that, in this test, the methane amount dissolved in dried mineral oil exceeds the normal limit and higher in comparison to wet mineral oil. The amount of hydrogen gas in dried oil also increased but still below the L1 limit. The results when two existing DGA interpretation methods are applied to the faults gases obtained in this investigation are shown in Table 3. The Duvals Triangle method gives 83.3% wrong predictions for biodegradable oil compared to only 33.3% for mineral oil. On the other hand, the Rogers ratio method fails to predict the fault in biodegradable oil in all tests. However this method successfully gives correct prediction for almost all (83.3%) tests for mineral oil.

CONCLUSION

Biodegradable oil has a higher moisture content than mineral oil plus it also has a different chemical chain and properties. These affect the amounts and pattern of hydrocarbon gas quantities dissolved in oil during faults. The experimental results indicate that existing methods for prediction of overheating fault in oil-filled transformers will give incorrect fault interpretations or fail to predict when they are applied to transformers using biodegradable oil for insulation. ACKNOWLEDGMENT The authors would like to acknowledge the support from Energy Australia Insulation Assessment Laboratory with the DGA and moisture measurements, and the Malaysian Government together with Universiti Teknologi Malaysia for sponsorship. REFERENCES 1. C57.104.1991, IEEE Guide for Interpretation of Gases Generated in Oil-Immersed Transformer,, E. The Institute of Electrical and Electronic Engineers, Editor. 1992, The Institute of Electrical and Electronic Engineers, Inc p. 27. N.A Muhamad, B.T. Phung, and TR Blackburn, Comparative Study and Analysis of DGA Methods for Mineral Oil Using Fuzzy Logic, in 8th International Power Engineering Conference. 2007: Meritus Mandarin Singapore. N.A Muhamad, et al., Comparative Study and Analysis of DGA Methods for Mineral Oil, in Power Tech 2007 Conference Switzerland. 2007. FITST3-31, Facilities Instructions, Standards and Techniques in Transformer Diagnostics. 2003, Bureau of Reclamation Hydroelectric Research and Technical Services Group Denver 2003. p. 5-13.

2.

3.

4.

2008 Australasian Universities Power Engineering Conference (AUPEC'08)

Paper P-179 page 4