Methanol and Furan Detections in a Gas to Liquid Hydrocarbon Oil for Indicating Cellulose Insulation

Ageing

S.Y. Matharage1, Q. Liu1*, Z.D. Wang1, Ch. Krause2, P.W.R. Smith3 and P. Mavrommatis4

1The University of Manchester, Manchester, M13 9PL, UK 2Weidmann Electrical Technology AG, Rapperswil, CH 8640, Switzerland

3Shell Global Solutions (UK), Manchester, M22 0RR, UK 4TJ/H2b Analytical Services Ltd, Bromborough, CH62 4SU, UK

*qiang.liu@manchester.ac.uk

Abstract—Ageing assessment of cellulose insulation in a transformer is normally conducted through indirect measurements of chemical markers in oil. In addition to the conventional paper ageing indicator 2-FAL, methanol in oil was proposed as an additional indicator to provide early ageing information. Relationships between the paper ageing and the concentration of the indicators in oil could be affected by the type of oil due to stability and partitioning issues. Therefore it is important to investigate the suitability of the paper ageing indicators for alternative insulating oils. This paper shows the results of a laboratory ageing experiment conducted to investigate the relationships between the paper ageing and the ageing indicators of 2-FAL and methanol in a Gas to Liquid (GTL) hydrocarbon oil. The ageing experiment was conducted at 100 °C for up to 126 days. A conventional mineral oil was also tested as benchmark. Variations of both methanol and 2-FAL against DP of paper aged in the GTL oil were found to be similar to those in the mineral oil.

Keywords—2-FAL, methanol, ageing, transformer, paper insulation, degree of polymerisation, cellulose

I. INTRODUCTION In addition to the roles of being an insulation and a heat

transfer medium, a transformer oil also acts as an information carrier. The information could include the conditions of oil, paper and other components in the transformer. This information is obtained by conducting different measurements in oil such as acidity, moisture, dissolved gases, furanic compounds, alcohols, etc [1-3]. Out of these chemicals, furanic compounds, dissolved carbon oxide gases and alcohols such as methanol and ethanol are used to indicate the ageing state of cellulose paper insulation.

Furanic compounds were first identified as paper ageing indicators during early 1980s [2]. Several furanic compounds including 2-furfuraldehyde (2-FAL), 5-hydroxymethyl-2-furfuraldehyde, 2-furoic acid, furfuryl alcohol, 5-methyl 2-furfuraldehyde and 2-acetyl furan were suggested to be related with paper ageing [2]. However, 2-FAL is commonly used due to its higher concentration in oil and better stability than the others. Experiments have shown that components in paper

including cellulose, hemi cellulose and the ageing by-products of cellulose such as levoglucosan can promote generation of 2-FAL from paper [4]. Furthermore, laboratory experiments have shown that 2-FAL would increase exponentially with paper ageing [3].

Compared to furanic compounds, methanol has a short history as a paper ageing indicator. Methanol was first confirmed as a paper ageing indicator in around 2007 [3]. Ageing experiments have shown that methanol could mainly be generated from cellulose rather than from the ageing by-products of cellulose such as levoglucosan or glucose [3]. Therefore, researchers have suggested that methanol could be generated during the scissioning of glycosidic bonds in cellulose polymer. However, the mechanisms involved with the production of methanol from paper have not been fully understood yet. In addition, methanol has shown higher yield than 2-FAL during the early ageing period and hence it has been suggested to use methanol to indicate early stage of paper ageing [3].

Even though both methanol and 2-FAL are mainly generated from paper, their concentrations have to be measured in oil. However, the concentration of the indicators in oil can be affected by the type of oil due to the stability and partitioning effects. Ageing indicators could react with components of the oil or the ageing by-products in the oil, which could result in misleading information of the ageing state of paper. Furthermore, the partitioning of the chemical indicators between oil and paper depends on the factors such as temperature, polarity of the oil, etc. Therefore it is important to investigate the suitability of the ageing indicators for new alternative insulating oils. This paper investigates the relationships between the paper ageing and the ageing indicators of 2-FAL and methanol in an alternative oil, which is manufactured from natural gas (methane) through gas-to-liquid (GTL) technology.

II. DESCRIPTIONS OF AGEING EXPERIMENT

The ageing experiment was conducted with kraft paper (thermally non-upgraded) and an inhibited GTL oil (Diala S4

ZX-I). In addition a conventional inhibited mineral oil (Gemini X) was used for the purpose of comparison.

In the manufacturing process of the GTL oil, methane is first reacted with oxygen (O2) generating a mixture of carbon monoxide (CO) and hydrogen (H2). Then the mixture is converted into liquid hydrocarbons in a Fischer-Tropsch process. Finally these long chain hydrocarbons are contacted with hydrogen in order to break them (hydrocrack) into smaller branched chain hydrocarbons [5]. Antioxidants are added to these predominantly iso-paraffinic oils before using them in the transformers. These predominantly iso-paraffinic transformer oils are purer and chemically consistent, which could provide better performance over the conventional mineral oils.

Both oil and paper samples were first preprocessed to remove the contaminations. Oil samples were first filtered through a nylon membrane of 0.2 µm. Then the filtered oil samples were dried inside a vacuum oven (<500 Pa) at 85 °C for 48 hours. In parallel paper samples were dried at 105 °C inside an air circulation oven for 24 hours. Next the samples prepared by mixing 200 g of oil with 10 g of paper inside 250 ml glass bottles were left inside the vacuum oven at 85 °C for another 24 hours to allow the impregnation of the paper. Finally the samples were sealed with thermoplastic polyester caps with PTFE sealing gaskets (withstand temperatures up to 180 °C) and left at 100 °C inside an air circulation oven (BINDER FD 115) for up to 126 days. Two samples, one from each oil type were taken out from the oven during each sampling period and left at room temperature for 72 hours before separating oil and paper for various measurements including DP of paper, 2-FAL and methanol in oil.

III. RESULTS AND DISCUSSIONS

A. DP of paperDP indicates the average number of monomer units in

cellulose polymer. It was obtained by measuring the intrinsic viscosity of paper dissolved in cupriethylenediamine solution according to the ASTM D4243 standard. DP measurement requires much less amount of paper samples compared to tensile strength (TS) measurement. Furthermore there is a well-established relationship between DP and TS of paper [6]. Fig. 1 shows the variation of the DP of paper aged in both the GTL oil and the mineral oil. DP of paper aged in both oil types reduces from 1126 to about 500 in a similar manner during the ageing period. This suggests that under the investigated conditions paper ageing in the GTL oil is similar to that of the mineral oil.

B. 2-FAL in oil2-FAL in oil is commonly measured through high

performance liquid chromatography (HPLC) technique. Two different injection techniques namely indirect injection, in which the 2-FAL is first extracted to a solvent before injection and direct injection, in which oil is directly injected into the system can be used to inject samples into the HPLC system. 2-FAL measurement in this work was conducted according to the direct injection technique described in ASTM D5837 standard.

Fig. 2 shows the variation of 2-FAL in both the GTL oil and the mineral oil. During the early ageing period the concentration of 2-FAL in both oil types is less than the detection limit of 0.01 ppm. However, after about 20 days of ageing 2-FAL in both oil types shows an exponential increase with ageing. At the end of the ageing period 2-FAL in both oil types are around 0.2 ppm. This exponential increase of 2-FAL has been observed in previous ageing experiments [3, 7], which could be due to the generation of 2-FAL from cyclic by-products of paper ageing [4].

Fig. 1. Variation of DP of paper aged in oil Gemini X and Diala S4 ZX-I during the ageing experiment conducted at 100 °C

Fig. 2. Variation of 2-FAL in Gemini X and Diala S4 ZX-I during the ageing experiment conducted at 100 °C

C. Methanol in oilAn international standard for measuring methanol in oil has

not been developed yet. Therefore, methanol measurements in this experiment were conducted through an in-house developed method using a headspace gas chromatography mass spectrometry (HS-GC-MS) system [8] based on [9]. Fig. 3 shows the variation of methanol with ageing in both the GTL oil and the mineral oil. Methanol in both oil types increases rapidly up to 1.2 ppm with ageing.

D. Relationship between 2-FAL and DP of paper Fig. 4 shows the variations of 2-FAL in the GTL oil and the

mineral oil against paper ageing measured through DP of paper obtained from the present ageing experiment conducted at 100 °C and a previous ageing experiment conducted at 120 °C [7]. Variations of 2-FAL against DP in the GTL oil and the mineral oil at both 100 °C and 120 °C are similar in nature. 2-FAL plotted in logarithm scale increases linearly up to about 1 ppm till DP reaches 400. However, at the latter stage of ageing and when DP is less than 400, the concentration of 2-FAL in both oil types increases linearly with a higher rate of change than the early ageing stage. Similar observations in the 2-FAL have been reported by other researchers [3].

Furthermore, the relationship between 2-FAL and DP of paper obtained from the present experiments were compared with several commonly used relationships [10]. All these relationships were derived based on laboratory ageing experiments. A commonly used formula is given in (1), in which logarithmic value of 2-FAL is linearly related with the DP of paper.

log10 [ 2FAL ]=A−B ×[DP ] (1)

Where, [2FAL] is the concentration of 2-FAL in oil in ppm, [DP] is the DP of paper and the constants A and B are the correlation constants.

The concentrations of 2-FAL measured in the present ageing experiments are in a similar range to those in the published equations. Furthermore the relationship suggested by Chendong seems to agree with most of the present experimental results better than the other equations. However, a clear deviation from the Chendong equation can be seen in the experimental results for DP values less than 400. Results obtained from the present laboratory ageing experiments suggest that a single linear relationship may not be the best option to relate 2-FAL and DP of paper. Therefore, two different linear relationships for an early stage of ageing (DP>400) and a later stage of ageing (DP<400), respectively, would result in a better fitting.

In addition to the uncertainties in the correlation between 2-FAL and paper ageing, some of the in-service transformers could have abnormally high concentrations of 2-FAL due to the oil contamination. New transformer oils could be contaminated with 2-FAL due to the incomplete removal of 2-FAL, which is used as a solvent during the refining process through solvent extraction method [11]. Furthermore, some recycled insulating oils used in transformers could contain residual 2-FAL [11]. In addition, the concentration of 2-FAL in the transformers with thermally upgraded paper are found to be considerably low compared to those with ordinary kraft paper [3]. All these reasons have resulted in more research on paper ageing indicators that could provide further information for transformer condition assessment.

E. Relationship between methanol and DP of paperMethanol in oil is a new ageing indicator that has shown

some promising results in both laboratory experiments and in-service oil measurements [3, 12]. Laboratory ageing

experiments have shown a higher yield of methanol than 2-FAL and a linear increase during the early ageing period. In addition, a significant amount of methanol has been observed in oil during the ageing experiments conducted with thermally upgraded paper [3]. Furthermore, in-service measurements have shown that more transformers have measurable amounts of methanol than of 2-FAL [3, 12].

Fig. 3. Variation of methanol in Gemini X and Diala S4 ZX-I during the ageing experiment conducted at 100 °C

Fig. 4. Relationship between 2-FAL and DP for the ageing experiments conducted with Gemini X and Diala S4 ZX-I at 100 °C and 120 °C

Fig. 5 shows the variations of methanol in both the GTL oil and the mineral oil against the DP of paper at both 100 °C and 120 °C [7]. Relationship between methanol and DP of paper aged in both oil types were similar till the DP reaches nearly 200. Furthermore, the relationship between methanol and paper ageing is independent from the investigated temperatures. This suggests that similar criteria would be applicable to assess paper ageing in both the GTL oil and the mineral oil through methanol measurement. Similar to those reported in the literature [3], the concentration of methanol in both oil types increased linearly with paper ageing up to about 1.5 ppm till the DP reaches about 400 (2).

[ MeOH ]=2.31−0.0023 × DP (2)

Where, [MeOH] is concentration of methanol in oil in ppm and [DP] is the DP of paper.

However, a reduction in methanol was observed at the latter stage of paper ageing, i.e. when DP reaches around 250 during the ageing experiment conducted at 120 °C. The exact reason for the reduction is yet to be understood. An instability of methanol is suspected to be the reason for the reduction. In acidic mediums, methanol can react with carboxylic acids producing esters, which is called the esterification process. Acidity in both oil and paper increases at the latter stage of ageing, which could react with methanol. Nevertheless, experiments are currently being conducted on this issue.

Fig. 5. Relationship between methanol and DP for the ageing experiments conducted with Gemini X and Diala S4 ZX-I at 100 °C and 120 °C

Furthermore, Fig. 6 shows the variation of methanol and 2-FAL in both the GTL oil and the mineral oil against the DP of paper till the DP reaches 300. From the results it can be seen that the concentration of methanol in both oil types are higher than the concentration of 2-FAL till the DP reaches about 400. These results confirm that methanol is suitable in indicating early paper ageing in both the GTL and the mineral oil.

Fig. 6. Variation methanol and 2-FAL against the DP of paper for the experiments conducted with Gemini X and Diala S4 ZX-I at 100 °C and 120 °C

IV. CONCLUSIONS

A laboratory ageing experiment was conducted at 100 °C to investigate the suitability of methanol and 2-FAL as paper ageing indicators in a new alternative transformer oil

manufactured from natural gas through gas-to-liquid (GTL) technology.

The rates of both paper ageing and production of ageing indicators in the GTL oil were similar to those in the conventional mineral oil under the investigated conditions. Furthermore, the relationships obtained between the paper ageing and the ageing indicators of methanol and 2-FAL in the GTL oil were similar to those of the mineral oil, which confirms that the same criteria used in conventional mineral oils to assess paper ageing through both methanol and 2-FAL can be applied to the GTL oil.

ACKNOWLEDGMENT

The authors would like to express their gratitude to M&I Materials, National Grid, Scottish Power, Shell Global Solutions, TJ|H2b Analytical Services, UK Power Networks and WEIDMANN Electrical Technology for their financial and technical contributions to the transformer research consortium at The University of Manchester. The authors would also like to thank Elizabeth Davenport and the staff in the Mill-B9 laboratory, School of Chemical Engineering & Analytical Science for their support during GC-MS measurement.

REFERENCES

[1] N. Azis, Q. Liu, and Z. D. Wang, "Ageing assessment of transformer paper insulation through post mortem analysis," IEEE Trans. Dielectr. Electr. Insul., vol. 21, pp. 845-853, 2014.

[2] P. J. Burton, J. Graham, A. C. Hall, J. A. Laver, and A. J. Oliver, "Recent developments by CEGB to improve the predictions and monitoring of transformer performance," presented at the CIGER conference paper, 1984.

[3] J. Jalbert, R. Gilbert, P. Tétreault, B. Morin, and D. Lessard-Déziel, "Identification of a chemical indicator of the rupture of 1,4-β-glycosidic bonds of cellulose in an oil-impregnated insulating paper system," Cellulose, vol. 14, pp. 295-309, 2007.

[4] J. Scheirs, G. Camino, M. Avidano, and W. Tumiatti, "Origin of furanic compounds in thermal degradation of cellulosic insulating paper," Journal of Applied Polymer Science, vol. 69, pp. 2541-2547, 1998.

[5] P. Smith, "The benifits of inhibited transformer oils using the latest gas-to-liquid based technology," presented at the Euro Techcon United Kingdom, 2013.

[6] A. M. Emsley, R. J. Heywood, M. Ali, and X. Xiao, "Degradation of cellulosic insulation in power transformers .4. Effects of ageing on the tensile strength of paper," IEE Proc. Science, Measurement and Technology, vol. 147, pp. 285-290, 2000.

[7] S. Y. Matharage, Q. Liu, E. Davenport, G. Wilson, D. Walker, and Z. D. Wang, "Methanol detection in transformer oils using gas chromatography and ion trap mass spectrometer," in 18th In'l Conf. Dielectr. Liquids (ICDL), pp. 1-4, 2014.

[8] S. Y. Matharage, Q. Liu, and Z. D. Wang, "Ageing Assessment of Kraft Paper Insulation through methanol in oil measurement," IEEE Trans. Dielectr. Electr. Insul., 2016 (in press).

[9] J. Jalbert, S. Duchesne, E. Rodriguez-Celis, P. Tétreault, and P. Collin, "Robust and sensitive analysis of methanol and ethanol from cellulose degradation in mineral oils," Journal of Chromatography A, vol. 1256, pp. 240-245, 2012.

[10] D. Feng, Z. Wang, and P. Jarman, "Transmission power transformer assessment using furan measurement with the aid of thermal model," in Condition Monitoring and Diagnosis (CMD), in In’l Conf. Condition Monitoring and Diagnostics (CMD), pp. 521-524, 2012.

[11] S. J. Tee, Q. Liu, Z. D. Wang, G. Wilson, P. Jarman, R. Hooton, et al., "Challenges in oil test database analysis for ageing assessment of in-

service transformers," in 50th In’l Universities Power Engineering Conference (UPEC), pp. 1-5, 2015.

[12] J. Jalbert, R. Gilbert, Y. Denos, and P. Gervais, "Methanol: a novel approach to power transformer asset management," Power Delivery, IEEE Transactions on, vol. 27, pp. 514-520, 2012.