Paper Duval

FE A T U R E A R T I C L E A Review of Faults Detectable by Gas-in-Oil Analysis in Transformers Key Words: Transformer, fault, dissolved gas analysis EC Publication 60599 [1] provides a coded list of faults detectable by dissolved gas analysis (DGA): PD = partial discharges, D1 = discharges of low energy, D2 = discharges of high energy, T1 = thermal faults of temperature < 300 ºC, T2 = thermal faults of temperature 300 ºC < T < 700 ºC, T3 = thermal faults of temperature > 700 ºC. The IEC TC10 databases of DGA results corresponding to faults identified by visual inspection of faulty transformers in service have been presented in a previous paper [2]. The present paper reviews these DGA results in a more user-friendly graphical form. It also reviews the DGA results of laboratory models attempting to simulate these faults, as published in the scientific literature or technical reports. The specific case of on-load tap changers (OLTC) is reviewed much more extensively, and separately, since DGA interpretation in this case must take into account the large background of residual gases resulting from the normal current-breaking operation of the OLTC. Particular attention is I Michel Duval IREQ, Canada Load tap changers, low-temperature hot spots, and partial discharges require particular attention. also given to DGA results related to PDs and low- temperature thermal faults. The Triangle graphical method of representation [1] is used to visualize the different cases and facilitate their comparison. The coordinates and limits of the discharge and thermal fault zones of the Triangle are indicated in Fig. 1, and are not repeated in Figs. 2-13 for clarity. Zone DT in Fig. 1 corresponds to mixtures of thermal and electrical faults. Readers interested in visualizing their own DGA results using the Triangle representation should preferably use triangular coordinate drawing paper, such as the one provided by Keffel & Esser Co (# 46 4493), for better precision. The Triangle coordinates corresponding to DGA results in ppm can be calculated as follows: % C2H2 = 100 x /(x+y+z); %C2H4 = 100y /(x+y+z); %CH4 = 100z /(x+y+z), with x = (C2H2); y = (C2H4); z = (CH4), in ppm. PD 98 T1 80 20 20 T2 % 4 13 60 23 40 H4 C2 CH 4 40 % D1 D2 50 60 Thermal Faults (T1, T2, T3) 40 Thermal Faults in Transformers in Service DT 29 T3 15 20 80 80 60 40 % C2H2 20 Fig. 1. Coordinates and fault zones of the Triangle. Thirty-five cases of thermal faults (hot spots) in faulty transformers in service, identified by visual inspection of the equipment, are indicated in Figs. 2 and 3. The corresponding values in ppm have already been published on page 37 of [2]. Figure 2 contains 16 cases of hot spots in paper/oil insulation. Seven cases where carbonized paper was found upon inspection (T > 300 ºC), and five cases where only brownish paper was found (T < 300 ºC), are indicated separately. PaIEEE Electrical Insulation Magazine 8 0883-7554/02/$17.00©2002IEEE or clamps. Figure 4 contains 12 cases of hot spots in paper or in paper/oil insulation. and those in the T2/T1 zones to a thermal fault in paper. If stray gassing oils are used during factory heat-run tests or in the early life of transformers. eroded. 5 to 250 ppm). : bad contacts in welds. as indicated. Thermal faults simulated in the laboratory. up to a concentration plateau (typically. Figures 2 and 3 suggest that DGA results appearing in the T3 zone correspond in general to a thermal fault in oil. and investigations are underway at CIGRE to identify the stray gassing oils presently in use and their gas concentration plateaus. 4. These were located in the laminations. These gases are formed the first time the stray gassing oil is heated at these temperatures. : brownish paper (T < 300 ºC). or the results of bad contacts in welds. or were the results of circulating currents in bolts. 18. Hot spots in oil at temperatures < 300 ºC require some discussion. which depends on the type of Fig. : circulating currents in bolts. After 9 May/June 2002 — Vol. However. and their references in Table IV . Carbonized paper tends to appear in or close to the T2 zone. but not always. : not mentioned. Thermal faults in the oil only of transformers in service. This behavior has been related to the oil refining technique used. Then. or terminals. 3. tank. identified during inspection by visual marks of: : carbonized paper (T > 300 ºC). They are located in the T1 and T2 zones. 3 . Most “normal” oils do not produce significant or measurable amounts of gases at these temperatures (except the carbon oxides). Fig. windings. Thermal faults in the paper/oil insulation of transformers in service. Hot spots in paper or paper/oil insulation at all temperatures are located in the T1 and T2 zones.T1 T2 T2 T3 T3 Fig. 5. no more gassing is observed. some types of oils (“stray gassing oils”) have been shown in recent studies [3]-[4] (especially among the new oils appearing on the market) to produce gases (mainly H2 and CH4) at the beginning of their life. identified during inspection by visual marks of: : laminations burnt. : in paper/oil insu- oil. 2. therefore this ratio should be used with caution when trying to predict whether or not paper is involved in a fault. it may thus be difficult to tell whether DGA results appearing in the T2/T1 zone indicate a hot spot in paper or stray gassing of the oil. Figure 3 contains 19 cases of hot spots in oil only (with no paper involved). <300° C 500° C 800° C Thermal Faults in Laboratory Models Twenty-two cases of hot spots simulated in the laboratory are indicated in Figs. The corresponding DGA results in ppm can be found in Tables I and II. Hot spots in oil at temperatures > 500 °C are located in the T3 zone. Hot spots in stray gassing oils at temperatures < 300 °C are indicated in Fig. or clamps. tank. terminals. 4 and 5. : in paper only. per involvement in these cases often results in a CO2/CO ratio < 3. No. if the oil is heated a second time at these temperatures. windings. or broken. Figure 5 contains 10 cases of hot spots in oil only. lation. 7. point-to-plane. increasing pC amplitudes of the discharges appear to pull the gas composition towards the D1/D2 zone boundary (cases 8-12). Thermal faults simulated in the laboratory.680 560 8950 670 51. Fig. The corresponding values in ppm can be found on page 35 of [2]. or after heat treatment of the oil. some months in service. oil/paper wedge.5 2. 6. 5.000 CO 112. D1 discharges of low energy such as tracking. in oil only. A lightning impulse discharge (case 17) appears as a discharge D2 in terms of DGA formation.000 15. Table I. Case # 1 2 3 4 5 6 7 8 9 10 11 12 X X X X X X X X X X X X Paper Paper/Oil Temp. Various configurations have been used in these laboratory models: needleto-sphere.000 11. ºC 800 500 300 150 100 300 300 300 225 200 150 140 H2 22. 6.5 3 2 22 C2H2 1570 C2H6 2690 45 2 1300 10 3 3 10 34 16 68 7500 3000 13. in oil and/or in paper-oil insulation. : sparking. sparking. and uninterrupted sparking discharges are usually easily detectable by DGA. Also.312 CO2 56. In transformers in service.400 670 224 900 470 65 53 219 37 34 14 CH4 22.200 2 1344 25..PD <300° C D1 500° C 800° C Fig. stray gassing will stop. : small arcing by visual marks of: (OLTC-type) . and their references in Table IV .5 3 5. Small breakdowns in oil produce relatively more C2H2 (cases 14-16 of Fig.440 67 900 45 448 5 4.5 9. small arcs.180 224 4637 20 39 44 47 30 62 C2H4 13.400 224 4. because gas formation is large enough. Twelve cases of discharges or PDs of the D1 type simulated in laboratory models are indicated in Fig. sparking or small arcs. or carbon particles in oil—are indicated in Fig. : < 250 ºC. Hot Spots in Paper and Paper/Oil Insulation Simulated in the Laboratory. small arcing. however. : corona-type PDs. 7). Discharges of low energy (D1) and partial discharges of the corona-type (PD) in transformers in service identified during inspection : tracking. Low-Energy Discharges (D1) Twenty-five cases of discharges of low energy (D1) in faulty transformers in service—as identified during inspection by visual marks of tracking. 10 IEEE Electrical Insulation Magazine .000 Unit µl / g µl / g µl / g µl / g µl / g µl / l % % ppm % % ppm 2 2 2 2 2 5 6 6 3 6 6 3 Ref. of temperatures: : > 300 ºC. carbonized pinholes in paper. The corresponding values in ppm can be found in Table III (cases 6-17). and DGA results in the T1 and T2 zones will indicate a hot spot in paper only. considered as PDs of high or very high amplitude.5 0.000 to 1. No. before they become damaging to the paper insulation [8] and produce detectable gases. For example. A discharge energy of 1 to10 mJ is estimated as the minimum required to produce damage or carbonization of paper insulation [6].6 2. but detectable amounts of gases related to these PDs are seldom found by DGA. however.000.1 ppm in the oil of a transformer. PDs of the corona-type (1-5). or about 0.6 C2H4 112 C2H2 16 0.Table II. May/June 2002 — Vol.8 1 0. detectable amounts of gases and damage might possibly be produced by a larger number of PDs of smaller amplitude occurring over a longer period of time.7 48 130 86 6 50 55 78 H2 34 1. 3 11 .8 0. which is the DGA analytical detection limit during factory tests. 1. performed over a relatively short period of time. may not be detected by DGA unless the PD activity is intense and/or occurring over a long period of time. This might explain why.5 120 2. during high-voltage factory tests on transformers. during factory or field tests. An energy of 10 mJ corresponds to several charge pulses of 100..5 0. 8.5 62 14 358 283 581 961 1772 2 400 C2H6 CO CO2 µl/g µl/g µl/g % ppm % ml ppm ppm ppm Unit 2 2 2 6 3 6 4 7 10 10 Ref. : : Fig. ºC 800 500 300 235 225 175 168 160 140 120 65 0. Discharges of high-energy (D2) in transformers in service identified by inspection. Similar levels of pC amplitude have been reported in [7] as necessary to produce visible damage to the paper insulation. The reason is most likely that PDs produce only a small amount of gases and damage: PDs of 100 mJ of energy will produce only 5 µl of gas [5].3 0. however. 18. At this stage. Hot Spots in Oil Simulated in the Laboratory. 7.7 2. PDs of the D1 type. : PDs of the lightning impulse-type (17). their location may be difficult to find because of the absence of visual marks of damage. In service.1 11 24 2.2 40 140 8 3 3 22 66 CH4 1. Case # 1 2 3 4 5 6 7 8 9 10 Temp. Electrical/partial discharges simulated in the laboratory: sparking/arcing (6-16). detectable amounts of gases and damage to the paper insulation are found only in relation to PDs of higher amplitudes [8].5 0. and visual damage is seldom found by inspection of the transformer. Monitoring by acoustic and electrical techniques is thus more efficient for detecting low-to-medium PDs in the incipient stage.000 pC [5]. low-to-medium levels of PD activity are often detected by acoustic or other electrical techniques.5 PD D1 6 D2 7 8 10 15 14 16 11 9 13 17 12 Fig.13 0. They are also often associated with x-wax formation (an un- Table III.000 CH4 3600 1000 9 168 123 157 8 4000 15. Samat Oommen Farber Borsi Sokolov Journal/Report Rev.Elec. PDs may also be harmful if they produce free bubbles in oil. Case # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Corona in oil Corona in ASTM D2300 Corona in gas bubbles 30 pC in gas occlusions in paper 30 pC in moist paper 30 pC in moist paper Pds in badly impregnated paper Needle-to-plane Pds > 100 pC Needle-to-plane Pds > 1000 pC 5. Electrical/PDs Simulated in the Laboratory. as identified during inspection by the formation of x-wax deposits. 6. because they are produced over very long periods of time and within large volumes of paper insulation. They are usually easily detectable by DGA.Equip.. Pugh Lampe et al.DGA is more useful for determining the onset of the next stage.000 pC in paper/oil wedge 10.Gen. Five cases of corona-type PDs simulated in laboratory models are indicated in Fig. 7. 1 2 3 4 5 6 7 8 9 10 Year 1970 1972 1974 1976 1982 1987 1997 1999 2000 2001 Authors Thibault et al.3 2240 4480 2240 200 60 890 41 16.Diag. Carballeira et al. References for Laboratory Results.2 169 403 450 170 6 84 254 8500 C2H4 14 19 2 45 12 8000 29. They all fall in the small PD zone of Fig. PDs of the corona type are of very low amplitude (10 to 30 pC) and in the µJ range [6]. The corresponding values in ppm can be found in Table III (cases 1-5) and their references in Table IV .7 360 560 560 230 5 110 112 4000 2 0.000 pC in oil 500.Gen. Galand et al. 7. They often generate large amounts of hydrogen. 27 CIGRE TF-15/12-01-11 12 IEEE Electrical Insulation Magazine .4 25 380 380 2 1 3 500 430 98 45 67 45 6 2240 403 672 160 20 29 45 90 56 67 180 14 CO CO2 790 400 Unit ppm ppm % l ml l % ppm ppm l µl l ppm % ppm ppm ppm 3 3 6 2 1 2 6 9 9 2 1 2 3 6 3 8 9 Ref.Elec. The corresponding values in ppm can be found on page 34 of [2]. 10-1205 CIGRE Paper 12-05 CIGRE TF-15-01 CIGRE TF-15-01 IEC-TC 10-WG13 EPRI Sub. when PDs just start damaging the paper insulation and may be detected by DGA and visual inspection. 188. Ref.81 (11) 727 Doble Conf. but at this stage enough gas is produced to be easily detectable by DGA.000 pC in paper/oil wedge Arcing in oil Point-to-plane discharge in oil Needle-to-sphere discharge in oil Rod-to-rod discharge Needle-to-plane lightning impulse Description H2 16.03 2 45 2 2000 11. Electra. are indicated in Fig.8 828 896 940 480 29 700 4536 16. however. They all fall in the small PD zone of Fig.000 6600 88 2240 1950 2240 73 5000 24. PDs of the Corona Type In terms of chemical degradation and DGA results. 80 (1) 46 Rev. Table IV. PDs of the corona-type occurring in the gas phase of voids or gas bubbles are very different from the PDs of the sparking type (D1) occurring in the oil phase. 6..Conf.000 C2H2 C2H6 670 38 25 38 90 4 2000 6. Nine cases of corona-type PDs in faulty transformers in service. 8. High-Energy Discharges (D2) Forty-seven cases of high-energy discharges (D2) in faulty transformers in service are indicated in Fig. 9. : 2750. etc. and 47. 12. The corresponding values in ppm can be found on page 36 of [2]. and their references in Table VI.000. at a typical rate of 1 to 4 ppm per 13 . Fig. 3600. identified by visual inspection of the equipment. Normal operation of tap changers in service OLTC).T2 D1 D1 DT T3 Fig. 9-12.). : tion by visual marks of: “Heating. The main gas formed is C2H2. Fig. Such cases cannot usually be simMay/June 2002 — Vol. : light coking.000 operations. : serious arcing. The corresponding DGA analyses in ppm are indicated in Table V .830 operations (from left to right of Triangle). No. laminations. 11. saturated hydrocarbon polymer). ulated in a general-purpose laboratory but only in a high-power laboratory. and 269. High-density current or power follow-through could be identified by the presence of molten copper or extensive damage to the windings or other elements of the transformers (tank. 202. 9.” D1 D1 T3 D2 Fig. which increases the tan δ of the oil and may lead to thermal runaways in instrument transformers and bushings. 10. Six cases identified as “normal operation” are indicated in Fig. Thermal faults in OLTCs in service identified during inspec: severe coking. The “normal” current-breaking operation of OLTCs corresponds to a discharge of low energy (D1). Arcing in OLTCs in service identified during inspection by visual marks of: : arcing. 8730.000. are indicated in Figs. OLTCs subjected to a large number of operations without changing the oil: : 500. 18. 49. 3 On-Load Tap Changers (OLTC) Forty-three cases of faults in OLTCs. carbon particles Severe thermal damage to diverter switch Heavy coking on reversing switch Reversing switch badly coked Contacts heavily coked Arcing tips burnt away Light coking and damage of reversing contacts Light coking Light coking Moveable contacts started to coke Excessive heating in nozzles.999 1317 47 210 35 391 CH4 1028 5386 2912 8 3610 617 1600 3255 53.705 580 1100 3130 458 2217 1210 658 217 244 1312 1481 19.191 117. stationary contacts Signs of heating on contacts 2750 operations 8730 operations 47.100 20.144 20.900 29.682 1094 2193 13.895 10.000 operations 202.258 1028 5386 10.766 37.300 188 3279 6803 608 12 43 6 164 C2H4 900 6400 4703 11 7136 2691 2010 5668 1486 235.420 18.125 17.434 7260 3167 749 1160 39.300 operations 269.319 31.839 126.146 7215 44.830 operations 500 operations 3600 operations 49.420 53.981 18.918 859 2251 150 17.00 0 6088 450 3498 13.548 1714 10.327 426 125 20.248 43 18. coking Serious coking Serious coking Severe coking Severe coking Severe coking Severe coking Contacts damaged.045 26 4191 33 1633 453 381 33 189 774 230 3490 193 41 1 12.600 200.259 5500 35.800 9278 1754 2354 120.320 18.670 60.000 166 9606 21.248 32.354 6870 12.381 8739 144 187 482 736 C2H6 79 400 963 2 557 146 221 580 116 55.500 388 403 4893 490 199 4769 3354 4576 381 1070 2240 72 CO2 388 403 4042 482 2606 618 1430 3176 1295 8534 1430 787 602 4905 2152 8270 934 837 5231 2212 Ref. Faults in Tap Changers (OLTCs) Identified by Inspection in Service.535 4130 1160 171 688 35.125 14. Case # N1 N2 N3 N4 N5 N6 T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 T13 T14 T15 T16 T17 T18 T19 T20 T21 A1 A2 A3 B1 B2 B3 B4 B5 D1 D2 D3 D4 D5 D6 D7 D8 500 normal operations 3600 normal operations 2750 normal operations Normal operation Normal operation Normal operation Pyrolitic carbon growth between selector contacts Severe coking and damage of reversing switch Tap shaft failure.800 10.500 769 8527 28.600 2626 137 2024 6779 3803 307 1221 843 887 643 963 4796 26.233 142.000 11.700 8 1136 3212 841 31 12 3 14 CO 29 37 672 191 374 16 732 323 1198 411 309 210 460 307 1110 294 164 773 230 1804 1165 180 1951 672 262 210 29 37 405 150 64 38 381 549 312 62 167 200 29 8400 3454 4758 4042 2954 10.433 329 4360 51.144 61 45.812 5155 18.352 900 6400 35.268 7535 956 6439 843 2159 990 2912 7830 15.490 18.608 23.417 2278 17 102 26 293 C2H2 5500 35.970 3574 6055 838 4703 15.024 17.Table V. 2 2 2 4 7 11 1 5 5 5 5 6 6 6 7 7 7 8 8 9 9 5 5 5 8 8 9 2 2 2 2 2 2 2 2 3 5 8 8 8 10 10 11 14 IEEE Electrical Insulation Magazine .444 30.740 16.947 15.000 operations Arcing on selector switch ring Serious arc damage 50% of arcing tips burnt away Arcing contact burnt off Arcing between springs of contacts Arc in selector switch Selector breaking current in selector tank Arcing on contacts Description H2 6870 12.032 79 400 3944 19.700 591 69 1025 30.100 9030 1084 9083 14. 8-16. Youngblood Deskins et al. Caldwell Youngblood et al.5 PU. Additional Cases of Faults Identified by Inspection of Equipment in Service. 6-14.1 Doble Conf.6 194 143 1660 259 500 243 4 68 308 3360 305 9630 2720 101 4. Web site * IEEE EI-Mag.” Table VII. Ref. Services Haupert et al.LTC Conf.1 Doble Conf. Haupert et al. Duval / Pablo Foata / Faucher Journal/Report Doble Conf. No. 3 15 . 10-205 CIGRE WG 15-01 CIGRE TF 15-01-01 Doble Conf.5 27 222 1140 640 10 8 1 9 191 10 2 - 200 32 40 25 114 1221 51 1570 579 157 4. no visible damage Hot spot in oil pump Overheating of leads Overheating of leads Overheating of bottom core cross bar Overheated core packets Overheating of lead crossover Hot spot in a conductor Overheating of cable to windings * Overheating of paper insulation * Circulating currents in laminations* Burnt oil pump Severe overheating of lead connection 200 101 81 30 465 50 3650 1040 305 12 93 107 220 78 700 184 57 70 200 3100 100 6730 2100 538 17. 1 2 3 4 5 6 7 8 9 10 11 Year 1975 1988 1991 1993 1994 1995 1996 2001 2001 2001 2001 Authors Rogers Carballeira et al.17 (2) 31 Hydro-Quebec report * Title of paper: “The Use of Diagnostic Testing of Insulating Oil for Fault Detection in Electrical Equipment. T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 T13 T14 T15 Circulating current in aluminum shields Flux overheating Loose bolts.Table VI.6 52 35 1880 117 300 61 72 57 219 1530 404 674 220 1900 554 2330 193 410 219 2000 298 203 243 1040 8060 3560 7230 366 4210 1710 6350 1330 2430 1827 1 3 3 3 4 4 4 4 4 5 6 6 6 7 7 O1 O2 Coking of OLTC Coked OLTC contacts 310 75 410 700 472 799 10 1 89 623 180 480 490 8690 4 4 * During heat-run tests at 1. H2b Anal.1 EPRI. 6-4. May/June 2002 — Vol. 18. Case # D1 D2 D3 Inspection Capacitive discharges on loose bolts Arcing in oil and wet paper Trace of overflashing between coils 400 2320 245 H2 40 616 30 CH4 C2H4 60 800 35 C2H2 200 822 245 6 72 5 C2H6 CO 200 9 CO2 1000 72 1 2 7 Ref. References for Cases of Faults in OLTCs. Fifteen such cases where severe coking was identified by inspection of the faulty equipment are indicated in Fig. D2. Conclusions One hundred and seventy-nine cases of faults in transformers in service. Dind et al. From these cases. In such cases. Thibault et al. Baker et al. with indications on how to reliably detect hot spots and abnormal arcing in this equipment. 1 2 3 4 5 6 7 Year 1970 1971 1981 1981 1982 1993 2001 Authors Fallou et al. zone D2 should not be considered as a zone of high-energy arcing but rather as a zone of mixture of faults (DT). faults occurring in this region can be clearly attributed to abnormal D1s. Michel Duval graduated from the University of Toulouse in 1966 with a B. he has worked for Hydro Quebec’s Institute of Research (IREQ) on electrical insulating oils. in chemical engineering. especially the relative percentage of C2H4. depending on the type of OLTC. then another DGA. carbonization of oil. increasing their resistance and resulting in a more and more severe hot spot. These cases have not been included in Figs. To make sure. or international standards. book chapters. Since then. 12. dissolved gas analysis. faults T3 in service tend to be related to hot spots in oil and faults T1 and T2 to hot spots in paper. and 11. as well as 19 cases of faults simulated in the laboratory. deeper into the D1 zone. A senior member of the IEEE. Additional Cases For the benefit of readers. 10. T1. Aubin Sokolov Journal/Report CIGRE Paper 15-07 Doble Conf. Those appearing to the left of normal operation. the fault is more likely to be a fault D2 than slight coking. which deposit on the contacts. He may be IEEE Electrical Insulation Magazine . change the oil. in two different types of OLTCs. 3. Figure 11 indicates eight cases where very large numbers of operations had been made without changing the oil. Ref. and carbon deposition on the contacts increasing further their resistance. Oommen et al. resulting in a gas composition intermediate between faults T3/T2 and D1. T3). When monitoring by DGA. but they confirm the conclusions drawn previously. T2. run a few current-breaking operations. Any faulty behavior of the OLTCs will alter their “normal” gas composition. 20 additional cases of inspected faults found in the literature are indicated in Table VII. Six cases where light coking or overheating of the contacts were reported by inspection are also identified in Fig. Among these. 6-1101 IEEE PES Paper # 81 SM 352-4 Doble Conf. and their references in Table VIII. 10. Finally.D. unless stray gassing oils are used. The chain of events leading to coking usually starts with worn-out contacts of increasing resistance. 2. When monitoring OLTCs by DGA. identified by visual inspection. Those appearing to the right of normal operation are related to more severe arcing or short circuits of the D2 type. faults occurring in this region may be attributed either to D2 or to slight coking. higher currents flowing through these contacts. and received a Ph. and is detectable by DGA if the background of gases does not interfere too much. and is very active in several CIGRE and IEC working groups. Eight cases where arcing was identified by inspection of faulty OLTCs are indicated in Fig. the relatively new application of DGA to load tap changers has been examined.10A-01 CIGRE Paper 12-01 Hydro Quebec private communication CIGRE TF 15/12-01-11 current-breaking operation of the OLTC. One way to make the distinction is to look at the number of cur16 rent-breaking operations since the last oil change: if it is low. and increasing their temperature. he holds 12 patents. 7.Table VIII. and lithium polymer batteries. Most of these cases are in zones T3/T2.Sc. Current-breaking operations increase the amount of carbon particles in oil. The two cases of burnt oil pumps reported (cases T4 and T11 of Table VII) are located in the T2 zone. meaning that there are more gases formed by the hot spot than by the normal operation of the OLTC. in polymer chemistry from the University of Paris in 1970. have been examined. 9. has authored over 60 scientific papers. D1. Their unexpected presence in the D2 zone may be explained by the formation of similar amounts of gases by the hot spot and by the normal operation of the OLTC. PDs potentially harmful to the equipment are detected by DGA but may otherwise remain undetected. clean the contacts. are related to abnormal arcs of the D1 type in the equipment. References for Cases of Table VII. six main types of faults detectable by DGA can be established (PD. TechCon Annual Conference. G.L. dePablo. P . V .R. A. “Transformer fluid: A powerful tool for the management of an ageing May/June 2002 — Vol.31-41. Vuarchex. “On-site partial discharge measurements in power transformers.V . Davies. Sokolov. 18.” CIGRE. R. Paris.R. T. or at duvalm@ireq. no. pp. L. B. Fournie. Golubev. 2001. References 1. 1970. Hanson. 2000. 4. Haupert and D. A. IEC Publication 60599. Savio. “Application of physico-chemical methods of analysis to the study of deterioration in the insulation of electrical apparatus.” CIGRE. and E. Lindgren. 8. Lundgaard. “Gassing in mineral insulating oils.” CIGRE TF 15/12-01-11. “Application of in-service monitoring systems during tests at Consolidated Edison’s Ramapo station. No.04. Abnormal hydrogen production. 2001. Viale. Moore. I. 1998.R. “Interpretation of gas-in-oil analysis using new IEC publication 60599 and IEC TC 10 databases.ca. 2. Reynolds. D. 2000. 2. Paris. transformer population. 7. Sokolov. Varennes.E. Galand. V . R. Chu. Kuchinsky. F. A.” in Proc. dePablo. Oommen.J.” IEEE Electrical Insulation Magazine. 1800 boul. 3 17 .01. vol.reached at IREQ. Mayakov. TJ/H2b. E. V . Dornenburg.” March 1999. Lionel Boulet.H. CIGRE. and S.” Doble Conferences. “Mineral oil-impregnated electrical equipment in service—Guide to the interpretation of dissolved and free gases analysis. Devaux. S. Nilsson. Fallou. Rogers. M. 6.” Report of TF 15. 17. 5. H. and A. T. J. Paris. Bassetto. L. Paper 12-205. “Partial discharges in transformer insulation. Duval and A. J3X 1S1. 3. Paper 15-07. 2001. Canada.