Aging of Kraft Paper in Natural Ester Dielectric Fluid

8/3/2019 Aging of Kraft Paper in Natural Ester Dielectric Fluid

1/5

Proceedings o f 14 Intemational Conference on D ielectric Liquids (ICDL 2002), Gra z (Austria), July 7-1 2,2 002

Aging of Kraft Paper in Natural Ester Dielectric FluidC.P. McShane K.J. Rapp J.L. CorkranDielectric Fluids Thomas A. Edison Technical Center Transformer Products

Franksville, WI 531 26 USACooper Power Systems Cooper Power Systems Cooper Power Systems

Waukesha, WI 5318 8 USA Waukesha, WI 5318 8 USAG.A. GaugerThomas A. Edison Technical CenterCooper Power SystemsFranksville, WI 531 26 USA

Abstract: Kraft transformer insulation paper aged in,natural ester(vegetable oil) diele cvic fluid was compared to identical paper agedin conventional transformer mineral oil. Sealed steel aging tubescontaining copper, aluminum, Kraft paper, and dielectric fluid(minera l oil and natural este r) were aged at 150C for 500, 1000,2000, and 4000 hours. The extent of paper degradation after agingwas determined using paper tensile strength, paper degree ofpolymerization, and furanic compounds in the aged fluid. Watercontents of fluids and paper were compared. Paper aged inconventional transformer oil degraded at a significantly faster ratethan in natural ester dielectric fluid. Paper in mineral oil reachedthree criteria for E E E end-of-life (50% retained tensile strength, 25%retained tensile streng th, and degree of polymerization of 200) withinthe first 1000 hours. After 4000 hours of aging, paper in natural esterdid not degrade to any of the E E E end-of-life criteria. At 4,000hours, the paper aged in natural ester,retained about 55% of theoriginal tensile strength and a degree of polymerization of about 280.Paper aged in conventional transformer oil degraded to the samevalues in about 315 and 390 hours, respectively - an order ofmagnitude faster. The reduced paper-aging rate in natural ester isprimarily attributed to the fluid maintaining the paper in a very drystate.

INTRODUCTIONNatural esters (vegetable oils) formlllated as dielectric fluidshave environmental and fue safety advantages overconventional transformer mineral oil [I]. In thermal evaluationcomparisons of transformer insulation systems, productiondistribution transformers filled with natural ester dielectricfluid exhibited less deterioration than did the correspondingtransformers using mineral oil 121.The environmental properties of natural esters are such that inGermany they are classified as non-hazardous to waters [3].Aquatic biodegradation tests [41 of the natural ester dielectricfluid used in this experiment found >99% metabolizedconversion to CO 2, equivalent to com pounds defined asultimately biodegradable. In acute trout toxicity tests [SI, he

J. LuksichDielectric FluidsCooper Power SystemsWaukesha, WI 531 88 USAsame fluid had no observable effect on fish at lOOOmgh, th ehighest concentration tested.Transformers using natural ester fluids deliver very importantimprovements in fire safety compared to those using mineraloil. Natural ester dielectric fluids have fn e points in the rangeof 350-360C; conventional mineral oil has a fire point ofabout 155C. Two natural ester dielectric fluids are recognizedas less-flammahle per Section 450.23 of the U S . NationalElectrical Code [6].Accelerated aging tests of distribution transformers gave earlyindications that the rate of paper aging is fluid-dependent 121.An earlier study quantified this dependence for thermallyupgraded Kraft insulation paper [7]. This work examines theaging rate dependence on fluid type for plain (not thermallyupgraded) Kraft insulation paper.

EXPERIMENTALTh e thermal a ging procedure and sam ple preparation methodsare identical to those previously described 171. Sealed steelaging tubes contained 28.4g of Kraft insulation paper dried toa water conten t of 0.76w t%, 350ml of dielectric fluid, andtypical transformer proportions of copper and aluminum. Thetubes were aged at 1 5 0 T and evaluated after 500, 1000,2000,and 4000 hours.Th e extent of paper aging wa s determined using changes in thetensile strength (TS) and the degree of polymerization (D,P).The total concentrations of four predominant furanic productsof paper degradation were measured in the aged fluids. Thewater contents of both the papers and fluids were alsodetermined. Whiteley Ltd. Grade K 0.255 presspaper ooper Power Systems Tranelec@nhibited insulatin g mineral oiland Envirotemp@FR3- natural ester dielectric fluid

0-7803-7350-2/0U$17,00 0.2002 EEE 173


2/5

e 1000 UFigure 1. Kraft paper insulation after 150C sealed tube aging innatural ester and mineral oil.

RESULTSThe results of aging at 150"C, summarized in Table 1, showthat paper insulation ages much slower in vegetable oil than inconventional transformer oil.Th e difference in appearance between th e papers aged in thetwo fluids is striking (Figure 1). At 1000 hours of aging, theTable 1. Results of Kraft paper sealed tube aging at 150C inmineral oil an d natural ester. Water content of dielectric fluid isgiven both as absolute content and percent saturation. Total furaniccompounds are given as mg furans per liter of fluid per initial kg ofpaper.Time (hrs) 0' 500 1000 2000 4000Water Content of Paper [wt%linmineraloil 0.76 0.67 2.44 3.39 3.36in natural ester 0.77 0.27 0.21 0.04 0.05Water Con tent of Fluid [% saturation @ 20C; (mgkg)]

mineral oil 2 ( I ) 30 (18) 58 (35) 61 (36) 75 (45)natural ester 2 ( 24 ) 9 (89) 11(116) 2 (21) 2 (25)Tensile Strength of Paper [MPa]

inmineraloil 119 53.2 21.9 17.8 6.9innatural ester 121 80.8 74.9 63.6 69.8Degree of Polymerization of Paperin&neraloil 1119 217 110 73 12innaturalester 1225 380 320 306 274Furanic Compound Content [mgN kg]mineraloil nd 1055 1462 670 602naturalester nd 191 234 151 62'baseline values

.\5 P 7 -w

0 lm 2ow 3 m 4 mAging Time (hours)

cc @-- -- - --0 " " " " " " " ' "

Figure 2. Retained tensile strength versus time for Kraft paper agedin natural ester and mineral oil. New paper tensile strength is aboutI20MPa. Vertical error bars are k 1 standard deviation; horizontalerror bars represent time-at-temperature uncertainty. Publishedretained tensile strength data 18.91 are included for comparison.

paper in mineral oil was dark brown and brittle; the paper inthe natural ester fluid remained flexible and was only slightlydiscolored. The 2000 an d 4000 hour mineral oil papers werefragile and difficult to handle. Th is was no t the case with thenatural ester-aged papers. T he color a nd fragility of the paperscorrelate well with changes in their TS and D,P, and theproportional furanic difference between the two fluids.Tensile StrengthThe decrease in TS over t ime (Figure 2) shows a steep initialslope of paper degradation in both natural ester and mineraloil. T he highest relative rate of pape r degradation takes placein the first 500 hours of aging in both fluids. T he relative ratesin both fluids decrea se substantially afte r 100 0 hours.In terms of absolute degradation, the paper in mineral oil,already below 50% retained TS at 500 hours, w as helow 20%retained TS value at 1000 hours. The decline in TS of thepaper aged in mineral oil is comparable to that seen in otherstudies [8,9]. The paper aged in vegetable oil degraded toahout 70% of its original TS at 500 hours. After 4000 hours,the retained TS w as about 58%.Degree of PolymerizationAnother measure of cellulose degradation is degree ofpolymerization (D,P). Th e chan ge in D,P o ver time is show nin Figure 3, and is similar to the change seen in TS. A rapidD,P decrease in the first 500 hours of aging was observed.Paper aged in mineral oil degraded to about 20% of the

174


3/5

'1Aging Time (hours)

Figure 3. Retained degree of polymerization (D,P) versus time forKraft paper aged in natural ester and mineral oil. New paper D,P isI175 k50. Horizontal error bars represent time-at-temperatureuncertainty. Published retained D,P data [10,11] are included forcomparison (data athibuted to S chroff& Stannett from [lo]).original D,P at 500 hours, in lin e with other published results[10,11]. T he retained D,P of pape r age d in natural ester fluidremained above 20% after 4000 hours of aging.Furanic Com poundsFuranic compounds in oil are used to observe the insulationdegradation ,of in-service transformers [121. Figure 4 showsthe concentrations of furanic compounds generated by paperdegradation over time. Th e furan evolution from paper aged inthe natural ester fluid was up to an order of magnitude lowerthan p aper aged in conventional transformer oil.Th e concentration of furan s reached a maximum at 1000 hoursin both fluids, supporting both the TS and D,P findings. After1000 hours, the concentration of furans in both fluidsdecreases.Water ContentThe water content of the paper in mineral oil increasedsignificantly. The water content of the paper in natural esterdecreased. By the end of the test, water content of paper agedin mineral oil was orders of magnitude greater than of paperaged in natural ester.In mineral oil, the water content of the paper at 500 hoursdecreased slightly from the initial 0.75wt%. At 1000hours, thewater content increased to 2.4wt%, and continued to increaseto about 3.3wt% fo r the remainder of the test.

," ~0 1ow 20w 3wo 4w oAglng Time (hours)Figure 4. Furanic compound content versus ime for natural ester andmineral oil after sealed tube aging of Kraft paper. Horizontal errorbars represent time-at-temperature uncertainty.

The w ater content of paper in natural ester decreased to 0.27%after 500 hours and continued to decrease. At 2000 and 4000hours, the paper contained abou t 0.05wt% water.Similar trends were seen in the water content of the fluids(Table 1) . Viewing the water content as percent saturation atroom temperature, both fluids started with water contents of2% saturation. The water in mineral oil increased steadily,reaching 75% saturation at 4000 hours. The water in thenatural ester rose gradually to 1 1 % saturation at 100 0 hours,then decreased to 2% saturation at 2000 an d 4000 hours.

-5 0.3 ~E0.2 --0.10.0

0 tw o 2wo 3w o 4w oAging Time (hours)

Figure 5. Water content versus ime for Kraft paper aged in naturalester and mineral oil. Paper was originally dried to a water contentof about 0.76wt90. Vertical error bars are i 1 standard deviation;horizontal error bars represent time-at-temperature uncertainty.

175


4/5

DISCUSSIONThe TS, D,P, and furanic content results establish that paperages slower in natural ester dielectric fluid than inconventional transformer oil. The paper aged in mineral oildegraded beyond three recognized IEEE end-of-life criteria[13]: 50% TS at 370 hours, 25% TS at 915 hours, andD,P=200 at 555 hours. At 4000 hours, the paper aged in thenatural ester fluid did not reach any IEEE end-of-lifeconditions.In both systems, the concentrations of furanic compounds roseto a maximum, then decreased, maintaining a factor of 4-10times more furans in mineral oil. The furans themselvesdecompose over time. Although the furanic contents of bothfluids decrease in a similar way, the rationale for the reductionis different for each fluid. In the case of paper in mineral oil,more of the cellulose was destroyed and converted to furanicproducts. In the natural ester fluid, less cellulose aging tookplace to produce furanic compounds.Th e water in paper data suggest an explanation for the contrastin paper degradation seen between fluids [12]. Theacceleration of aging due to water has been known fo r at least40 years [14]. As paper thermally degrades it produces water,promoting further degradation through hydrolysis. In aconventional transformer paper-in-oil system, the degradationis autocatalytic [15]. Th e partitioning of the water between thepaper and fluid depends on the characteristic polarity (affinityfor water) of each.An essentially non-polar fluid, such as mineral oil, preventsmost of the water generated during cellulose degradation fromleaving the paper. Although the water content of the mineraloil steadily increased to 75% saturation at 4000 hours, this isonly 45mgkg in terms of absolute water content . The paperalso was at approximately 75 % water saturation.The opposite effect is seen for paper in the more polar naturalester. The natural ester has a higher affinity for water thanmineral oil. The water, attracted to the ester fluid, is liberatedfrom the paper. Removing water from the paper limits thepaper degradation rate.The water generated during cellulose degradation migrates outof the paper and into the natural ester. One would expect thewater in'natural ester to increase proportionally. However, thenatural ester remains dry at less than 200C'[17].

CONCLUSIONSThese results sho w that Kraft insulation paper takes at least anorder of magnitude longer to reach IEEE end-of-lifeconditions in natural ester than in conventional transformeroil. Applying natural ester dielectric fluids in transformersshould result in improved insulation performance.

REFERENCES[ I ] McShane, C.P., Gauger, G.A., Luksich, J. "Use of NaturalVegetable Oil Esters as Dielectric Coolant", CIRED 15IhInternational Conference on Electricity Dishibution, June 1-4,

1999, Nice, France[21 McShane, C.P., Gauger, G.A., Luksich, J. "Fire ResistantNatural Ester Dielectric Fluid and Novel Insulation System forIts Use", EEWPES Transmission & Distribution Conference,April 12-16,1999 , New Orleans, USA

Administrative Regulation on the Classification of SubstancesHazardous to Waters, Water Hazard Classes (VwVwS), Annex1 Substances Non-Hazardous to Waters Pursuant to Number1.2a, Index No. 660,1999, GermanyAerobic Aquatic Degradation, US EPA OPPTS 835.3100, TestReport No. MI.-304.98, Thomas A. Edison Technical Center,Cooper Power Systems, Franksville, USA, 1998

[3]

[4]

[5 ] "Final Report: Acute Trout Toxicity Testing for TwoEnvirotemp FR3 Formulations": Procedure 203. OECD"Guidelines for Toxicity of Chemicals", 1993a, and "BiologicalTest Method: Acute Lethality Test Using Rainbow Trout",Environment Canada, 1990 (1996 Amendments); GlobalToxInternational Consultants, Guelph, Canada, Dec. 1, 1999

176


5/5

[6] National Electrical Code, NFPA 70, 2002 Edition, NationalFire Protection Association, Quincy, MA, USA171 McShane, C.P., Rapp, K.J., Corkran, J.L., Gauger, G.A.,Luksich, J. Aging of Paper Insulation in Natural EsterDielectric Fluid, 2001 IEEEEE S Transmission& DistributionConference, Oct. 28 -Nov. 2,20 01, A tlanta, USA

Darveniza, M., Saha , T.K., Hill, D.J.T., Le,T.T. Investigationsinto Effective Methods for Assessing the Condition ofInsulation in Aged Power Transform ers, IEEE T rans PD, Vol.13, No. 4, pp. 1214.1223, Oct. 1998Hill, D.J.T., Le,T.T., Darveniza, M., Saha, T. A Study ofDegradation of Cellulosic Insulation Materials in a PowerTransformer. Part 2: Tensile Strength of Cellulose InsulationPaper, Polymer Degradation and Stability, Vol. 49, pp. 429-435.1995

[IO1 Em sley, A.M.. S tevens, G.C.. A Reassessment of the LowTemperature Thermal Degradation of Cellulose, EE E 6*International Conference on Dielectric Materials,Measurem ents, and Ap plications, pp.229-32, Sept. 7-10, 1992,Manchester, UK[Ill Miyoshi, A., A New Additive for Improving the ThermalAeging Characteristics of Insulating Paper, IEEE Trans EI,Vol. El-IO (I), 13, 1975

[SI

[91

[121 Unsworth, J., Mitchell, F. Degradation of Elecmcal InsulatingPaper Monitored with High Performance LiquidChromatog raphy, IEEE Trans E I, Vol. 25, No. 4, pp. 737-746,Aug 1999[I31 IEEE Guide for Loading Mineral-Oil-ImmersedTransformers, IEEE Std 257.91-1995, April 25 , 1996, IEEE,New York, USA[141 Clark, F.M. Insulatine Materials for Design and Enaineeringpp. 287-290 , Wiley, USA, 1962[151 Hill, D.J.T ., Le,T.T., Darveniza, M., Saha, T. A Study of theDegradation of C ellulosic Insulation Materials in a PowerTransformer. Part 3: Degradation Products of CelluloseInsulation Paper, Polymer Degradation and Stab ility, Vol. 5 I ,pp. 211-218, 199 6[161 Huang, M.R., Li, X.G. Thermal Degradation of Cellulose andCellulose Esters, J. Appl. Polym er Sci.. Vol. 68, pp, 293-304,1998.U71 Oommen, T.V., Lindgren, S.R. Bubble Evolution fromTransformer Overload, IEEWES Transmission &

Distribution Conference, April 12-16, 1999, New Orleans, USA

Documents

Aging of Kraft Paper in Natural Ester Dielectric Fluid