F

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 C, T2 = thermal faults of temperature 300 C < T < 700 C, T3 = thermal faults of temperature > 700 C. 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

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

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 C), and five cases where only brownish paper was found (T < 300 C), are indicated separately. PaIEEE Electrical Insulation Magazine

0883-7554/02/$17.002002IEEE

T1

T2

T2

T3

T3

Fig. 2. Thermal faults in the paper/oil insulation of transformers in service, identified during inspection by visual marks of: : carbonized paper (T > 300 C); : brownish paper (T < 300 C); : not mentioned.

Fig. 3. Thermal faults in the oil only of transformers in service, identified during inspection by visual marks of: : laminations burnt, eroded, or broken; : circulating currents in bolts, tank, or clamps; : bad contacts in welds, windings, terminals.

per involvement in these cases often results in a CO2/CO ratio < 3, but not always, therefore this ratio should be used with caution when trying to predict whether or not paper is involved in a fault. Figure 3 contains 19 cases of hot spots in oil only (with no paper involved). These were located in the laminations, or were the results of circulating currents in bolts, tank, or clamps, or the results of bad contacts in welds, windings, or terminals, as indicated. Figures 2 and 3 suggest that DGA results appearing in the T3 zone correspond in general to a thermal fault in oil, and those in the T2/T1 zones to a thermal fault in paper. Carbonized paper tends to appear in or close to the T2 zone.

<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. 4 and 5. The corresponding DGA results in ppm can be found in Tables I and II, and their references in Table IV . Figure 4 contains 12 cases of hot spots in paper or in paper/oil insulation. Figure 5 contains 10 cases of hot spots in oil only. Hot spots in paper or paper/oil insulation at all temperatures are located in the T1 and T2 zones. Hot spots in oil at temperatures > 500 C are located in the T3 zone. Hot spots in oil at temperatures < 300 C require some discussion. Most normal oils do not produce significant or measurable amounts of gases at these temperatures (except the carbon oxides). However, 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. These gases are formed the first time the stray gassing oil is heated at these temperatures, up to a concentration plateau (typically, 5 to 250 ppm), which depends on the type of

Fig. 4. Thermal faults simulated in the laboratory; lation; : in paper only.

: in paper/oil insu-

oil. Then, if the oil is heated a second time at these temperatures, no more gassing is observed. This behavior has been related to the oil refining technique used, and investigations are underway at CIGRE to identify the stray gassing oils presently in use and their gas concentration plateaus. Hot spots in stray gassing oils at temperatures < 300 C are indicated in Fig. 5. They are located in the T1 and T2 zones. If stray gassing oils are used during factory heat-run tests or in the early life of transformers, 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. After 9

May/June 2002 Vol. 18, No. 3

PD

<300 C

D1
500 C

800 C

Fig. 5. Thermal faults simulated in the laboratory, in oil only, of temperatures: : > 300 C; : < 250 C.

Fig. 6. Discharges of low energy (D1) and partial discharges of the corona-type (PD) in transformers in service identified during inspection : tracking; : sparking; : small arcing by visual marks of: (OLTC-type) ; : corona-type PDs.

some months in service, however, or after heat treatment of the oil, stray gassing will stop, and DGA results in the T1 and T2 zones will indicate a hot spot in paper only.

Low-Energy Discharges (D1) Twenty-five cases of discharges of low energy (D1) in faulty transformers in serviceas identified during inspection by visual marks of tracking, small arcing, sparking, carbonized pinholes in paper, or carbon particles in oilare indicated in Fig. 6. The corresponding values in ppm can be found on page 35 of [2]. Twelve cases of discharges or PDs of the D1 type simulated in laboratory models are indicated in Fig. 7. The corresponding values in ppm can be found in Table III (cases

6-17), and their references in Table IV . Various configurations have been used in these laboratory models: needleto-sphere, point-to-plane, oil/paper wedge, sparking or small arcs, in oil and/or in paper-oil insulation. Small breakdowns in oil produce relatively more C2H2 (cases 14-16 of Fig. 7). Also, increasing pC amplitudes of the discharges appear to pull the gas composition towards the D1/D2 zone boundary (cases 8-12). A lightning impulse discharge (case 17) appears as a discharge D2 in terms of DGA formation. In transformers in service, D1 discharges of low energy such as tracking, small arcs, and uninterrupted sparking discharges are usually easily detectable by DGA, because gas formation is large enough.

Table I. Hot Spots in Paper and Paper/Oil Insulation Simulated in the Laboratory.
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., C 800 500 300 150 100 300 300 300 225 200 150 140 H2 22,400 670 224 900 470 65 53 219 37 34 14 CH4 22,400 224 4.5 9,180 224 4637 20 39 44 47 30 62 C2H4 13,440 67 900 45 448 5 4.5 3 5.5 2.5 3 2 22 C2H2 1570 C2H6 2690 45 2 1300 10 3 3 10 34 16 68 7500 3000 13,000 CO 112,000 11,200 2 1344 25,312 CO2 56,000 15,680 560 8950 670 51,000 Unit l / g l / g l / g l / g l / g l / l % % ppm % % ppm 2 2 2 2 2 5 6 6 3 6 6 3 Ref.

10

IEEE Electrical Insulation Magazine

Table II. Hot Spots in Oil Simulated in the Laboratory.

Case # 1 2 3 4 5 6 7 8 9 10 Temp., C 800 500 300 235 225 175 168 160 140 120 65 0.7 48 130 86 6 50 55 78 H2 34 1.5 0.2 40 140 8 3 3 22 66 CH4 1.8 0.13 0.5 120 2.5 0.7 2.6 2.6 C2H4 112 C2H2 16 0.3 0.1 11 24 2.5 0.8 1 0.5 62 14 358 283 581 961 1772 2 400 C2H6 CO CO2 l/g l/g l/g % ppm % ml ppm ppm ppm Unit 2 2 2 6 3 6 4 7 10 10 Ref.

PDs of the D1 type, however, may not be detected by DGA unless the PD activity is intense and/or occurring over a long period of time. For example, during high-voltage factory tests on transformers, performed over a relatively short period of time, low-to-medium levels of PD activity are often detected by acoustic or other electrical techniques, but detectable amounts of gases related to these PDs are seldom found by DGA, and visual damage is seldom found by inspection of the transformer. 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 l of gas [5], or about 0.1 ppm in the oil of a transformer, which is the DGA analytical detection limit during factory tests. A discharge energy of 1 to10 mJ is estimated as the minimum required to produce damage or carbonization of paper insulation [6]. An energy of 10 mJ corresponds

to several charge pulses of 100,000 to 1,000,000 pC [5], considered as PDs of high or very high amplitude. Similar levels of pC amplitude have been reported in [7] as necessary to produce visible damage to the paper insulation. This might explain why, during factory or field tests, detectable amounts of gases and damage to the paper insulation are found only in relation to PDs of higher amplitudes [8]. In service, 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. Monitoring by acoustic and electrical techniques is thus more efficient for detecting low-to-medium PDs in the incipient stage, before they become damaging to the paper insulation [8] and produce detectable gases. At this stage, however, their location may be difficult to find because of the absence of visual marks of damage.

1.5

PD

D1 6

D2
7 8 10 15 14 16 11 9 13 17 12

Fig. 7. Electrical/partial discharges simulated in the laboratory: sparking/arcing (6-16); : PDs of the lightning impulse-type (17); PDs of the corona-type (1-5).

: :

Fig. 8. Discharges of high-energy (D2) in transformers in service identified by inspection.

May/June 2002 Vol. 18, No. 3

11

DGA is more useful for determining the onset of the next stage, when PDs just start damaging the paper insulation and may be detected by DGA and visual inspection. PDs may also be harmful if they produce free bubbles in oil, but at this stage enough gas is produced to be easily detectable by DGA.

PDs of the Corona Type In terms of chemical degradation and DGA results, 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. Nine cases of corona-type PDs in faulty transformers in service, as identified during inspection by the formation of

x-wax deposits, are indicated in Fig. 6. The corresponding values in ppm can be found on page 34 of [2]. They all fall in the small PD zone of Fig. 6. Five cases of corona-type PDs simulated in laboratory models are indicated in Fig. 7. The corresponding values in ppm can be found in Table III (cases 1-5) and their references in Table IV . They all fall in the small PD zone of Fig. 7. PDs of the corona type are of very low amplitude (10 to 30 pC) and in the J range [6]. They are usually easily detectable by DGA, however, because they are produced over very long periods of time and within large volumes of paper insulation. They often generate large amounts of hydrogen. They are also often associated with x-wax formation (an un-

Table III. Electrical/PDs Simulated in the Laboratory.

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 5,000 pC in paper/oil wedge 10,000 pC in oil 500,000 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 16,000 6600 88 2240 1950 2240 73 5000 24.3 2240 4480 2240 200 60 890 41 16,000 CH4 3600 1000 9 168 123 157 8 4000 15.7 360 560 560 230 5 110 112 4000 2 0.03 2 45 2 2000 11.2 169 403 450 170 6 84 254 8500 C2H4 14 19 2 45 12 8000 29.8 828 896 940 480 29 700 4536 16,000 C2H2 C2H6 670 38 25 38 90 4 2000 6.4 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 l l ppm % ppm ppm ppm 3 3 6 2 1 2 6 9 9 2 1 2 3 6 3 8 9 Ref.

Table IV. References for Laboratory Results.

Ref. 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. Galand et al. Pugh Lampe et al. Carballeira et al. Samat Oommen Farber Borsi Sokolov Journal/Report Rev.Gen.Elec., 80 (1) 46 Rev.Gen.Elec.,81 (11) 727 Doble Conf. 10-1205 CIGRE Paper 12-05 CIGRE TF-15-01 CIGRE TF-15-01 IEC-TC 10-WG13 EPRI Sub.Equip.Diag.Conf. Electra, 188, 27 CIGRE TF-15/12-01-11

12

IEEE Electrical Insulation Magazine

T2

D1
D1

DT

T3

Fig. 9. Normal operation of tap changers in service OLTC).

Fig. 10. Thermal faults in OLTCs in service identified during inspec: severe coking, : light coking; : tion by visual marks of: Heating.

D1 D1 T3

D2

Fig. 11. OLTCs subjected to a large number of operations without changing the oil: : 500, 3600, 49,000, 202,000, and 269,000 operations; : 2750, 8730, and 47,830 operations (from left to right of Triangle).

Fig. 12. Arcing in OLTCs in service identified during inspection by visual marks of: : arcing; : serious arcing.

saturated hydrocarbon polymer), which increases the tan of the oil and may lead to thermal runaways in instrument transformers and bushings.

ulated in a general-purpose laboratory but only in a high-power laboratory.

High-Energy Discharges (D2) Forty-seven cases of high-energy discharges (D2) in faulty transformers in service are indicated in Fig. 8. The corresponding values in ppm can be found on page 36 of [2]. 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, laminations, etc.). Such cases cannot usually be simMay/June 2002 Vol. 18, No. 3

On-Load Tap Changers (OLTC) Forty-three cases of faults in OLTCs, identified by visual inspection of the equipment, are indicated in Figs. 9-12. The corresponding DGA analyses in ppm are indicated in Table V , and their references in Table VI. Six cases identified as normal operation are indicated in Fig. 9. The normal current-breaking operation of OLTCs corresponds to a discharge of low energy (D1). The main gas formed is C2H2, at a typical rate of 1 to 4 ppm per
13

Table V. Faults in Tap Changers (OLTCs) Identified by Inspection in Service.

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, coking Serious coking Serious coking Severe coking Severe coking Severe coking Severe coking Contacts damaged, 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, stationary contacts Signs of heating on contacts 2750 operations 8730 operations 47,830 operations 500 operations 3600 operations 49,000 operations 202,300 operations 269,000 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 12,125 17,248 43 18,705 580 1100 3130 458 2217 1210 658 217 244 1312 1481 19,700 591 69 1025 30,682 1094 2193 13,918 859 2251 150 17,248 32,444 30,354 6870 12,125 14,320 18,100 9030 1084 9083 14,999 1317 47 210 35 391 CH4 1028 5386 2912 8 3610 617 1600 3255 53,434 7260 3167 749 1160 39,981 18,191 117,00 0 6088 450 3498 13,268 7535 956 6439 843 2159 990 2912 7830 15,258 1028 5386 10,740 16,100 20,300 188 3279 6803 608 12 43 6 164 C2H4 900 6400 4703 11 7136 2691 2010 5668 1486 235,024 17,800 9278 1754 2354 120,319 31,233 142,000 11,433 329 4360 51,608 23,548 1714 10,970 3574 6055 838 4703 15,766 37,352 900 6400 35,839 126,600 200,000 166 9606 21,417 2278 17 102 26 293 C2H2 5500 35,420 18,144 61 45,947 15,045 26 4191 33 1633 453 381 33 189 774 230 3490 193 41 1 12,327 426 125 20,812 5155 18,490 18,144 20,895 10,259 5500 35,420 53,670 60,800 10,500 769 8527 28,381 8739 144 187 482 736 C2H6 79 400 963 2 557 146 221 580 116 55,535 4130 1160 171 688 35,146 7215 44,600 2626 137 2024 6779 3803 307 1221 843 887 643 963 4796 26,032 79 400 3944 19,900 29,700 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 10,500 388 403 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. 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

Table VI. References for Cases of Faults in OLTCs.

Ref. 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. Caldwell Youngblood et al. Youngblood Deskins et al. Haupert et al. H2b Anal. Services Haupert et al. Duval / Pablo Foata / Faucher Journal/Report Doble Conf. 10-205 CIGRE WG 15-01 CIGRE TF 15-01-01 Doble Conf. 6-4.1 Doble Conf. 6-14.1 Doble Conf. 8-16.1 EPRI- LTC Conf. Web site * IEEE EI-Mag.17 (2) 31 Hydro-Quebec report

* Title of paper: The Use of Diagnostic Testing of Insulating Oil for Fault Detection in Electrical Equipment.

Table VII. Additional Cases of Faults Identified by Inspection of Equipment in Service.

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.

T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 T13 T14 T15

Circulating current in aluminum shields Flux overheating Loose bolts, 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 17.6 194 143 1660 259

500 243 4 68 308 3360 305 9630 2720 101 4.5 27 222 1140 640

10 8 1 9 191 10 2 -

200 32 40 25 114 1221 51 1570 579 157 4.6 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 1.5 PU.

May/June 2002 Vol. 18, No. 3

15

Table VIII. References for Cases of Table VII.

Ref. 1 2 3 4 5 6 7 Year 1970 1971 1981 1981 1982 1993 2001 Authors Fallou et al. Dind et al. Oommen et al. Baker et al. Thibault et al. Aubin Sokolov Journal/Report CIGRE Paper 15-07 Doble Conf. 6-1101 IEEE PES Paper # 81 SM 352-4 Doble Conf.10A-01 CIGRE Paper 12-01 Hydro Quebec private communication CIGRE TF 15/12-01-11

current-breaking operation of the OLTC, depending on the type of OLTC. Any faulty behavior of the OLTCs will alter their normal gas composition, especially the relative percentage of C2H4, and is detectable by DGA if the background of gases does not interfere too much. Fifteen such cases where severe coking was identified by inspection of the faulty equipment are indicated in Fig. 10. The chain of events leading to coking usually starts with worn-out contacts of increasing resistance, higher currents flowing through these contacts, and increasing their temperature, carbonization of oil, and carbon deposition on the contacts increasing further their resistance. Most of these cases are in zones T3/T2, meaning that there are more gases formed by the hot spot than by the normal operation of the OLTC. Six cases where light coking or overheating of the contacts were reported by inspection are also identified in Fig. 10. 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, resulting in a gas composition intermediate between faults T3/T2 and D1. In such cases, zone D2 should not be considered as a zone of high-energy arcing but rather as a zone of mixture of faults (DT). Figure 11 indicates eight cases where very large numbers of operations had been made without changing the oil, in two different types of OLTCs. Current-breaking operations increase the amount of carbon particles in oil, which deposit on the contacts, increasing their resistance and resulting in a more and more severe hot spot. Eight cases where arcing was identified by inspection of faulty OLTCs are indicated in Fig. 12. Those appearing to the left of normal operation, deeper into the D1 zone, are related to abnormal arcs of the D1 type in the equipment. When monitoring by DGA, faults occurring in this region can be clearly attributed to abnormal D1s. Those appearing to the right of normal operation are related to more severe arcing or short circuits of the D2 type. When monitoring OLTCs by DGA, faults occurring in this region may be attributed either to D2 or to slight coking. 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, the fault is more likely to be a fault D2 than slight coking. To make sure, change the oil, clean the contacts, run a few current-breaking operations, then another DGA.

Additional Cases For the benefit of readers, 20 additional cases of inspected faults found in the literature are indicated in Table VII, and their references in Table VIII. These cases have not been included in Figs. 2, 3, 7, 9, and 11, but they confirm the conclusions drawn previously. The two cases of burnt oil pumps reported (cases T4 and T11 of Table VII) are located in the T2 zone. Conclusions One hundred and seventy-nine cases of faults in transformers in service, identified by visual inspection, have been examined, as well as 19 cases of faults simulated in the laboratory. From these cases, six main types of faults detectable by DGA can be established (PD, D1, D2, T1, T2, T3). Among these, faults T3 in service tend to be related to hot spots in oil and faults T1 and T2 to hot spots in paper, unless stray gassing oils are used. PDs potentially harmful to the equipment are detected by DGA but may otherwise remain undetected. Finally, the relatively new application of DGA to load tap changers has been examined, with indications on how to reliably detect hot spots and abnormal arcing in this equipment.
Michel Duval graduated from the University of Toulouse in 1966 with a B.Sc. in chemical engineering, and received a Ph.D. in polymer chemistry from the University of Paris in 1970. Since then, he has worked for Hydro Quebecs Institute of Research (IREQ) on electrical insulating oils, dissolved gas analysis, and lithium polymer batteries. A senior member of the IEEE, he holds 12 patents; has authored over 60 scientific papers, book chapters, or international standards; and is very active in several CIGRE and IEC working groups. He may be
IEEE Electrical Insulation Magazine

reached at IREQ, 1800 boul. Lionel Boulet, Varennes, J3X 1S1, Canada, or at duvalm@ireq.ca.

transformer population, in Proc. TechCon Annual Conference, TJ/H2b, 2001. 5. V . Sokolov, V . Mayakov, G. Kuchinsky, and A. Golubev, On-site partial discharge measurements in power transformers, Doble Conferences, 2000. 6. L.E. Lundgaard, Partial discharges in transformer insulation, Report of TF 15.01.04, CIGRE, Paris, 2000. 7. B. Fallou, I. Davies, R.R. Rogers, E.H. Reynolds, F. Viale, A. Devaux, R. Fournie, J. Galand, P . Vuarchex, and E. Dornenburg, Application of physico-chemical methods of analysis to the study of deterioration in the insulation of electrical apparatus, CIGRE, Paris, Paper 15-07, 1970. 8. D. Chu, L.J. Savio, S.R. Lindgren, H.R. Moore, and S.L. Nilsson, Application of in-service monitoring systems during tests at Consolidated Edisons Ramapo station, CIGRE, Paris, Paper 12-205, 1998.

References
1. IEC Publication 60599, Mineral oil-impregnated electrical equipment in serviceGuide to the interpretation of dissolved and free gases analysis, March 1999. 2. M. Duval and A. dePablo, Interpretation of gas-in-oil analysis using new IEC publication 60599 and IEC TC 10 databases, IEEE Electrical Insulation Magazine, vol. 17, no. 2, pp.31-41, 2001. 3. A. dePablo, Gassing in mineral insulating oils. Abnormal hydrogen production, CIGRE TF 15/12-01-11, 2001. 4. V . Sokolov, A. Bassetto, T.V . Oommen, T. Haupert and D. Hanson, Transformer fluid: A powerful tool for the management of an ageing

May/June 2002 Vol. 18, No. 3

17